Philosophical Foundations of The Nature of Law

PHILOSOPHICAL FOUNDATIONS
O F T H E N A T U R E O F L AW
This page intentionally left blank
Philosophical
Foundations of the
Nature of Law
Edited by
WIL WALUCHOW
and
STEFAN SCIARAFFA
1
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
# The several contributors, 2013
The moral rights of the authors have been asserted
First Edition published in 2013
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Crown copyright material is reproduced under Class Licence
Number C01P0000148 with the permission of OPSI
and the Queen’s Printer for Scotland
British Library Cataloguing in Publication Data
Data available
ISBN 978–0–19–967551–7
Printed in Great Britain by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Contents
List of Contributors vii

Abbreviations ix
Introduction xi
PART I. FURTHERING DEBATE BETWEEN

LEADING THEORIES OF LAW
1. The Explantory Role of the Weak Natural Law Thesis 3
Mark C. Murphy
2. In Defense of Hart 22
Matthew H. Kramer
3. Law’s Authority is not a Claim to Preemption 51
Kenneth M. Ehrenberg
4. The Normative Fallacy Regarding Law’s Authority 75
Arie Rosen
5. The Problem about the Nature of Law vis-à-vis Legal Rationality
Revisited: Towards an Integrative Jurisprudence 101
Imer B. Flores
P A R T II . TH E P O W E R O F L E G A L S Y S T E M S
6. Law as Power: Two Rule of Law Requirements 129
Bruno Celano
7. A Comprehensive Hartian Theory of Legal Obligation: Social
Pressure, Coercive Enforcement, and the Legal Obligations of Citizens 152
Kenneth Einar Himma
8. Law and the Entitlement to Coerce 183
Robert C. Hughes
PART III. CONCEPTUAL ANALYSIS

9. Farewell to Conceptual Analysis (in Jurisprudence) 209
Andrei Marmor
10. What Do We Want Law to Be? Philosophical Analysis and
the Concept of Law 230
Natalie Stoljar
vi Contents
PART IV. NEW DIRECTIONS
11. Legal as a Thick Concept 257
David Enoch and Kevin Toh
12. Making Old Questions New: Legality, Legal System, and State 279
Keith Culver and Michael Giudice
13. Legal Disagreements and the Dual Nature of Law 301
Andrea Dolcetti and Giovanni Battista Ratti
14. Is There One Right Answer to the Question of the Nature of Law? 322
Dan Priel
Index 351
List of Contributors
Bruno Celano is Professor of Philosophy of Law at the Faculty of Law, University of
Palermo.
Keith Culver is Professor and Director of the Okanagan Sustainability Institute, University
of British Columbia.
Andrea Dolcetti is a doctoral student in Law at St Hilda’s College, University of Oxford.
Kenneth M. Ehrenberg is Assistant Professor of Philosophy at the University of Alabama.
David Enoch is Professor of Philosophy and Jacob I. Berman Professor of Law at The
Hebrew University of Jerusalem.
Imer B. Flores is Professor-Researcher at the Legal Research Institute and the Law School of
the Universidad Nacional Autónoma de México.
Michael Giudice is Associate Professor of Philosophy, York University, Canada.
Kenneth Einar Himma lectures at the University of Washington School of Law.
Robert C. Hughes is a Postdoctoral Scholar in the Law and Philosophy Program at the
University of California, Los Angeles.
Matthew H. Kramer is Professor of Legal and Political Philosophy at the University of
Cambridge and a Fellow of Churchill College, Cambridge.
Andrei Marmor is Professor of Philosophy and Maurice Jones Jr. Professor of Law at the
University of Southern California.
Mark C. Murphy is the McDevitt Chair of Religious Philosophy in the Department of
Philosophy at Georgetown University.
Dan Priel is Assistant Professor at Osgoode Hall Law School.
Giovanni Battista Ratti is Lecturer in Legal Philosophy at the Faculty of Law, University of
Genoa.
Arie Rosen is an Emile Noël Post-Doctoral Fellow at New York University School of Law.
Stefan Sciaraffa is Assistant Professor in the Department of Philosophy, McMaster Univer-
sity.
Natalie Stoljar is Associate Professor at the Department of Philosophy, McGill University.
Kevin Toh is Associate Professor of Philosophy, San Francisco State University.
Wil Waluchow is Professor in the Department of Philosophy, McMaster University; the
Senator William McMaster Chair in Constitutional Studies; and an Adjunct Member of the
Graduate Faculty of Osgoode Hall Law School, Toronto.
Abbreviations
CA Court of Appeal
CAL conceptual analysis of law
HTSO Hartian Theory of Social Obligation
Ill. Rev. Stat. Illinois Revised Statutes
KB King’s Bench
NDPR Notre Dame Philosophical Reviews
NJT normal justification thesis
PCIJ Permanent Court of International Justice
RoL Rule of Law
SCR Supreme Court Reports (Canada)
Introduction
In May 2011, the McMaster University Program in Legal Philosophy sponsored a

conference on the theme The Nature of Law: Contemporary Perspectives (<tnl.
mcmaster.ca>).* Over 100 specialists and leading figures in legal theory partici-
pated. The conference featured seven keynote addresses and twenty-one presenta-
tions culled from a pool of almost eighty submissions. As the number and quality of
conference submissions and participants attest, the nature of law is a popular and
vibrant area of inquiry. With the aim of advancing inquiry in this area, we have
compiled a number of papers from the conference and three commissioned works
on the conference theme.
Part I of the volume comprises five chapters that advance traditional debates
within analytic jurisprudence. In the first contribution to this part, Mark Murphy
argues that natural law legal theory does not respect the boundaries of one of these
debates—the debate that pits positivists who hold that social facts alone are the
fundamental determinants of what norms count as law in any given legal system
against non-positivists who hold that moral considerations necessarily couple with
social facts to determine the laws of any such system.
Hartian legal theory is the leading contemporary exemplar of positivist legal
theory. On the Hartian view, the laws of a legal system are those specified by that
system’s rule of recognition. The rule of recognition is the standard that the
system’s officials converge in following and that specifies the system’s criteria of
legal validity. As such, the rule of recognition depends on social fact, for it is a
matter of social fact what criteria of legal validity a system’s officials converge in
following. Hence, on the Hartian view, legal systems’ laws are fundamentally
determined by social fact, i.e. the relevant system’s rule of recognition.
Ronald Dworkin’s interpretive theory of law is a leading exemplar of non-
positivist theory.1 Dworkin holds that in any legal system, the truth or falsity of
propositions of law, e.g. propositions to the effect that some action or class of
actions is legally permitted or forbidden, turns in part on whether the prohibition
or permission in question is consonant with the morally best account of the
principles of justice that underlie and unify the legal system’s constitutive complex
of rights and duties.
*
We would like to thank the Social Sciences and Humanities Council of Canada for generously
supporting this conference.
1 See Dworkin (1986).
xii Introduction
Mark Murphy argues that those who group natural law theory with Dworkinian
legal theory as a version of non-positivism are mistaken. On Murphy’s view, only
those versions of natural law theory that subscribe to the strong natural law thesis
are non-positivist theories. This thesis holds that unjust or egregiously unjust
norms are not law. Murphy argues that though many mistakenly hold otherwise,
the strong natural law thesis is not one of natural law theory’s defining commit-
ments. Rather, the key and defining commitment of natural law theory is the weak
natural law thesis.
The weak natural law thesis holds that there is some normative feature N such
that all instances of law either possess that normative feature or are defective
instances of law. According to this thesis, a norm need not possess its distinctive
normative feature to be law, but it must be constitutionally capable of possessing it.
To see this idea, consider that a hunk of pulverized metal is no clock, for it is
constitutionally incapable of telling time, whereas a stopped clock in need of
winding is a clock, for it is constitutionally capable of telling time. Similarly, a
norm may be law though it fails to fulfill its distinctive normative function (on
Murphy’s account, this function is to provide decisive reasons for action) so long as
it remains constitutionally capable of doing so.
Murphy argues that a secondary distinction within natural law theory is between
strong natural law theorists who hold that norms that are unjust or egregiously
unjust are constitutionally incapable of fulfilling law’s function and those natural
law theorists who hold that unjust norms may nonetheless be constitutionally
capable of fulfilling law’s function. Only the former set of natural law theorists
are non-positivists (in the sense we have described); the latter, perhaps surprisingly,
are positivists (in the sense we have described).
Murphy adds that though non-positivism is not a defining commitment of
natural law theory, the natural law approach to legal philosophy qualitatively differs
from traditional legal positivism in another respect. Whereas natural law theory
holds that a full understanding of law requires understanding both law’s distinctive
normative function and the ways in which a norm might fail to achieve this
function while remaining constitutionally capable of performing it, traditional
positivist legal theory overlooks these crucial facets of law’s nature.
A second debate within analytic jurisprudence concerns the proper role of moral
evaluation in formulating theories about the law. For the most part, both sides of
this debate agree that the legal theorist should develop an account of law that is true
to the understanding and behavior of the participants in legal practice.2 However,
they disagree about the standards that govern the process of distilling the unruly
and indeterminate observational data regarding these practices into an account of
the law. One line of theorists, descriptivists (also referred to as methodological
positivists), hold that the appropriate standards of theory construction include only
metatheoretic standards, such as coherence, simplicity, clarity, and consilience,
2 Cf. Leiter (2007: 171–3).

Introduction xiii
familiar from other domains of inquiry.3 On the descriptivist account, the best legal
theory is the one that best satisfies this complex of metatheoretic standards. A rival
school, methodological non-positivism, seeks to add moral considerations to the list
of metatheoretical standards that govern the formulation of an account of law. On
one version of methodological non-positivism, the governing moral standard favors
the account of law that, if widely accepted, would lead to better moral consequences
overall than rival accounts.4 On a second version, the governing moral standard
favors accounts of law that cast the law in a better moral light than its rivals. For
example, one account of law is to be preferred to another if it provides a more
satisfactory explanation of why laws necessarily are legitimately authoritative.5
Many methodological positivists are also first-order positivists (referred to here
simply as positivists).6 In parallel, a number of methodological non-positivists are
also first-order non-positivists (referred to here simply as non-positivists). Dworkin
is a clear example of a member of this category. However, these categories need not
line up so neatly. For example, L. Murphy (2001) and MacCormick (1985) have
argued for a form of first-order positivism on the ground that its adoption would
lead to better consequences. A question remaining for Mark Murphy is whether he
would not only deny (as we have detailed) that first order non-positivism is a
defining commitment of natural law theory, but also if he would similarly deny
(contra Finnis) methodological non-positivism’s status as a defining commitment.
If he denies this latter commitment as well, then, from his perspective, a natural law
theorist could occupy three (I-III) of the four quadrants depicted in Table 1 that are
commonly taken to divide the major schools within analytic jurisprudence.7
In his contribution, Imer Flores argues for a version of first-order non-positivism.
Moreover, at one point he purports to argue for this claim on the basis of a
methodological non-positivist approach. Thus, if we were to take Flores at his word,
we would have to group him with Dworkin as both a methodological and a first-order
non-positivist. However, we might take issue with Flores’s characterization of his
argument. It appears that his argument distinctively and interestingly occupies a
relatively sparsely populated quadrant in the logical space described herein. That is,
he seems to argue for a version of first-order non-positivism on the basis of a
methodologically positivist approach.
One of Flores’s key premises is that the proper way to develop an account of law
is to begin with important and valuable features of legal practice, some of which
3 Some leading descriptivists are Waluchow (1994: 19–27), Dickson (2001), and Leiter (2007: chs.
6 and 7). See also the intramural skirmish between Dickson, who holds that importance (including
moral importance) is a metatheoretic standard governing the construction of a descriptive theory of
law, and Leiter, who rejects this view.
4 See e.g. L. Murphy (2001) and MacCormick (1985).
5 See Waluchow (1994), Dickson (2001), and Leiter (2007: particularly 164–72) for discussions
critical of a number of leading methodological non-positivists. See Dworkin (1986: chs. 1 and 2) and
Finnis (1980: ch. 1) for seminal defenses of methodological non-positivism.
6 See Table 1.
7 In light of Murphy’s discussion, Finnis’s natural law theory would probably be best categorized as
a first-order positivist theory, for on his view unjust norms may be law, though, in his terms, they
would not be central instances of law. See Finnis (1980: ch. 1).
xiv Introduction
Table 1
First-Order Positivism First-Order Non-

Positivism
Quadrant Ia Quadrant IIb
Methodological Positivism Hart, Raz, Waluchow, Dickson, Greenberg?, Toh? Imer
Leiter Flores?
Methodological Non- Quadrant IIIc Quadrant IVd
Positivism L. Murphy, MacCormick, Finnis Dworkin, Perry
a See Hart (1994), Raz (2009: chs. 1 and 2) and n. 3 of this Introduction.
b See Greenberg (2004) and Toh (2008). We are not certain that either theory belongs in Quadrant III, but
we think the case could be plausibly made.
c MacCormick (1985) and L. Murphy (2001).
d Dworkin (1986) and Perry (1995).
may be important or valuable because they are morally important or valuable.

Flores holds that the legal theorist should then reflect on these important features
with an eye to developing an account of law’s essential features. Upon such
reflection, the theorist may conclude that some of these important and valuable
features are also essential features.
In accordance with this approach to theorizing about law, Flores identifies five
features of legal rationality that he takes to be important. He adduces these features
through a close reading of a number of American judges’ and legal scholars’
discussions of legal rationality. These features are standards of legal rationality
that these judges and scholars hold out as standards that govern judicial decision
making.
Flores argues that the law is constituted by the conclusions that result from the
integrated application of these five standards of legal rationality to a background of
legal materials, such as court decisions or statutes. Because one of these law-
determining rational standards (the standard of ethical rationality) requires the
judge to look to moral considerations to determine legal content, Flores’s conclu-
sion is a form of first-order non-positivism. Note, however, that nowhere in Flores’s
argument can one find premises characteristic of methodological non-positivism.
That is, Flores does not seem to argue that the theorist must hold that the law is, in
part, determined by the foregoing standards because such an account of law casts
legal practice in the best possible moral-political light, nor does he seem to argue for
his account on the basis that adopting it would lead to better moral consequences.
Hence, he seems to reach a first-order non-positivist view on the basis of a
methodological positivist approach to developing his account of law.
In their respective entries, Arie Rosen and Kenneth Ehrenberg address aspects
of a third major debate within analytic jurisprudence—the debate between soft
(inclusive) and hard (exclusive) legal positivists. Hard positivist legal theory
holds that in no case may moral considerations play a role in determining what
is law in a legal system, whereas the soft positivist legal theory holds that though
moral considerations may not be fundamental determinants of law, they may be
Introduction xv
derivative determinants. A soft positivist allows that moral principles may be

conditions of legal validity, but only insofar as the relevant law-determining
social facts assign them a law-determining role.8 For example, a relevant social
fact might be the provisions of the legal system’s written constitution specifying
that legislative enactments that violate fundamental tenets of equality are not law.
By contrast, hard positivists hold that despite their superficial appearance as
conditions of legality, the moral considerations enumerated in such provisions
do not play a law-determining role. Rather, provisions of this sort specify a limited
legislative power to the judiciary to overturn laws that violate the enumerated
moral considerations.
Joseph Raz offers the seminal defense of hard positivism.9 Ehrenberg and Rosen
target a key premise of Raz’s central argument for this view. This premise holds that
the law claims legitimate authority. This claim is shorthand for the more complex
claim that legal officials, particularly judges, systematically act as if laws are
authoritative; hence Raz imputes to these officials an operative belief that the law
is authoritative.
The claims the law makes for itself are evident from the language it adopts and from the
opinions expressed by its spokesmen, i.e., by the institutions of the law. The law’s claim to
authority is manifested by the fact that legal institutions are officially designated as “author-
ities,” by the fact that they regard themselves as having the right to impose obligations on
their subjects, by their claims that their subjects owe them allegiance, and that their subjects
ought to obey the law as it requires to be obeyed (i.e., in all cases except those in which some
legal doctrine justifies breach of duty).10
Though it is impossible to do full justice here to Raz’s authority-based argument for
exclusive legal positivism, its broad outlines are as follows. Raz holds that a norm
cannot, as a matter of conceptual necessity, be authoritative if those subject to it
must engage in moral reasoning to determine the norm’s existence and content. Raz
adds that though legal officials individually may be mistaken about the content of a
legal system’s laws, they cannot be systematically confused about law’s nature. He
then concludes that it must be possible for subjects to determine the existence and
content of any law without recourse to moral reasoning, for were it the case that any
norm’s status as law could be determined only through moral reasoning, then,
contra judges’ systematic operative belief in the authority of all laws, such norms
could not, as matter of conceptual necessity, be authoritative. Thus, soft positivism
(as well as non-positivism) must be false, for this theory countenances the possibil-
ity of laws whose existence and content can only be determined through moral
reasoning.
Ehrenberg observes that on Raz’s account, to say that the law claims authority
entails that each directive of the authority is a reason for its subjects to refrain from
acting for at least some of their reasons that favor not conforming to the directive in
8 See e.g. Kramer (2004) and Waluchow (1994).

9 Raz (1985). See also Marmor (2001) for a more recent discussion and defense of exclusive legal
positivism.
10 Raz (1985: 300).
xvi Introduction
question. In other words, the law claims to provide its subjects with preemptive
reasons. He argues that though Raz correctly observes that subjects often acquiesce
to the law’s directives because they take them to be authoritative, and hence
preemptive reasons, he calls into question Raz’s claims that the law claims author-
ity. The core of Ehrenberg’s contribution consists in a number of arguments (as he
counts them, 2½ arguments) in support of the thesis that the law claims a right to
direct its subjects’ actions but does not claim authority.
Ehrenberg’s argument threatens to undermine Raz’s authority-based argument
for exclusive legal positivism, for as we have seen, the argument relies on the
premise that law claims authority. Though Ehrenberg has set the stage for the
development of this further implication of his argument, he does not pursue it here.
Unlike Ehrenberg, Rosen does not deny that the law claims authority. Rather, he
contests Raz’s understanding of this claim. Rosen argues that legal officials entertain
a variety of views about what law’s claim to authority entails and the conditions
under which such claims are justified. Thus, on Rosen’s view, the law claims
authority, but it only does so ambiguously. By contrast, and in Rosen’s view
mistakenly, Raz interprets law’s claim to authority in terms of a particular meaning
of authority and understanding of the condition of its justification. For Raz, to say
that law claims authority is to say that the law claims to provide its subjects with
content-independent and preemptive reasons and that such a claim is justified
because its subjects will likely better comply with the first-order reasons that apply
to them by trying to follow the law’s directives rather than the first-order reasons
that apply to them directly. Thus, on Raz’s account law claims authority, where
authority is to be understood in terms of Raz’s service-conception of authority.
Rosen argues that Raz imputes this content to law’s claim to authority on the
ground that this is the correct theory of authority’s content and ground of
justification. Rosen describes Raz’s imputation as an instance of a normative fallacy.
He holds contra Raz that the proper characterization of the content of law’s claim to
authority is not the claim that legal officials would make if they reasoned correctly
about authority, but rather the claims they do make. And, as we have seen, Rosen
argues that, as a matter of social fact, the understanding of the content of law’s
claim to authority varies from official to official, and hence law’s claim to authority
must be characterized as an ambiguous one. Among the implications that Rosen
draws from his conclusions is that Raz’s argument for exclusive positivism fails, for
this argument relies on the premise that law claims authority, understood in terms
of Raz’s service-conception of authority.
Kramer’s contribution is an extended response to Scott Shapiro’s recent book,
Legality.11 Shapiro argues that the Hartian account of the rule of recognition suffers
from a number of irremediable deficiencies, and he offers his planning theory of law
as an alternative basis of positivist legal theory that does not suffer from these same
failings. Kramer rebuts Shapiro’s criticisms of Hart’s theory in detail and concludes
11 Shapiro (2010).
Introduction xvii
that the basic Hartian picture is in no need of the significant modifications that
Shapiro urges.
Part II of the volume comprises three works that address the nature of the
normative ties that bind the law’s subjects to their legal duties. John Austin, an early
legal positivist, is well known for his reductivist account of this bond. On his view,
the bond of legal obligation is constituted by a threat of coercive enforcement; a
citizen is bound by a legal duty only insofar as there is a probability that he will
suffer the sovereign’s sanction should he fail to conform to the duty. Hart famously
rejects Austin’s conception of legal obligation, holding that Austin provides an
account of the way in which law obliges yet fails to explain legal obligation.
Hart introduces the notion of a social rule to explain the nature of legal
obligation. On Hart’s account, a social rule exists within a group if the group’s
members take the internal point of view with respect to a particular pattern of
behavior. An agent who takes the internal point of view takes the relevant pattern of
behavior to be a standard that she and her fellow members of the relevant group
ought to follow. According to Hart, the defining standard of behavior of a social
rule is an obligation for the group’s members insofar as the group’s members take
the internal point of view with respect to the pattern, they take conforming to the
standard to concern matters of great social importance, and they readily apply social
pressure to enforce conformity with the rule.
In his contribution, Ken Himma seeks to resolve a problem that he finds with
Hart’s account of legal obligation. Himma worries that though Hart’s account is
well suited to explain the legal obligations of a legal system’s officials, it fails to
explain the nature of citizens’ legal obligations. As Himma notes, Hart’s legal
officials participate in the rule of recognition (a social rule constituted by a standard
of treating certain norms as legally valid) from the internal point of view, they take
the rule of recognition to address matters of great societal importance, and they
readily enforce the rule by means of insistent social pressure. Thus, the rule of
recognition that the rule recognizes can be readily characterized as a Hartian
obligation binding on legal officials. However, argues Himma, Hartian theory
cannot similarly explain the sense in which citizens are subject to a legal obligation,
for unlike legal officials, citizens need not take the internal point of view with
respect to their system’s rule of recognition or any of its laws. In sum, the challenge
Himma poses for Hartian legal theory is to explain the sense in which citizens
who neither participate in the legal system’s rule of recognition nor accept, from
the internal point of view, the laws it validates are nonetheless bound by legal
obligation.
Himma argues that resolving this puzzle requires modifying the Hartian con-
ception of legal obligation with the help of resources found within the Hartian
conception of obligation itself. A key element utilized in Himma’s modified
account is Hart’s notion of social pressure. On Himma’s reconstruction, a key
defining element of citizens’ legal obligation is felt social pressure in the form of the
threat of sanction. Thus, Himma argues that Hartian theory must reintroduce
elements of Austin’s conception of obligation if it is to explain the sense in which
citizens are bound by legal obligations.
xviii Introduction
Whereas Himma focuses on whether the threat of coercion is an element of legal
obligation that cannot be eliminated, Robert Hughes examines whether the right to
enforce the law through coercion is an essential element of the entitlement to make
law. Hughes argues that these two rights are conceptually separable and that the
state may enjoy the latter entitlement, yet lack an accompanying right to enforce
the law. Hughes adds that whether an entitlement to make law is accompanied by a
right of coercion is dependent on a number of contingent factors that may very well
vary across legal systems or from law to law within a legal system; hence, the state
may have a right to make and coerce citizens into obeying some laws but have only
the entitlement to make others.
In his contribution, Bruno Celano focuses on the power of a legal system,
understood as the legal system’s ability to direct and influence the behavior of its
citizens. His contribution develops at length the idea that the power of a legal
system that instantiates the rule of law is distinctive. In short, Celano argues that
such a system influences and guides the behavior of its subjects in a way that pro
tanto respects their dignity as persons. By contrast, argues Celano, legal systems that
do not realize these rule of law virtues treat their subjects as things to be manipu-
lated rather than rational agents worthy of guidance by rules.
The volume’s third part comprises two contributions that directly address
conceptual analysis of law, the method of theorizing about law that is commonly
recognized as the predominant method practiced by contemporary positivist phil-
osophers of law.12 Conceptual analysis of the law (CAL) involves testing candidate
analyses of law against intuitions. According to CAL, the legal theorist should reject
analyses that have counterintuitive implications about whether some object counts
as law or whether some feature is a necessary feature of law in favor of analyses that
better accord with such intuitions.
If the common view that CAL is the predominant methodology in contemporary
positivist legal theory is correct, then one would think that the seminal figure in
contemporary analytic jurisprudence, H. L. A Hart, must have primarily employed
this approach in developing his theory of law. Provocatively, Andrei Marmor argues
that the widely held belief that Hart was engaged in conceptual analysis of law is
false. As Marmor sees it, Hartian legal positivism seeks to explain law in terms of
certain social facts, which in turn can be explained by people’s actual conduct,
beliefs, and attitudes. Marmor argues at length that this approach is not conceptual
analysis; rather it is a reductive analysis of legal phenomena. One important
implication that Marmor draws from his argument is that many criticisms of
Hart fail to engage with his project insofar as they target Hart’s alleged commit-
ment to conceptual analysis rather than his reductivist project.
Unlike Marmor, Natalie Stoljar joins the ranks of theorists who take conceptual
analysis to be the predominant approach to legal theory; however, she joins Sally
Haslanger in distinguishing between three kinds of conceptual analysis. As Stoljar
puts it, this first form of conceptual analysis is an a priori inquiry that sets forth the
12 See e.g. Coleman (2002) and Leiter (2007: particularly, 175–99).

Introduction xix
concept’s users’ understanding of their concept. Hence, this first form of inquiry
mirrors CAL described herein. The second is a descriptive inquiry that investigates
the phenomena in the world that the concept at issue tracks, with the aim of
providing an analysis of an operative concept.13 The third is an ameliorative inquiry
that theorizes about our “legitimate purposes” in using a concept and yields a target
concept. Stoljar observes that we find each of these methodological approaches in
contemporary legal theory.
We can relate Stoljar’s typology to the distinction between methodological
positivism and methodological non-positivism discussed herein. Stoljar’s first two
forms of conceptual analysis are compatible with methodological positivism in the
sense that moral considerations do not directly govern these forms of analysis. That
is, under these approaches, it does not count in favor of an analysis that it describe a
morally better object or that the widespread acceptance of the concept would lead
to better consequences. By contrast, the ameliorative approach is a version of
methodological non-positivism.
Stoljar argues for methodological pluralism. On her view, any of the three forms
of conceptual analysis may be appropriate, depending on the context of inquiry and
the interests of the theorist. From this perspective, the question at the heart of the
debate between methodological positivists and non-positivists is ill formed. The
question is not whether methodological positivism or non-positivism is correct;
rather, the better formed question queries the circumstances under which these
differing approaches are warranted.
The fourth and final part of the volume comprises four works that suggest novel
approaches to inquiry into law’s nature.
David Enoch and Kevin Toh explore the possibility that “law” is a thick moral
term. A thick moral concept is a term that contains both descriptive and evaluative
elements. Enoch and Toh offer “courageous” as an example of a thick moral
concept. To say that some action is courageous is not merely to say that it ought
to be done or endorsed (the evaluative element of the concept); it is also to describe
it to some extent, for example, perhaps, it is a case of overcoming one’s fears (the
descriptive element of the concept).
The central motivation behind Enoch and Toh’s project is the conjecture that if
“law” is a thick moral term, then philosophers of law could learn much by applying
the extensive and sophisticated literature on thick moral terms to the legal context.
Specifically, Enoch and Toh explain why extant debates and discussions regarding
thick moral concepts promise to shed new light on (i) “Hart’s distinction between
legal judgments from the internal point of view and legal judgments from the
external point of view, and the corresponding distinction between internal and
external statements;” (ii) Dworkin’s methodological non-positivism; (iii) “the
judgments from ‘the legal point of view’ that Raz posited and highlighted;” and
(iv) “the notion of a basic norm that played a crucial role in Kelsen’s jurisprudential
theorizing.”
13 See Leiter (2007: 183–202) for a discussion and argument for a version of this form of
conceptual analysis.
xx Introduction
The contribution of Andrea Dolcetti and Giovanni Ratti discusses the nature of
theoretical disagreement amongst legal officials, an issue that Dworkin has brought
to the forefront of debate in legal theory. As Dworkin stipulates, theoretical
disagreement is disagreement among a legal system’s officials with respect to the
system’s grounds of law, or in Hartian positivist terms, criteria of legal validity.14
Dworkin argues that Hartian legal positivism should be rejected, for it cannot
explain this form of disagreement amongst legal officials.
On the positivist view, agreement about the criteria of legal validity is a necessary
precondition of law. Insofar as officials fail to generally converge in accepting
criteria of legality, there is no law. Such a failure can be global or local. A global
failure might be failure to agree on any criteria at all, in which case there would be
no legal system and no law.15 A local failure would occur with the context of global
agreement. For example, a system’s officials might generally agree with respect to a
number of key criteria of legal validity, e.g. statutory enactments and judicial
precedents are law, but disagree about whether, say an entrenched statutory
enactment is law. In such a case, there would be a legal system in which statutory
enactments and judicial precedents are law; however, it would be indeterminate
whether entrenched enactments are law in this legal system.
Dworkin observes that judges embroiled in local theoretical disagreements do
not act as if the law is indeterminate. Rather, they purport to argue about what the
law is. Thus, the Hartian positivist cannot take judicial behavior at “face value” in
such cases; rather, the Hartian positivist must hold that the judges are confused or
insincere. Dworkin holds that such an explanation is unsatisfying, and hence
Hartian positivism cannot adequately explain theoretical disagreement.16
Dolcetti and Ratti affirm Dworkin’s contention that legal officials disagree about
the criteria of legal validity; however, they argue that such theoretical disagreement
is limited in scope. Dolcetti and Ratti distinguish between two types of criteria of
legal validity: criteria that set out the sources of law (e.g. legislative enactments and
court decisions) and criteria that set out the proper method of interpreting those
sources. Dolcetti and Ratti observe that Dworkin only identifies disagreements
about methods of interpretation of sources. Moreover, they argue at length that
whereas disagreement about methods of interpretation is a commonplace feature of
legal systems, disagreement about sources is not. They further observe that despite
widespread disagreement about interpretive method, there is nonetheless massive
agreement at a decisional level for all but the hardest cases, i.e. those that reach the
highest courts of appeal. Dolcetti and Ratti argue that such massive decisional
convergence is compatible with widespread interpretative disagreement because, in
most cases, the easy cases, the variety of interpretative approaches lead to the same
result.
14 Dworkin (1986: chs. 1 and 2). See Shapiro (2007) and Smith (2011) for recent discussions of
theoretical disagreement.
15 See e.g. Hart (1994: 120–2) for discussion that suggests this distinction between kinds of
16 See Leiter (2009) for a detailed reconstruction and rebuttal of Dworkin’s argument from
Introduction xxi
Dolcetti and Ratti acknowledge that with respect to hard cases, the cases where
the various interpretative methods lead to conflicting results, the positivist must
hold that the law is indeterminate. Moreover, they acknowledge that judges
typically do not act as if the law is indeterminate in such cases. On the contrary,
in these circumstances, they typically write their opinions as if they are identifying
rather than making the law. However, Dolcetti and Ratti deny that these observa-
tions pose any deep problem for Hartian legal theory. Rather, they argue that such
judicial posturing is nothing more than an expression of “judicial dogma.”
We could have placed Dolcetti and Ratti’s contribution in the first section of this
volume, for they enter into a line of ongoing debate between Dworkinian and
Hartian legal theorists. Also, as we have just noted, these authors side with the
Hartian camp. However, they do so in an interesting and distinctive way. They
argue that law has a dual nature. On their view, whereas the criteria of legal validity
that specify legal sources are settled in well-functioning legal systems, the criteria of
legal validity that specify interpretative approaches to these sources typically are not.
Moreover, this conclusion suggests (though the authors likely would not follow this
suggestion) the possibility of a novel hybrid view—Hartian positivism for legal
sources and Dworkinian interpretive legal theory for canons of interpretation.17
Culver and Giudice’s contribution is a subtle and complex argument in support
of legal pluralism. It is also a call for legal theorists to direct their attention to the
problem of legal continuity, as originally formulated by Raz. Raz argues that Hart’s
legal theory has no resources for distinguishing between changes to a legal system’s
rule of recognition that amount to mere changes or modifications of the system
from more radical changes that amount to revolutionary breaks or discontinuity
between legal orders. Though Raz does not hazard a full account of this distinction,
he suggests the beginnings of one. Raz posits that all legal systems supervene on a
larger system of political and social norms that constitute a single state. Moreover,
Raz holds that discontinuities in the larger systems of norms that constitute any
given state explain discontinuities in the state’s supervening legal system. Hence, on
Raz’s view, a necessary preliminary to resolving the problem of continuity of legal
systems is further theorizing about the nature of the political and social norms that
constitute the state.
Culver and Giudice observe that the political and social norms that constitute
the state are neither confined to particular state boundaries nor are they clearly
unified. For example, in part, norms that constitute any given state include rules of
sovereignty and recognitional legitimacy that govern state relations and define the
jurisdictional boundaries of states. Moreover, Culver and Giudice note that some of
these social and political norms carve out exceptions to the reach of a legal system in
ways that contradict the legal system’s norms. They note, for example, the existence
in some societies of state-constituting social norms that carve out autonomous
zones where the state legal system’s laws do not apply (e.g. with respect to the
17 Cf. Sciaraffa (forthcoming).

xxii Introduction
favelas in Brazil and communities in Micronesia), despite the fact that from the
perspective of the legal system, the laws actually do apply.
As the examples suggest, the systems of social and political norms that carve out
state-based legal systems also carve out domains of normative order at the intrastate,
transstate, and suprastate levels that are outside of the scope of any particular state’s
normative order. Culver and Giudice suggest that because non-state and state
normative orders similarly supervene on social and political norms, there is no
obvious principled reason to label only one class of these normative orders, the
state-level normative order, as a legal system; rather it may be more illuminating to
describe each level of normative order as a different form of legal order. Thus,
Culver and Giudice tentatively embrace a form of legal pluralism.
The final contribution to the volume is Dan Priel’s skeptical reflection on
inquiry into law’s nature. Priel holds that traditional contemporary analytic juris-
prudence is misguided insofar as it seeks to provide a unitary account of law’s
nature. Priel’s case in point is a comparison of the American and British legal
systems. Priel argues that despite their shared history, British and American forms
of life and traditions have led to different conceptions of the relationship between
politics and law that animate these respective countries’ legal systems. On Priel’s
account, in American legal institutions, the domains of law and politics are fused,
whereas British legal institutions preserve the autonomy of each vis-à-vis the other.
Hence, the British and American legal systems’ respective fundamental natures
differ. Priel argues that this difference is telling, for if the nature of these two
historically connected legal systems differ fundamentally, it is unlikely that other
systems without such close historical ties do not similarly differ.
Note that Priel does not deny that there may be a bird’s-eye perspective from
which the theorist can discern a common nature between legal systems. His point is
that the traditional focus on this common ground has obscured fundamental
differences between legal systems. Moreover, this focus has led theorists, particu-
larly Dworkinian theorists and Hartian positivists, to engage in needless debate
about law’s unitary nature when, in fact, they have been reporting on the divergent
particular natures of different legal systems—i.e. Dworkinians reflecting the Ameri-
can frame and positivists the British.
References
Coleman, J. (2002). “Methodology,” in J. Coleman and S. Shapiro (eds.), The Oxford

Handbook of Jurisprudence and Philosophy of Law. Oxford: Oxford University Press.
Dickson, J. (2001). Evaluation and Legal Theory. Oxford: Hart Publishing.
Dworkin, R. (1986). Law’s Empire. Cambridge, MA: Harvard University Press.
Finnis, J. (1980). Natural Law and Natural Right. Oxford: Oxford University Press.
Greenberg, M. (2004). “How Facts Make Law,” Legal Theory, 10: 157–98.
Hart, H. L. A. (1994). Concept of Law. Second edition. Oxford: Oxford University Press.
Kramer, M. (2004). Where Law and Morality Meet. Oxford: Oxford University Press.
Leiter, B. (2007). Naturalizing Jurisprudence. Oxford: Oxford University Press.
Introduction xxiii
Leiter, B. (2009). “Explaining Theoretical Disagreement,” University of Chicago Law
Review, 76: 1215–50.
MacCormick, N. (1985). “A Moralistic Case for A-Moralistic Law,” Valparaiso Law Review,
20: 1–41.
Marmor, A. (2001). Positive Law and Objective Value. Oxford: Oxford University Press.
Murphy, L. (2001). “The Political Question of the Concept of Law,” in J. Coleman (ed.),
The Postscript: Essays on the Postscript to the Concept of Law. Oxford: Oxford University
Press.
Perry, S. (1995). “Interpretation and Methodology in Legal Theory,” in A. Marmor (ed.),
Law and Interpretation. Oxford: Oxford University Press.
Postema, G. (1998). “Jurisprudence as Practical Philosophy,” Legal Theory, 4: 329–57.
Raz, J. (1985). “Authority, Law and Morality,” The Monist, 69: 295–324.
Raz, J. (2009). Between Authority and Interpretation. Oxford: Oxford University Press.
Sciaraffa, S. (forthcoming). “Explaining Theoretical Disagreement and Massive Decisional
Agreement: The Justificatory View,” Problema.
Shapiro, S. (2007). “The ‘Hart–Dworkin’ Debate: A Short Guide for the Perplexed,” in
A. Ripstein (ed.), Ronald Dworki. Cambridge: Cambridge University Press.
Shapiro, S. (2010). Legality. Cambridge, MA: Harvard University Press.
Smith, D. (2011). “Theoretical Disagreement and the Semantic Sting,” Oxford Journal of
Legal Studies, 30: 635–61.
Toh, K. (2008). “An Argument Against the Social Facts Thesis (and Some Additional
Preliminary Steps Toward a New Conception of Legal Positivism),” Law and Philosophy,
27: 445–504.
Waluchow, W. (1994). Inclusive Legal Positivism. Oxford: Oxford University Press.
PART I
FURTHERING DEBATE
BETWEEN LEADING
THEORIES OF LAW
1
The Explanatory Role of the Weak
Natural Law Thesis*
Mark C. Murphy
My topic is a thesis of analytical jurisprudence that I have labeled elsewhere the

“weak natural law thesis.”1 This thesis has come in for various sorts of abuse: that it
cannot be clarified in a way that does not collapse into some other, better
understood thesis about the law; or that it is not really a thesis of analytical
jurisprudence at all; or that it is a confusion to think of this thesis as a distinctively
natural law thesis.
I think that these criticisms of the weak natural law thesis are baseless. But I also
think that defenders of the weak natural law thesis, like me, have invited them
by the way that they have put that thesis forward. Sometimes natural law theorists
put that thesis forward as if its interest were borne simply by the thesis itself and
not by the thesis as playing a specific role in theorizing about the nature of law.
So what I propose to do is this. Following a brief reminder of what the weak natural
law thesis claims (section 1), I will address these various criticisms of the weak
natural law thesis (section 2). After making a few unoriginal remarks about the aims
of theories of law (section 3), I will make clear, and offer some defense, of the
explanatory role that the weak natural law thesis has in natural law jurisprudence
(section 4). I will conclude by drawing out some implications—predictably, anti-
positivist implications—for theories of law (section 5).
1. What the weak natural law thesis claims

Assertions represent the world as it is. Frogs have four legs. Clocks tell the time. Each of
these claims is, if construed as a universal generalization, false, and a fortiori if
construed as a necessary universal generalization. Some assertions are false, some
* I presented drafts of this chapter at the McMaster Conference on the Nature of Law and at a
meeting of the Australian Legal Philosophy Association. I am grateful to those who attended these talks
and offered criticisms and comments on the argument. I am particularly indebted to Stefan Sciaraffa,
whose careful written comments helped me to avoid some errors.
1 Murphy (2006: 25).
4 Mark C. Murphy
frogs have been de-legged, and some clocks are broken. Nor are these claims about
the preponderance of true assertions, four-legged frogs, and properly functioning
clocks.
Rather, each of these claims is about a kind (assertion, frog, clock), and the claim
holds that there are certain non-defectiveness conditions that belong to the kind.2
Because the claim concerns non-defectiveness conditions, it of course does not
follow either that all or most instances of that kind—or, indeed, any of that kind—
will exhibit those features. So it is no objection to the “frogs are four-legged” claim
that there are some three-legged frogs, and it is no objection to the “clocks tell the
time” claim that most clocks do not tell the correct time. But while it need not be
true that every instance of the kind exhibits the features that make them non-
defective instances, it is true that every member of the kind will have those features
as non-defectiveness conditions.3 While it does not follow from being a frog that
one must be four-legged, it does follow from being a frog that to be non-defective it
must be four-legged, or, in other words, that necessarily every frog is either defective
or four-legged.
I will not try to offer a very precise characterization of the notion of defectiveness.
But it is important for the discussion here that I make explicit the relation of the
notion of defect to other notions: in particular, the notions of kind, and ought, and
good. With respect to the notion of kind: the criteria for defectiveness are always in
relation to a kind; as it is sometimes said, these criteria are “internal” to the kind in
question. “Defective” is an attributive rather than a predicative adjective; to call
something a defective X is not to call it defective and an X, for the criteria for
defectiveness are fixed by the kind in question.4 Indeed, these criteria are internal to
the kind, in the sense that they at least partially define the kind. To be an assertion
is, in part, to be something that is defective if false.
Connected to this notion of defectiveness is a distinctive sort of “ought,” the
ought of kind-membership. It is true that a frog ought to have four legs. But when
one makes this claim one is not making some practical assessment of the world,
holding that it is better from the point of view of the universe (or, indeed, from
anybody’s practical perspective) for frogs to be four-legged; and it is also true that
when one makes that claim one is not affirming that anyone ought to act for the
sake of seeing to it that frogs be four-legged. That the ought of kind-membership is
not one of these practical oughts is clear from the fact that one could try to build a
machine that realizes outcomes that are in every way worse from the point of view
of the universe, or a machine the proper operations of which no one has any reason
to favor. Of such a machine it is true that it is supposed to, it ought, to realize those
2 For a discussion of the logical features of such claims, see Thompson (2008: 63–82).
3 This isn’t quite right. There can be deviant instances of the kind—decorative chairs, etc.—the
non-defectiveness conditions of which are not nondefectiveness conditions of the kind. We might call
these, as Mark Lance and Maggie Little call them, “riffs” on the kind (see their 2006). This point raises
complications for the statement of the natural law theses with respect to law, complications that
I discuss in Murphy (2011). Without rehearsing the discussion in that paper, I simply note that I think
that the argument in this chapter does not turn on any of these complications.
4 See Geach (1956).
The Role of the Weak Natural Law Thesis 5
bad outcomes; that there is no conceptual failure here indicates that this “ought” is
not an ought of practical rationality.
Also connected to the notion of defect is the notion of good, again, read
attributively, not predicatively; what makes something a good X is, at least in
part, that it is a non-defective X. It would be confusion here to think that by
bringing in this notion of “good” we are moving to the normative in a strong sense
of involving reasons for action, obligations, or any sort of practical considerations.
If I fashion a Pointless Doomsday Device that destroys all living things, but only
when doing so would have no point, it is a bad Pointless Doomsday Device if it fails
to generate the right sort of destruction in those circumstances. This badness is
obviously not any sort of moral badness, or indeed any badness from a relevant
practical perspective, for from every reasonable practical point of view it is better if
the machine works badly than if it works well.
I hope that this is enough on the notion of defect. For one to call something
defective requires one to make reference to some kind to which it belongs; it is to
say that it lacks something that things of that kind ought to have, and so it is in
some way not a good instance of that kind.
Now, characteristic natural law theses are of the following form: Law exhibits N,
where N is some normative feature. Instances of this form include Law is a legitimate
practical authority; or law is just; or law is not too unjust. The interpretation of such
theses that positivists know and love take these to be or to entail necessary universal
generalizations: necessarily, if x is a law, then x is legitimately authoritative, or just, or
not too unjust. We can call these strong natural law theses. But it is plausible, I have
argued elsewhere, that the sort of thesis that is central to the natural law tradition of
theorizing about law’s nature is not these characteristic natural law theses interpreted
strongly but rather interpreted weakly. On this weaker interpretation, that law
exhibits some normative feature N is understood as the claim that, necessarily,
anything that does not exhibit that normative feature is either not law or is defective
as law. To see the difference between the strong and weak readings it does not
particularly matter what the relevant normative feature is. But as I will later make use
of a particular weak natural law thesis, I may as well introduce it now. My own view is
that the weak natural law thesis central to the tradition of natural law theorizing, and
which is also, by the way, true, is that Law is backed by decisive reasons for compliance,
which I will abbreviate as Law is a rational standard.5
The strong natural law thesis is, then, a claim of this sort: Law is a rational
standard, understood as the thesis that anything that is not a rational standard
cannot be law. The weak natural law thesis is, then, a claim of this sort: Law is a
rational standard, understood as the thesis that anything that is not a rational
standard is either not law or is defective law. The weak thesis, unlike the strong
thesis, does not entail that every law must exhibit the feature being a rational
standard, much less that being a rational standard is part of what it is to be a law.
What it entails is that part of what it is to be good law is to be a rational standard,
5 See Murphy (2006: 8–20).

6 Mark C. Murphy
and that law ought to be a rational standard, in the sense of “ought” in which
assertions ought to be true and frogs ought to be four-legged and my Pointless
Doomsday Device ought to destroy all life. For each law, that law has being a
rational standard among its non-defectiveness conditions—nothing is a law unless
it is defective if not a rational standard.
2. Some bad objections to the weak natural law thesis
The initial objections that I want to discuss do not bear on any of the arguments
that have been offered for the weak natural law thesis, so I will put both the
exposition and further defense of those arguments to the side.6 Rather, these
objections concern whether the weak natural law thesis is itself a distinctive thesis
and whether it is a thesis that those who are interested in giving an account of the
nature of law ought to take an interest in.
Brian Bix, for example, wonders whether the weak natural law thesis is a
distinctive thesis; in particular, he wonders whether the weak natural law thesis
really is distinct from the strong natural law thesis on one side and what we may call
“the moral natural law thesis” on the other. We have already described the strong
natural law thesis. Call the “moral natural law thesis” the thesis that there are some
normative features that law must have if it is to count as a rational standard. Now,
the moral thesis as I have just stated it is excruciatingly uninteresting. Since being a
rational standard is a normative feature, it is pretty obvious that the moral thesis is
true. One could make it somewhat more interesting by specifying the relevant
normative feature—say, by declaring that law must be just if it is to count as a
rational standard. The worry that Bix wants to raise is independent of how
interesting the moral thesis is as a moral thesis, so I put that issue to the side.
What Bix is worried about is whether we can make sense of the claim that law
that is not a rational standard is defective as law is a distinctive thesis—perhaps the
only ways to understand it are either as the strong natural law thesis (that being a
rational standard is a necessary feature for each law to exhibit) or as the moral thesis
(that exhibiting some normative feature is necessary for any law to be a rational
standard). His specific way of raising the question is by focusing on the very idea of
legal defect: “Is it either sensible or tenable to have an idea of ‘defective’ law that is
not reducible to ‘legally valid but immoral’ or ‘legally invalid’?”7
The answer is simply Yes: it is both sensible and tenable. There are multiple
reasons for this, but we can begin with the plainest: since there is nothing in the
very idea of legal defect that requires that legal defect arise from any sort of
immorality, it is plain that we cannot reduce the notion of “legally defective” to
“legally valid but immoral”; and since a law’s being defective as law presupposes
that law’s existence, it is plain that we cannot reduce the notion of “legally
defective” to “legally invalid.”
6 I offer such defenses in Murphy (2006: 25–60), and in Murphy (2011).

7 Bix (2006: 146).
If the sketch of the notion of defectiveness that I offered above is on the right
track, it is plain why one cannot do the reduction toward which Bix gestures. To be
a legal defect is to be a failure to exhibit some feature that an instance of law ought
to have in virtue of being a member of the kind law. Nothing in the notion of “legally
valid but immoral” captures the idea that there are some features that laws ought to
have in virtue of being laws, and that is true even if being morally good is one of those
features that law ought (in the sense of the ought of kind-membership) to have.
The concept of legal defect is clearly not the concept of legal invalidity or
the concept of immoral (etc.) legality. It is open to Bix and others to deny that the
concept of legal defect has application, but it is fruitless to insist that the concept must
be reducible. After all, we have the concept of defect. We know how to use it
with respect to frogs, lawnmowers, assertions, rules of games, and so forth. What
would be mysterious is not the concept of legal defect but that concept’s being
reducible in the way that Bix suggests that it might be. Indeed, given the prevalence
of the application of the notion of defect with respect to artifacts—and laws are of
course artifacts—it would also be passing strange if the notion of legal defect lacked
application.
Let me turn now to a different objection. This objection does not, or need not,
call into question either the concept of legal defectiveness or its application in the
actual world. This objection is not against the weak natural law thesis as such but
against that thesis as brought forward as an important contribution to the tasks of
analytical jurisprudence. Leiter, for example, has recently claimed that natural law
theorists like Finnis and me, when proposing conditions for non-defective law, are
engaged in a “transparent change of the subject.”8 Dickson likewise claims that by
providing an account of the normative conditions that must be met for something
to be law in the focal sense—that is, non-defective—Finnis’s account is “problem-
atic for the reason that it does not seem to take seriously the enterprise of
identifying what law’s essential properties are.”9 Since the task of identifying
law’s essential properties is the central task of analytical jurisprudence, it seems
right to treat Dickson as also holding that Finnis—and other defenders of the weak
natural law thesis as central to jurisprudence—is guilty of changing the subject.
But this charge of changing the subject is groundless. What is the subject that is
allegedly being changed when the weak natural law thesis is put forward? The
subject is the existence conditions for law—as Dickson puts it, the identification of
law’s essential properties. The argument is that since the weak natural law thesis
gives the non-defectiveness conditions of law—and some laws may fail to exhibit
those non-defectiveness conditions—one putting forward the weak natural law
thesis cannot be putting it forward as a contribution to the task of giving the
essential properties of law. But this is a mistake. For, as I noted above, when one
gives the non-defectiveness conditions for a kind, one is also thereby giving some of
the existence conditions for the kind: if some feature is a non-defectiveness condi-
tion for some kind, then it is an existence condition for members of that kind that,
8 Leiter (2011: 677). 9 Dickson (2001: 77).

8 Mark C. Murphy
unless defective, they exhibit that feature. So the weak natural law thesis entails
that, necessarily, each and every instance of law instantiates this property: being a
rational standard or defective.
Now, I suppose that one might respond that this is sheer sophistry. If I were to
say that because it is one of the efficiency conditions of law that law maximize
wealth, it follows that an existence condition of law is that it be such that if it is
efficient, then it maximizes wealth, and conclude from that that one who makes a
claim about what it is for law to be efficient is not thereby changing the subject,
I might rightly be charged with sophistry. But there is a relevant difference, a
difference that comes from the fact that we’re talking about legal defect. To be a
defect is defined specifically, in terms of the kind in question; what counts as
defectiveness is, as they say, internal to the kind. So I am not importing some
alien criterion, conditionalizing, and putting that forward as an important existence
condition, part of the nature of law.
Another way to make this claim—that in giving non-defectiveness conditions for
law we are also giving existence conditions—plausible is by noting that it is
extremely plausible in the case of other kinds for which we give non-defectiveness
conditions. Surely part of what it is to be a paper clip is that if it cannot do a certain
job, then it is defective; and surely part of what it is to be an assertion is that if it is
false, or unjustified, then it is defective. We are not changing the subject from paper
clips to ideal paper clips when we say that paper clips hold pieces of paper together,
nor are we changing the subject from assertions to ideal assertions when we say that
assertions depict the world as it is. Part of what it is to successfully produce a paper
clip is to produce something that is, if unable to hold papers together, defective, and
if one devotes oneself to the task of explaining what makes it true that one has
successfully produced a paper clip one must explain what it is about one’s activity
such that one has made it true that this object has this non-defectiveness condition.
Part of what it is to successfully produce an assertion is to produce something that
is, if untrue, defective, and if one devotes oneself to the task of explaining what
makes it true that one has successfully produced an assertion one must explain what
is it about one’s activity such that one has made it true that this speech act has this
non-defectiveness condition.
To affirm the weak natural law thesis is therefore not to change the subject. But
Shapiro has recently suggested that it is a mistake to refer to it as the key thesis of
natural law jurisprudence. Here is the relevant stretch of argument:
My reason for rejecting this characterization is threefold. First . . . the weak reading of the
natural law thesis is consistent with the central tenet of legal positivism. A positivist can
coherently maintain that the law is ultimately grounded in social facts alone but that immoral
laws are defective as law. . . . Second, the weak reading of the natural law thesis cannot explain
why so many natural lawyers have claimed that unjust rules are not laws. . . . Third, the weak
reading of the natural law view, while not uninteresting, is not so interesting that the dispute
over it should constitute the major debate in analytical jurisprudence.10
10 Shapiro (2011: 408–9).

On the first point. One would think that whether a view is defining of a position is
to be understood by looking to the tradition of thinkers who self-define as
advocates of that position. So one would think that whether a view is defining of
the natural law position is not to be decided by asking what positivists need in order
to set up an opposition between them and some other theory, but rather by asking
what natural law theorists themselves have held—by, say, asking what seems to be
central and organizing of their position on law. After all, this is how we have gotten
to the contemporary characterization of positivism. When Gardner characterizes
certain propositions as “myths” about what legal positivism is,11 he does not begin
with an assumed, well-articulated account of what natural law jurisprudence is, and
just define positivism as the contradictory of that. Rather, he looks to what the
central, organizing commitments of paradigmatic positivists are, and he is not
hesitant to do some trimming here and there in order to clean things up.
On the second point. I think that Shapiro’s argument here bites back. I can—
and, later in this chapter, will—use the weak reading of the natural law thesis to
explain why so many natural law theorists have endorsed a strong natural law thesis;
as I will show, it is in part because natural law theorists endorse the weak natural law
thesis that some of them take a further step and endorse a strong natural law thesis
as well. So Shapiro is wrong when he says that one who characterizes the natural law
view in terms of the weak natural law thesis cannot explain why so many have
affirmed the strong formulation. But there’s more: it is in part because natural law
theorists endorse the weak natural law thesis that some of them have declined to
take that further step to the strong natural law thesis. So by identifying the key
thesis of natural law jurisprudence as the weak thesis, I can explain why so many
self-identified natural law theorists have affirmed the strong view; by identifying the
key thesis of natural law jurisprudence as the strong thesis, Shapiro cannot explain
why some self-identified natural law theorists have denied it.
On the third point. I am not sure why there is any need to set up one debate in
jurisprudence as the major debate, as if the major question (assuming that there is
one major question) must be answerable as True or False. It is also strange that
Shapiro identifies the strong natural law thesis as the False view contrasted with his
True view while defining the issue as “whether law ultimately rests on moral
facts”—it could be true that law ultimately rests on moral facts even if the strong
natural law thesis is false.12 So neither the strong nor the weak reading of the
natural law thesis may give Shapiro what he wants, that is, an unambiguous
opponent for the sort of positivism that he defends.
3. What a theory of law should accomplish

These criticisms of the weak natural law thesis as a central natural law contribution
to analytical jurisprudence are mistaken. But I think that what needs to be made
11 See Gardner (2001). 12 I discuss this point in more detail in section 5.

10 Mark C. Murphy
explicit is the role that the weak natural law thesis plays in natural law jurispru-
dence, and how that role places some pressure on certain sorts of theories of law.
Some quick remarks about what I take theories of law to be and what counts as
success in a theory of law. I follow Raz in holding that a theory of law is supposed to
consist of necessary truths about law that explain it.13 And I follow Dickson in
holding that part of the character of a theory of law as explanatory is its capturing
our ordinary legal experience.14
And I also follow, with a caveat, Shapiro’s distinction between the “Identity
Question” and “Implication Question” with respect to law.15 Shapiro is of course
right to distinguish between the task of characterizing the property of legality, of
being law, with the task of drawing out the implications of characterizing being law
in that way. And it is important to take both of these questions to be relevant to a
theory of law if one wants to follow Raz and Dickson in their way of describing its
tasks, for our ordinary legal experience is explained both by what being law is and
what follows from being law. Here, though, is the caveat. I do not think that we
should be happy distinguishing truths about law, even necessary truths, simply into
answers to the Identity or Implication Questions. That is because the label
“Implication Question” elides some very important distinctions. The distinction
that matters most for the purposes of the argument of this chapter is that between
implications about how law is realized from implications that are not about how law
is realized. How some property is or even must be realized is a distinct question
from that of the identity of that property, and it is easy to confuse the identity
conditions for some property with its realization conditions. It is a mistake to do so.
But it also seems to me that realization conditions are so important that an adequate
theory of law should not treat claims about those necessary features of law that are
realization conditions simply as some implications of law among others. Rather,
they should be broken out and acknowledged as such.16
13 Raz (2004: 324).

14 Dickson (2001: 23–4). This does not make legal theory inherently conservative with respect to
our ordinary reflective judgments about law—while this constraint could be satisfied by showing why,
given one’s preferred theory, ordinary legal actors’ judgments about law are true, it could, alternatively,
be satisfied by explaining why, though ordinary reflective judgments are systematically confused, those
judgments tend to be held. But one’s legal theory must handle these data by explaining why the data
are as they are.
15 “In asking the Implication Question, we are not concerned with why something counts as an
instance of law [the Identity Question] but rather with what necessarily follows or does not follow from
that fact.” Shapiro (2011: 12).
16 Suppose that in some world Triheighters come in three possible heights: 3 meters, 2 meters, and
1 meter. The property being a Triheighter of unsurpassed height is the property of being a Triheighter of
such a height that no Triheighter’s height is greater than it. It is not to be identified with being 3 meters in
height, though in the actual world one is a Triheighter of unsurpassed height if and only if one is a
3-meter Triheighter. The most obvious reason is that perhaps in some other world Triheighters come
in other heights. But it would be weird to treat being 3 meters in height as being just one implication of
being a Triheighter of unsurpassed height; it is by being a 3-meter Triheighter that one is a Triheighter
of unsurpassed height. And this would not change a bit, I say, if it were a necessary truth about
Triheighters that they come in these three heights. My views here require the rejection of sparsism
about properties. I hereby reject.
That is sufficient preliminary discussion of what a theory of law should be up to.

It should consist of necessary truths about law that explain the law, including folks’
characteristic responses to it, and it should properly characterize these truths in
terms of whether they are or are part of informative property identifications about
what being law is, or provide realization conditions for being law, or are otherwise
necessary (in some sense of “necessary”17) implications of being law.
4. Hypothetical necessity and the weak natural law thesis
Natural law theorists hold that some weak natural law thesis is a necessary truth
about law. But there is another salient feature of natural law theories, and that is
that some weak natural law thesis plays a privileged role in specifying the other
necessary truths that belong in a theory of law. The explanation of other truths
about law goes through the weak natural law thesis by way of a distinctive form of
explanation: hypothetical necessity.18
What is the structure of explanations that appeal to hypothetical necessity? Here
are some of Aristotle’s remarks, which are the locus classicus of the idea:
No hand of bronze or wood or constituted in any but the appropriate way can possibly be a
hand in more than name. For like a physician in a painting, or like a flute in a sculpture, in
spite of its name it will be unable to do the office which that name implies.19
Again:
If a piece of wood is to be split with an axe, the axe must of necessity be hard; and, if hard,
must of necessity be made of bronze or iron. Now exactly in the same way the body, which
like the axe is an instrument—for both the body as a whole and its several parts individually
have definite operations for which they are made—just in the same way, I say, the body, if it
is to do its work, must of necessity be of such and such a character, and made of such and
such materials.20
There are multiple ways to construe the explanatory moves being made here, and
I am not interested in trying to adjudicate among them. I want to consider the
merits of one admissible reading of these arguments, which I will later claim to be
the canonical form of explanation within natural law jurisprudence generally. The
form of argument is
17 There are lots of sorts of necessity that could be invoked here: metaphysical, physical, practical,
for-all-practical-purposes, etc. The important point is that legal theory is not concerned with what
holds merely accidentally of law; and if it claims that some features are necessary for law, it should make
clear the sort of necessity at issue.
18 This should probably be qualified, for there are other features of law the explanation of which
seems not to go through the weak natural law thesis. Perhaps the features of law that are concerned with
its scope—that law is supposed to be comprehensive, or for a complete community, or the like. The
range over which law purports to be a rational standard is not explained by the law’s purporting to be a
rational standard. I am far from sure about how best to think about this.
19 Aristotle, Parts of Animals, 640b35–641a2.
20 Aristotle, Parts of Animals, 642a10–14.
12 Mark C. Murphy
(1) It is the office of Xs to ç
(2) Only things that are Y are constitutionally able to ç
(3) Nothing is an X unless it is Y.
So Aristotle takes it that the hand is primarily characterized by its office, which is
(I suppose) grasping; and Aristotle takes it that a hand-shaped object made of bronze
or wood is, by its nature, not capable of doing that job; and so it is a hand in no more
than name, no more a hand than a representation of a hand is. And Aristotle takes
it that an axe is primarily characterized by its office, which is (I suppose) chopping;
and Aristotle takes it that an axe-shaped object made of (say) sludgy cottage cheese
is, by its nature, not capable of doing that job; and so it is an axe in no more than
name.
Now there are cases that make this form of explanation look very plausible. Take,
for example, artifacts, and the differences that we recognize between a not-fully-
successful attempt to create an instance of an artifactual kind and a botched attempt
that results only in scrap.21 If things go awry in my paper clip factory and it
produces little globules of liquid aluminum rather than nicely coiled pieces of
aluminum wire, it is clear that much as I have tried to make paper clips, what I have
here are no paper clips at all. This is true regardless of whether I think that I tried to
make paper clips, whether I tried to make something with the function of paper
clips, and even whether I believe that what I have produced are paper clips.
Or take illocutionary acts. Some theorists of illocutionary acts have distinguished
between the success and non-defectiveness conditions for illocutionary acts, where
a success condition is a condition for performing an act of that type at all and a non-
defectiveness condition is a condition for performing a perfect instance of that
kind.22 Consider assertions, for example. There are various conditions that one has
to meet in order to perform an assertion at all, at least under standard conditions:
one has to be speaking a language, to get the words out, be able to be recognized as
in a context in which speech acts can be of the asserting kind, and so forth. These
are to be distinguished from non-defectiveness conditions—for example, truth and
justification—that may fail to be met in a given case of genuine assertion. The
explanation for the success conditions being in place for there to be an assertion at
all, regardless of what the would-be asserter thinks—one can be massively confused
about whether one has made any assertions—is that such failures would count as
not even being the right kind of thing to be able to do the job of representing to
someone the world as it is, as being doomed to egregious inability to perform the
office of assertions.
One might also think that we have some further evidence for the plausibility of
this notion from the fact that we are inclined to say that machines the function of
which is some impossible task are not merely invariably defective but necessarily
non-existent. You know what I mean when I say that some pathetic would-be but
never-was Orville Wright produced several prototypes but never succeeded in
making an airplane—nothing he or she ever produced was even of the right sort
21 Hilpinen (1993: 155). 22 See e.g. Searle and Vanderveken (1985: 51–9).
to do that job.23 You also know what I mean when I say that there has never been
and never will be a perpetual motion machine. I am saying not merely that all such
will be defective, but rather that there is no such thing as “the right kind of thing”
to be a perpetual motion machine.
It is not just that there are plausible judgments about cases that seem well
captured by the appeal to hypothetical necessity. One can offer a principled account
of why that argument form is successful in terms of an explicit account of what it is
to have an “office”—think of it as a function, a characteristic activity that a type of
thing is supposed to carry out. But to have a function is, in part, to tend toward
some equilibrium state in some set of privileged conditions. (That the function of a
vacuum cleaner is to remove dirt in a particular way is, in part, for the conditions in
which the vacuum cleaner is to operate (plugged in, etc.) to tend toward dirtless-
ness.) But if something is constitutionally unable to perform the office of some
kind, there could not be such a tendency in any set of privileged conditions. So
there is a plausible reason to endorse the requirement on constitutional ability in
determining what things can be instances of such kinds.
Now, there is no doubt but that the notion of “constitutionally able” is the weak
link here and a source of real discomfort for the defender of hypothetical necessity
as a successful argument form. It is clear why the Aristotelian should want some
such notion: we should want to distinguish between local failures of a paper clip or
an assertion to do its job, making the object count as defective or perhaps operating
in an unfavorable environment, and a deeper sort of inability that makes an object
not even a candidate for the role of paper clip or assertion.24 While I don’t think
that it is incumbent on the user of hypothetical necessity arguments to give a tidy,
vagueness-free account of constitutional inability, still we should want some clearer
sense of what “constitutionally able” amounts to.
I take it that, first, what something is constitutionally able to do is not a product
of adventitious circumstances: if a paper clip cannot do its job because it has been
placed in a vat of liquid nitrogen and thus is so brittle that it will crumble when
papers are pushed into its coil, that does not count against its being a paper clip.
“Constitutional ability” is about, unsurprisingly, constitution—the stuff that some-
thing is made of, and how it is configured.
With respect to the former—the material—the focus of the examples from
Aristotle that I offered, we can say that there is some stuff that is just unsuitable
to do certain jobs; perhaps we might say, strongly, that since replacing that matter
23 Here I disagree with Ehrenberg; see his (2009: 94). It makes perfect sense to say that those
inventors whose designs were total non-starters never built an airplane.

24 I reject Ehrenberg’s characterization of Aristotelian function ascription on which to hold that
anything in its current condition is unable to do a certain job entails its not being an instance of the
functional kind in question (Ehrenberg (2009: 95)). What my rejection presupposes, though, is a
distinction between being merely unable to do something and being constitutionally incapable of doing
that thing. Since the view that Ehrenberg ascribes to Aristotle is not a very plausible view—could
Aristotle, being a non-moron, have possibly thought that there is nothing that is an X but a defective
X?—I take it that I am not shouldering a burden that Aristotelians don’t have already, that is, the task
of making sense of the difference between a mere momentary contingent inability and a deep consti-
tutional incapacity to perform some office.
14 Mark C. Murphy
with more suitable matter would involve not improving the object’s ability to carry
out the office in question but rather substituting a new object for the old one, that’s
what makes it true that this object (a cottage cheese “axe”) could never perform the
office in question (axe-ing).
With respect to the latter—the configuration—things are more difficult. I think
that there are still clear cases here—where the material is fine but the configuration
just egregiously wrong—like in the case of aluminum configured as a liquid rather
than as a solid. And I take it that we want to say, as in the case of the would-be
airplane inventor who never did make an airplane, that there are some design flaws
so egregious that we would say that there is not even an instance of the kind in
question present, even though the materials used may be impeccable. (Perhaps a
test here is whether we would say that the design can be tweaked or whether the
idea itself needs to be scrapped.) But I think some vagueness on the sort of
configuration that counts as constitutional inability is inescapable, or at least
I am unable to escape it.
So much for the character of appeals to hypothetical necessity in establishing
necessary truths about what things must be like in order to be instances of particular
kinds. Do truths thus established count as telling us more about the identity of these
kinds—about what it is to be a paper clip or an axe, what sort of property analysis
we should give for being a paper clip or being an axe? My inclination is not to think
of truths of the sort “Axes cannot be made of cottage cheese, but must be made of
something harder than that,” “Paper clips cannot be liquid, but must be solid,” etc.,
as answering what we might think of as the Identity Question regarding axes and
paper clips. To be an axe or a paper clip is to be characterized functionally; these
truths about what an axe or a paper clip must be like in order to be constitutionally
able to carry out its function tell us what something must be like in order to realize
the features that are essential to being an axe or a paper clip. These, then, are
realization conditions for being a paper clip or being an axe, not parts of a property
analysis of being a paper clip or being an axe.
The pattern of explanation that helps us to see what something must be like (or
must not be like) to be a paper clip or assertion is the key pattern of explanation in
natural law jurisprudence. Unlike the office of the paper clip, the office of law is to
be characterized normatively, as providing a rational standard for conduct. But it is
open for legal theorists to argue from this thesis about law’s office to necessary
conditions on being law that are either normative or non-normative. These other
conditions that make for law are conditions that make for law because their
presence is necessary for law to perform its office as a rational standard for conduct.
So, some natural law theorists affirm some formulation of a strong natural law
thesis. But their arguments for some version of the strong natural law thesis run
from a weak natural law thesis, via the premise that there are failures to exhibit
certain normative features that make would-be legal norms constitutionally incap-
able of being rational standards for conduct, to the conclusion that any such would-
be laws are laws “in name only.” This is Aquinas’s pattern of argument, at least on
the standard reading of Aquinas. Aquinas allows that it is the office of law to be a
rational standard for conduct. But on Aquinas’s view, nothing that is incompatible
with the natural law could possibly be a rational standard for conduct;25 to be
incompatible with the natural law is to be the wrong sort of thing to be a law, just
like a puddle of aluminum is the wrong sort of thing to be a paper clip.26 This is
clearly Michael Moore’s account,27 and it seems to be the most charitable way to
read Fuller’s claim that his “eight ways to fail to make law” specify the “internal
morality” of law.28 If you take the eight ways to fail to make law to count as failures
because would-be rules that egregiously fail in these ways cannot be rational
standards for those under the law, then what Fuller is giving is hypothetical
necessity argument from a weak natural law thesis to a strong natural law thesis.29
But it is not only that some natural law theorists take the argument from the
weak natural law thesis and the explanation pattern of hypothetical necessity to give
necessary normative conditions for the realization of law. All natural law theorists,
strong or weak, take the weak natural law thesis to be central to hypothetical
necessity arguments for the non-normative conditions on law. Their view is not that
there are some non-normative conditions on law explained in one way, and some
normative conditions on law explained in some other way, which we then simply
conjoin for a complete account. Rather, the weak natural law thesis explains
whatever normative conditions and whatever non-normative conditions that
there are on law.30 So the natural law theorist appeals to the weak natural law
thesis also to explain why law must be grounded in some sort of social practice,
because law cannot serve as a rational standard for conduct unless it is realized at
least in part through social facts obtaining. There are various explanations from the
weak natural law thesis for this: some appeal to the weakness of our ability to grasp
the requirements of practical rationality; some appeal to the weakness of our ability
to grasp the non-normative facts to which these practical requirements are to be
applied, or the weakness of our ability to carry out those applications well; some
25 I think Aquinas is committed to rejecting the view that being a standard that a person could
rationally decide to follow entails being a rational standard for conduct. Rationality, as Parfit would later
put it, is not “inherited” this way (Parfit 1984: 40). So the fact that (for example) blasphemy is simply
ruled out by practical rationality shows that a would-be law that requires blasphemy is just the wrong
sort of thing to count as a law. That one might be in some position such that the rational thing to do is
to allow another person to guide one’s conduct and that person could direct one to blaspheme would
not give any basis for thinking that “one should blaspheme” could count as a rational standard for
conduct.
26 Aquinas, Summa Theologiae IaIIae 95, 2; see Murphy (2004: 15–18).
27 Moore (2001).
28 Fuller (1964: 39). In my view, Fuller’s formal conditions on law have as their unifying rationale
that those conditions must be present in order for legal norms to function as reasons for action for the
parties on whom they are imposed; see 105, 110.
29 One might offer the following objection: But the eight ways are ways to fail to make law not
because they preclude rules from being authoritative, but because they preclude these from even
functioning as rules at all, whether genuinely authoritative or not. To which I say: This is irrelevant.
What is calling the shots here, ultimately deciding what can be law and what cannot, is the truth about
law specified by the weak natural law thesis. If Fuller’s arguments work by appealing only to a formal
aspect that would have to be realized by any would-be norm that satisfies that thesis, that is an
interesting point but does not make any difference to the claim that Fuller’s view is part of the
mainstream natural law tradition that I am identifying and in essentials endorsing.
30 Subject to the qualification in n. 18.
16 Mark C. Murphy
appeal to the normative openness of the principles of practical rationality them-
selves, and the need for fill in that openness through authoritative intervention.
So, for example: Aquinas holds that law is a rational standard for conduct, and
also holds that the law provides a comprehensive ordering for the life of a political
community. But what is necessary to provide a comprehensive ordering for the life
of a political community goes beyond what is provided by natural standards of
practical rationality, and that is why Aquinas says that law consists both of
standards of conduct grounded in the natural law by deduction and of standards
of conduct grounded in the natural law by determination.31 If law is to be
grounded in the natural law by way of determination yet be a rational standard
for conduct, there must be some way of specifying the existence and content of the
law that goes beyond what is given by normative facts.32 And so Aquinas allows that
law consists of what is laid down in the relevant sense by those who in fact have care
of the community. How this takes place within a given community is a matter of
custom, of what is accepted in that community, and is contingent.33 But that law
does not exist without the holding of certain social facts is not contingent, and the
broad type of social facts that make the difference—in Aquinas’s case, de facto
authorities giving commands, prohibitions, etc.34—is not contingent either.
The basic point is that the weak natural law thesis is the basis for affirming
whatever ought to be said about both of these sorts of necessary conditions on law,
and the argument to these conditions is, on the natural law view, an argument from
hypothetical necessity.
5. Anti-positivist implications of the weak natural

law thesis and hypothetical necessity
While I have offered some defense of hypothetical necessity arguments as a way of
giving necessary conditions on instances of kinds that have an “office,” I have not
defended any particular arguments that employ hypothetical necessity to move
from the weak natural law thesis to truths about what law must be like. One might
doubt whether any further interesting theses of legal theory could be defended
without offering some such particular hypothetical necessity arguments. But while
of course one should be very interested in pursuing the details of such arguments,
we can say something of jurisprudential interest even without having those details
at hand.
I begin with something that would not follow from the endorsement of the weak
natural law thesis and hypothetical necessity. It would not follow that one could not
give necessary and sufficient conditions for law that are entirely non-normative, and
it would not even follow that one could not give necessary and sufficient realization
31 Aquinas, Summa Theologiae IaIIae 95, 2.

33 See Murphy (1997).
conditions for law that are non-normative. Shapiro’s account of law in terms of
plans of action, Hart’s account in terms of the rule of recognition, Austin’s account
in terms of a weird account of commands and a weird account of sovereignty—it is
compatible with affirming both the weak natural law thesis and the appeal to
hypothetical necessity to affirm that any of these gives a correct account of what is
necessary, and what is sufficient, for legality to be realized, at least in our world.35
Now, to what does follow. First, it is very plausible that if it is essential to a kind
of thing to have a certain office, then having that office is part of the answer to the
Identity Question for that kind. So it is part of what it is to be law to be such that it
is, if not a rational standard for conduct, defective. What follows from this is that no
attempt to answer the Identity Question without appealing to a property that
includes the property ascribed to law by a weak natural law thesis can succeed.36
Even if such an account could, as conceded above, provide a full non-normative
account of the realization conditions for law that does not include the weak natural
law thesis, an account of realization conditions is not an account of identity
conditions.
Second, the weak natural law thesis together with hypothetical necessity would
entail the falsity of the view that facts about law’s existence do not depend on moral
facts—or, more accurately and less parochially, facts about good reasons for action.
The natural law argument goes from what is needed to successfully carry out the
office of law to the necessary conditions for law’s being in place, and what is involved
in successfully carrying out the office of law is a matter of fact about right reason. So
any of the other putative identity conditions for law, or any realization conditions for
law, or indeed any further implications about law, that are derived from the weak
natural law thesis by way of hypothetical necessity depend on facts about right reason.
If some strong natural law theses are true, perhaps the facts upon which facts
about law’s existence depend are specific facts about what actions are ruled out by
practical rationality; or perhaps what is relevant is the more general fact that there are
some facts about what courses of action are more reasonable than others.37 So any
35 I confess to being deeply skeptical of all of these views as accounts of the necessary realization
conditions for law simpliciter. I think that folks have not adequately reflected on the fact that other
species of rational beings could have law, and the sorts of practical cognition and motivation exhibited
by those species could make something like a straight natural law view the true account of the
realization conditions for law for those species. This is not a cheap trick. If by “necessary” is meant
“metaphysically necessary,” then the metaphysical possibility of other sorts of rational being is of course
relevant. And because on my own view, as on Aquinas’s (Summa Theologiae Ia QQ. 50–64), this is not
just a possibility, but an actuality, there is further reason to reflect on whether law among rational
non-humans might exhibit importantly different realization conditions.
36 This conclusion is Shapiro’s as well. But unless I misunderstand him he offers different grounds
for the conclusion. He appeals to what is necessary to carve the cases correctly: we cannot correctly
distinguish the cases pretheoretically identified as instances of law from others without including the
weak natural law thesis as part of the answer to the Identity Question; Shapiro (2011: 214–17). I am
appealing here to a connection between an office’s being a necessary feature of a kind and its being at
least part of the identity of that kind. Having a certain function is, I say, either not necessary to a kind
or is part of the identity of that kind.
37 I take it that there are facts about right reason is a fact about right reason.
18 Mark C. Murphy
view that is formulated as the claim that law’s existence does not depend on facts
about good reasons for action is a false view.
Here is another way to put the point.38 I say that if you want to know whether a
theory of law really holds that facts about law’s existence do not depend on facts
about good reasons for action, then you should ask what the implications of that
theory would be were it conjoined with error theory about reasons for action.39 If
you think that facts about law’s existence do not depend on facts about good
reasons for action, then your view of what law exists should not be altered a whit by
the supposition that an utter and uncompromising nihilism is the truth about
reasons for action.
There is more than one way to resist here. One way is to hold that while a weak
natural law thesis is part of the answer to the Identity Question, the natural law
theorist is wrong to think that the weak natural law thesis explains the other
conditions for law via hypothetical necessity; rather, these are simply independent
conditions conjoined. This is a possible reading of Shapiro’s view in Legality,
though I don’t think that it is an admissible reading of Raz’s view in “Authority,
Law, and Morality,” where the explanatory role of something like the weak natural
law thesis is explicit with respect to the social facts that make for legality.40 I have
already said all that I aim to say here defending the claim that conditions of legality
can be explained by appeal to the weak natural law thesis, so I will just add one
thing: if one does not appeal to the weak natural law thesis to explain other
conditions of law, but one nevertheless includes it in one’s account of law along
with other conditions, it makes being law look much more gerrymandered and
artificial, not displaying the unity that we would expect even in a social kind.
Another way to resist is to say that while one should appeal to a weak natural law
thesis to explain other necessary features of law, one should not appeal to what must
be the case for law to successfully satisfy its claim to carry out its office of being a
rational standard—only to the fact of the claim itself. This move still appeals in
some way to the office of law to explain law’s other features. And it is a prima facie
plausible reading of Raz’s “Authority, Law, and Morality” argument, where he
seems to move directly from law’s necessary claim to authority to a partial account
of what would have to be the case to explain the possibility of that claim’s being
true;41 and so one might thus say that Raz need not appeal to facts about right
38 I was pretty proud of coming up with this way of putting things, until I found a similar
formulation in Greenberg (2004: 158); I must have gotten the idea from him.
39 Here is a fun way of seeing why this matters. Above I suggested that we should say not that
perpetual motion machines are possible but inevitably defective, but that there are no such machines,
for nothing could be constitutionally able to generate perpetual motion. If it were true that being a
rational standard were intrinsically impossible to realize—if the right view to take of being a rational
standard were that being a rational standard is like being a square circle, say—then I would say that there
could never have been law, not that all law is defective. The fact that one could have practices on their
face indistinguishable from our legal practices does not make trouble for this view, any more than the
imagined existence of a thriving practice of attempting to make perpetual motion machines, carried on
by folks who actually think that some such are successful, makes trouble for the idea that none of them
has actually produced one.
40 Raz (1994: 199–204). 41 Raz (1994: 202–4).
reason to get to his hard positivist conclusions. But I do not think that view is
sustainable. There must be reasons for action that can be mediated and upon which
would-be rules of action can depend, and there must be reasons to act on such
mediations of dependent reasons, for law’s claim to authority to be possibly true,
and it is the possibility of such conditions for successful mediation to which Raz
appeals in arguing for the sources thesis.42
Now, one might cry foul: even if Raz appeals to what is required for a successful
claim to authority to defend hard positivism, that appeal is only to certain aspects
of that success, aspects that are not substantive, and indeed are compatible with
the most horrible, stupid, or pointless attempts to exercise authority. But I do not see
why this is a relevant objection. If law’s claim is a claim to success in authoritative
imposition of norms, and the conditions of legality are what are necessary to get at
that success, it does not matter a whit that the conclusion that one is arguing for is
that law’s identity (or realization) conditions do not include facts about merits; the
point is that what makes those identity (or realization) conditions what they are is
facts about right reason. Facts about right reason thus determine what can be law.
Perhaps they do so by determining that there are substantive constraints on what
legal norms there could be, or perhaps they do so by determining that there could
not possibly be any such constraints. But either of these would be “subordinate
theorems” of a recognizably natural law account of the conditions of legality.43
Here is another way to try to resist the claim that the weak natural law thesis,
together with hypothetical necessity, makes trouble for positivism. As Gardner
writes, legal positivism may be understood not as a large-scale philosophical theory
about law,44 which is how I have construed natural law theory, but simply as a
single thesis about law: that the existence of law is a matter of sources, not merits.45
And one might say that only if it turns out that the best natural law theory yields the
thesis that law’s existence is a matter of some merits being realized in law will there
be any tension between legal positivism and the natural law view as I have
characterized it, even if in some sense law’s existence conditions depend on facts
about right reason.
It is true that it does not follow simply from the weak natural law thesis and the
employment of hypothetical necessity arguments that law must exhibit any merits
of the following sorts: being a rational standard, being just, being authoritative, and
so forth. As I have allowed, it depends on further, even more controversial
arguments to move from the weak natural law thesis to the strong natural law
thesis, and it is merits of this sort that strong natural law theorists have included as
existence conditions for law, thus placing themselves clearly in opposition to the
positivist thesis as Gardner characterizes it. But I don’t think that this leaves the
positivist in the clear. For consider the following property: being constitutionally
capable of being a rational standard. According to the view I have defended here, this
42 Thus writers like Goldsworthy and Kramer seem to me to be right in thinking that Raz’s view of
law’s claim to authority is a Trojan horse for natural law theory to get inside the positivist stronghold.
See Goldsworthy (1990: 449–86) and Kramer (1999: 79).
43 Finnis (1980: 351). 44 Gardner (2001: 223–4). 45 Gardner (2001: 199).
20 Mark C. Murphy
is an essential feature of law; something’s counting as law consists, in part, of its
being constitutionally capable of being a rational standard. And being constitution-
ally capable of being a rational standard certainly looks like a merit. If that property
counts as a merit, then the weak natural law thesis together with hypothetical
necessity entails the falsity of the positivist thesis.
Third, I think that treating the weak natural law thesis, or some nearby
functional thesis about the law, as the master explanatory principle with respect
to conditions for legality via hypothetical necessity should motivate legal theorists
to be more careful about what is included in law’s identity conditions and what are
merely realization conditions for law. If the weak natural law thesis has this master
role, then it may be that proposed identity conditions for law should be thought of
as, instead, realization conditions. Suppose that some such feature of law is wholly
explained by the weak natural law thesis—that it is just entailed by it. What
motivation would we have to treat it as part of the property being law? Not to
carve the cases correctly, anyway; one would get no more case-carving work done by
including these explained conditions as part of the property being law than one
would get from the weak natural law thesis alone. It seems, then, that the only
properties we would want to include in the property analysis of being law would be
those not wholly explained by the weak natural law thesis.
So, to take one application of this point: if there are conceptual arguments for
positivist conditions on law that proceed from a weak natural law thesis, that gives
reason to think that those positivist conditions are not even part of the identity of
law, but merely realization conditions for it. For example: if law must be identifi-
able apart from the reasons on which it depends, and that is established by a
conceptual argument from an appeal to law’s office,46 then we have reason to think
that being identifiable apart from the reasons on which it depends is not part of what it
is to be law, but rather a feature that something must have to realize legality.
So there are some interesting results from the weak natural law thesis’s being
conjoined with the in-principle soundness of hypothetical necessity explanations.
This conjunction threatens some standard formulations of positivist theses about
law. And it can demote certain positivist theses about law’s identity conditions to
realization conditions. It is pretty plain, then, that the natural law theorist’s appeal
to the weak natural law thesis does not change the subject. It is more accurate to say
that it defines the subject.
References
Aquinas, T. (1981). Summa Theologiae, trans. Fathers of the English Dominican Province.
Westminster, MD: Christian Classics. References given by part, question, and article
number.
Aristotle (2009). Parts of Animals, trans. W. Ogle. Whitefish: Kessinger.
46 Raz (1994: 202).

Bix, B. (2006). “Robert Alexy, Radbruch’s Formula, and the Nature of Legal Theory,”
Rechtstheorie, 37: 139–49.
Ehrenberg, K. (2009). “Defending the Possibility of a Neutral Functional Theory of Law,”
Oxford Journal of Legal Studies, 29: 91–113.
Finnis, J. (1980). Natural Law and Natural Rights. Oxford: Oxford University Press.
Fuller, L. (1964). The Morality of Law. New Haven: Yale University Press.
Gardner, J. (2001). “Legal Positivism: 5½ Myths,” American Journal of Jurisprudence, 46:
199–227.
Geach, P. (1956). “Good and Evil,” Analysis, 17: 33–44.
Goldsworthy, J. (1990). “The Self-Destruction of Legal Positivism,” Oxford Journal of Legal
Studies, 10: 449–86.
Hilpinen, R. (1993). “Authors and Artifacts,” Proceedings of the Aristotelian Society, 93:
155–78.
Kramer, M. (1999). In Defense of Legal Positivism: Law without Trimmings. Oxford: Oxford
University Press.
Lance, M. and Little, M. (2006). “Defending Moral Particularism,” in J. Dreier (ed.),
Contemporary Debates in Moral Theory. Malden, MA: Blackwell, 304–21.
Leiter, B. (2011). “The Demarcation Problem in Jurisprudence: A New Case for
Skepticism,” Oxford Journal of Legal Studies, 31: 663–77.
Moore, M. (2001). “Law as Justice,” Social Philosophy and Policy, 18: 115–45.
Murphy, M. C. (1997). “Consent, Custom, and the Common Good in Aquinas’s Theory
of Political Authority,” Review of Politics, 59: 323–50.
Murphy, M. C. (2004). “Natural Law Theory,” in M. Golding and W. Edmundson (eds.),
Blackwell Guide to the Philosophy of Law and Legal Theory. Malden: Blackwell, 15–28.
Murphy, M. C. (2006). Natural Law in Jurisprudence and Politics. Cambridge: Cambridge
University Press.
Murphy, M. C. (2011). “Defect and Deviance in Natural Law Jurisprudence,” in Matthias
Klatt (ed.), Institutionalized Reason: The Jurisprudence of Robert Alexy. Oxford: Oxford
University Press, 45–60.
Parfit, D. (1984). Reasons and Persons. Oxford: Oxford University Press.
Raz, J. (1994). “Authority, Law, and Morality,” in Ethics in the Public Domain. Oxford:
Oxford University Press, 194–221.
Raz, J. (2004). “Can There Be a Theory of Law?,” in M. Golding and W. Edmundson
(eds.), Blackwell Guide to the Philosophy of Law and Legal Theory. Malden: Blackwell,
324–42.
Searle, J. and Vanderveken, D. (1985). Foundations of Illocutionary Logic. Cambridge:
Cambridge University Press.
Shapiro, S. J. (2011). Legality. Cambridge, MA: Harvard University Press.
Thompson, M. (2008). Life and Action. Cambridge, MA: Harvard University Press.
2
In Defense of Hart
Matthew H. Kramer
In his important and engaging book Legality, Scott Shapiro (2011) seeks to provide
the motivation for the development of his own elaborate account of law by
undertaking a critique of H. L. A. Hart’s jurisprudential theory. Hart maintained
that every legal system is underlain by a Rule of Recognition through which the
officials of the system identify the norms that belong to the system as laws. Shapiro
argues that Hart’s remarks on the Rule of Recognition are confused and that his
model of law—though commendably more sophisticated than any model pro-
pounded by earlier legal positivists—is consequently untenable. Having thus
endeavored to establish that Hart’s exposition of the nature of legality is unsustain-
able, Shapiro contends that a new approach is vital for progress in the philosophy of
law. With his lengthy presentation of his own Planning Theory of Law, he aspires
to pioneer just such an approach.
Except for a very terse observation in the final main section, this article will not
directly assess the strengths and shortcomings of Shapiro’s piquant Planning
Theory. Instead, I will defend Hart against Shapiro’s charges and will thereby
undermine the motivation for the development of the Planning Theory. Admit-
tedly, while arguing that Hart’s jurisprudential theory is broadly sound and that
Shapiro’s alternative is therefore superfluous, this article will not attempt to uphold
every detail of Hart’s ruminations on law. Having criticized Hart on a number of
points elsewhere (Kramer (1999: 21–36), (2003: 312–13), and (2004: 249–94)),
I am scarcely inclined to suggest here that his writings are entirely beyond reproach.
Nonetheless, nearly all the objections to Hart’s work posed by Shapiro are inappo-
site, or so this article will aim to show.
In what follows, I will not take issue with everything in Shapiro’s discussions of
Hart that strikes me as doubtful. Nor will I take issue with any strands of his
book—such as his espousal of Exclusive Legal Positivism in preference to Inclusive
Legal Positivism1—that do not pertain directly to his animadversions on the
Hartian Rule of Recognition. This chapter will concentrate not only on Shapiro’s
book but also on a recent essay of his concerning the Rule of Recognition (2009).
1 For my principal defenses of Inclusive Legal Positivism against Shapiro and other Exclusive Legal
Positivists, see Kramer (2004: chs. 1–4).

In Defense of Hart 23
Both his book and his essay seriously underestimate the resourcefulness of Hart’s
account of law, and both of them misrepresent many aspects of Hart’s ideas.
1. A matter of belonging
Quite early in Legality’s chapter on Hart’s model of law, Shapiro attributes to Hart
the following conception of the Rule of Recognition:
It might be wondered, then, why Hart presents the rule of recognition as a test of legal
validity when in reality it is a duty-imposing norm. The answer, I believe, is that, according
to Hart, the law of a particular system consists of all the norms that legal participants of
that system are under a duty to apply in their official capacities. In other words, the rule
of recognition sets out the criteria of legal validity, and hence picks out the set of legal rules
for a particular legal system, because the law of a particular system just is the set of rules that
officials of a certain system are under a duty to apply, and the rule of recognition sets out the
content of this duty.2
Neither in his book nor in his essay does Shapiro cite any pages from Hart’s
writings in support of his attribution to Hart of the view summarized in this
quotation. The absence of citations is not very surprising, because Shapiro’s
attribution is not borne out by any of Hart’s statements. Hart in his remarks on
the Rule of Recognition talked not about the set of rules which the officials of a
certain system are under a duty to apply, but about the rules that belong to a system
of governance as its laws. Of course, the officials are indeed duty-bound under their
Rule of Recognition to apply those laws, but they can also be duty-bound to apply
other norms (such as laws of other jurisdictions in conflict-of-law cases). Hence,
although the Rule of Recognition does impose duties on officials, its distinctive and
pivotal role is to pick out the norms that belong to a given legal system as the laws
thereof.
Whereas no statements by Hart support the claim that he equated the laws of a
system of governance with the norms which the system’s officials are duty-bound to
apply, numerous statements in his writings support my understanding of his
position. Let us look briefly at a few of the relevant passages.
1.1. Hart’s own words

At the outset of his opening discussion of the Rule of Recognition, Hart recounted
its basic function: “The simplest form of remedy for the uncertainty of the regime of
primary rules is the introduction of what we shall call a ‘rule of recognition’. This
will specify some feature or features possession of which by a suggested rule is taken
as a conclusive affirmative indication that it is a rule of the group to be supported by
the social pressure it exerts.”3 Slightly later in that same discussion, Hart declared
2 Shapiro (2011: 85), emphasis in original. For an almost identical passage, see Shapiro (2009: 240).
3 Hart (1994: 94, emphasis in original).
24 Matthew H. Kramer
that even a very simple Rule of Recognition “brings with it many elements
distinctive of law. By providing an authoritative mark it introduces, although in
embryonic form, the idea of a legal system: for the rules are now not just a discrete
unconnected set but are, in a simple way, unified.”4
In Chapter 6 of The Concept of Law, in which Hart focused principally on the
Rule of Recognition, he maintained that the use of criteria “of recognition, by
courts and others, in identifying particular rules of the system is characteristic of the
internal point of view” (1994: 102). In the next paragraph of that same chapter,
Hart explicated the notion of legal validity in a passage that should be quoted at
some length:
[T]he word ‘valid’ is most frequently, though not always, used, in just such internal
statements, applying to a particular rule of a legal system, an unstated but accepted rule of
recognition. To say that a given rule is valid is to recognize it as passing all the tests provided
by the rule of recognition and so as a rule of the system. We can indeed simply say that the
statement that a particular rule is valid means that it satisfies all the criteria provided by the
rule of recognition.5
Hart went on to note that a duty-imposing norm can be valid as a law even if it is
very widely flouted with impunity, unless “the rule of recognition of the system
includes among its criteria, as some do, the provision (sometimes referred to as a
rule of obsolescence) that no rule is to count as a rule of the system if it has long
ceased to be efficacious” (1994: 103). With his understanding of validity as the
property of “count[ing] as a rule of the system,” Hart summarized the role of the
Rule of Recognition: “The rule of recognition provid[es] the criteria by which
the validity of other rules of the system is assessed” (1994: 105). As he remarked a
few pages later: “We only need the word ‘validity’, and commonly only use it, to
answer questions which arise within a system of rules where the status of a rule as
a member of the system depends on its satisfying certain criteria provided by the
rule of recognition” (1994: 108–9, emphasis in original). Again he wrote that “in a
mature legal system, we have a system of rules which includes a rule of recognition
so that the status of a rule as a member of the system now depends on whether it
satisfies certain criteria provided by the rule of recognition” (1994: 110).
These and other statements by Hart indicate that he regarded the Rule of
Recognition as “the rule providing criteria for the identification of the other rules
of the system” (1994: 111). That is, Hart took the Rule of Recognition to be the
array of ultimate standards by reference to which the officials of a legal system
identify the norms that belong to the system as laws. Nowhere did he suggest that it
is instead an array of standards that pick out all the norms which the officials of a
legal system are duty-bound to invoke and apply. Having rebutted Shapiro’s
misrepresentation of Hart’s model of the Rule of Recognition on this point, my
present discussion has punctured one of the principal reasons adduced by Shapiro
for rejecting that model.
4 Hart (1994: 95). 5 Hart (1994: 103).

1.2. Hart’s undeceivedness

Although the officials of a legal system are indeed duty-bound to invoke and apply
the norms that belong to their system as laws, they are also duty-bound in some
circumstances to invoke and apply other norms such as foreign laws. Hence, had
Hart taken the view which Shapiro attributes to him, he would have been in error.
As Shapiro proclaims:
For it can easily be shown that Hart’s rule of recognition transforms the law into a vortex
that sucks the rules of other normative systems into its voracious maw. As Joseph Raz
argued, judges are often under an obligation to apply laws of other jurisdictions in conflict-
of-law cases. In contract litigation involving a New York plaintiff and New Jersey defendant,
a New York court may be required to apply the New Jersey Statute of Frauds. On Hart’s
treatment of the rule of recognition as a duty-imposing rule, however, the New Jersey law
would automatically become incorporated into New York’s law because judges would be
under a legal obligation to apply it in certain cases. But this is clearly wrong: New York law
does not annex New Jersey law simply because there are occasions when New York officials
are required to apply the rules adopted by New Jersey officials.6
Had Hart indeed posited a state of equivalence between the norms identified by the
Rule of Recognition and the norms which legal officials are duty-bound to invoke
and apply, he would have been guilty of the blunder for which Shapiro indicts him.
In fact, however, he did not posit any such equivalence. Not a single statement in
Hart’s writings is vulnerable to Shapiro’s indictment, whereas a myriad of state-
ments bear out my claim that the function of the Rule of Recognition is to require
and enable the identification of the norms that belong to a given legal system as its
laws. (Note, incidentally, that the Rule of Recognition does not itself directly guide
all instances of law-ascertainment. Some laws are directly ascertained by reference
to other laws which in turn are directly ascertained either by reference to still deeper
laws or by reference to the ultimate Rule of Recognition itself. Though all the
norms directly identified by the Rule of Recognition in a legal system belong to the
system as laws thereof, not all the norms that belong to the system as laws are
directly identified by its Rule of Recognition.7)
Another reason for rejecting Shapiro’s criticism of Hart is that Hart himself in an
essay on the work of Hans Kelsen displayed a sophisticated awareness of the fact
that laws from other jurisdictions are not incorporated into a given legal system
when they are invoked and applied there in conflicts-of-laws cases. Hart dwelt on
that point—in opposition to Kelsen—over the course of several pages.8 Though his
essay on Kelsen was published somewhat after The Concept of Law, there is no reason
to think that in his great book he was unattuned to a point on which he insisted
lengthily a few years later (without the slightest indication that he himself had ever
adopted any position that rendered him liable to his own criticism of Kelsen).
6 Shapiro (2009: 246–7), footnote omitted.

7 This point is well made at several junctures in Greenawalt (1987).
8 Hart (1983: 314–21). Shapiro himself cites a couple of these pages in Shapiro (2009: 247 n. 45).
2. The overarching unity of the Rule of Recognition
In his essay on the Rule of Recognition, Shapiro submits that Hart’s account of law-
ascertainment leads to the conclusion that each legal system contains a multiplicity
of rules of recognition rather than a single overarching Rule of Recognition:
Hart does not explain what makes the rule of recognition a rule, as opposed to rules, of
recognition. Why think that the rule that validates executive orders of the Governor is the
same one that validates the regulations enacted by the New York State Legislature? Hart, it
seems, is able to establish the content of the law only by helping himself to the oneness of the
rule of recognition. But without establishing the unity of the New York rule of recognition,
he cannot show why the Governor’s orders ought to be included within the set of New York
law. Indeed, on Hart’s own theory of rule-individuation, according to which rules that guide
different audiences ought to be considered separate rules, many of the provisions of a
Hartian rule of recognition do not properly belong to the same rule. In any complex system,
different officials will be under duties to apply different rules. When this is so, there will be
multiple rules of recognition and hence the rules that they validate will not be part of the
same legal system.9
Attached to the final sentence of this passage is the following footnote: “At one
point, Hart sought to establish the unity of the rule of recognition by claiming
that any rule of recognition that sets out multiple criteria of legal validity will
also contain a provision determining the order of precedence in cases of conflict.
‘The reason for still speaking of “a rule” at this point is that, notwithstanding
their multiplicity, these distinct criteria are unified by their hierarchical arrange-
ment.’ . . . I must confess to not understanding Hart’s argument. Why is the mere
fact that a rule ranks certain criteria sufficient to incorporate those criteria into
the rule?”10
2.1. A puzzling footnote

Let us begin with Shapiro’s footnote. His expression of puzzlement is itself
puzzling, since Hart’s argument is straightforward and convincing.11 On the one
hand, as Hart was aware, the phrase “Rule of Recognition” is misleading in some
important respects. For example, as will become apparent later in this paper, the
so-called Rule of Recognition is an array of power-conferring norms and duty-
imposing norms rather than a single norm.12 Moreover, as will also become
apparent later, the officials who run a system of legal governance can differ with
one another about some elements of the Rule of Recognition that underpins their
9Shapiro (2009: 246), emphases in original, footnotes omitted.

10 Shapiro (2009: 246 n. 44), quoting Hart (1965: 1293).
11 I endorse Hart’s position in Kramer (2004: 107–8).
12 Of course, with enough logical connectives and operators, this array of norms can undoubtedly
be formulated as a single hideously unwieldy principle. (In a like fashion, though less messily, we could
formulate John Rawls’s principles of justice as a single principle.) However, it is best to understand the
Rule of Recognition as an array of norms in which criteria for legal validity are embedded.
system. Nevertheless, Hart’s stated reason for using the phrase “Rule of Recogni-
tion” is clear and persuasive. Whether the hierarchization of the ultimate law-
validating standards in a legal system ensues from prescriptions within those
standards themselves or from the ordering introduced by supplementary rules of
prioritization that are also comprised by the Rule of Recognition, it ties the
standards together as a coherently interrelated set of directives. The integratedness
which it bestows upon them is what justifies our designating those standards and
their rankings as an overarching Rule of Recognition. Although Hart’s label is
misleading in the respects already noted, it well captures the unity of a throng of
law-ascertaining touchstones that stand in quite clear relationships of superiority
and subordination to one another. (Of course, the order of priority among the law-
validating standards need not be perfectly comprehensive. The unity of the Rule of
Recognition in a particular legal system is consistent with a modicum of indeter-
minacy in the rankings among those standards.)
2.2. Differing norms and differing addressees

Shapiro contends that Hart malgré lui committed himself to the proposition that
every legal system comprises multiple rules of recognition addressed to differing sets
of officials. This contention by Shapiro is neither novel nor correct. I have else-
where replied at length to similar claims by Joseph Raz and John Finnis.13 Though
I will not here reiterate the whole of my rejoinder to those precursors of Shapiro’s
objection, the gist of the matter is as follows.
Notwithstanding that there is ample room for divergences on matters of detail
among the sundry officials in any particular legal system, those officials adhere to a
single Rule of Recognition. Their Rule of Recognition includes some standards that
instruct subordinate officials to treat the law-ascertaining determinations of super-
ior officials as binding. Lower-tier officials manifest their adherence to those
standards by indeed treating such determinations as dispositive and by criticizing
any lower-tier officials who do not conform to the upper-level determinations.
Naturally, the officials in the uppermost echelon of the system manifest their
acceptance of the deference-prescribing standards rather differently, since such
officials are not subordinate to anyone else. Norms in their Rule of Recognition
that call for deference to superior officials do not apply to them and therefore do
not directly guide their own ascertainment of laws. Nevertheless, their acceptance
of the deference-prescribing norms emerges in other ways. They are disposed to
criticize any deviations from those norms by lower-ranking officials, and they are
disposed to go further if the deviations are not rapidly corrected. Were the
deviations to persist, the topmost officials would arrange suitable penalties for the
contumacious subordinates. Through their preparedness to engage in rebukes and
(if necessary) in punitive measures, the officials at the top of a legal system’s
hierarchy join the lower-echelon officials in upholding a single Rule of Recognition.
13 Kramer (2004: 108–10).

They do not themselves seek to comply with the deference-prescribing directives in
that Rule of Recognition—since those directives are not applicable to them—but
they are firmly inclined to ensure that those directives are heeded by their subordin-
ates, to whom the directives are applicable. In that crucial respect, the upper-
echelon officials are adhering to the same Rule of Recognition to which the
compliant subordinate officials adhere. The overarching unity of their Rule of
Recognition is comparable to that of a religious code of appropriate observances
which includes some provisions that apply only to men and some provisions that
apply only to women; everyone in a society can be upholding that one code even
though its precise bearing on each person’s behavior will differ between the sexes.
In sum, so long as we keep in mind that the deference-requiring prescriptions in
a Rule of Recognition carry different implications for the officials in different strata
of a legal system’s hierarchy, we should encounter no difficulty in discerning the
cohesiveness of the system that is undergirded by that Rule of Recognition. Those
deference-requiring prescriptions directly guide some of the law-ascertaining deter-
minations of officials in the lower strata. Pari passu, they directly guide some of the
second-order determinations whereby the officials in the paramount stratum (and
in the subordinate strata) gauge the correctness of the law-ascertaining decisions
that have been reached by the lower-echelon officials. Thus, although the defer-
ence-requiring portions of a Rule of Recognition impinge on the officials’ behavior
in varying ways and to varying degrees, they impinge on every official’s behavior
to some extent. They constitute a common point of reference for the officials’ law-
ascertaining endeavors—that is, they set standards toward which the officials orient
themselves for varying purposes in those endeavors—just as do the other norms
in a Rule of Recognition. Hence, while comprising those deference-ordaining
provisions, a Rule of Recognition can perfectly well function as an integrated
assemblage of touchstones that are presupposed and upheld by authoritative actions
of the officials at every level.
3. The Rule of Recognition as a complex hybrid
Like many other critics of Hart, Shapiro asserts that “the Hartian framework . . .
famously acknowledges only two types of legal rules, namely, duty-imposing and
power-conferring” (2009: 239). Whereas Lon Fuller took the Rule of Recognition
to be exclusively power-conferring, Raz and Neil MacCormick have taken it to be
exclusively duty-imposing.14 Shapiro sides with Raz and MacCormick:
Is it possible, then, to understand the rule of recognition as either a power-conferring [rule]
or a duty-imposing rule? I think that the first option cannot be Hart’s position. For if we
suppose that the rule of recognition in Britain is ‘The Queen in Parliament has the power to
14 For my retort to the views taken by Fuller and Raz and MacCormick (with citations to their
relevant writings), see Kramer (2004: 104–5). If one had to choose between their positions, there is a
more substantial textual basis for Fuller’s understanding of the Rule of Recognition than for Raz’s and
MacCormick’s understanding.
create British law,’ we inadvertently convert Britain’s rule of recognition into its rule of
change. Moreover, the rule of recognition can validate certain types of customs, and since
customs need not be (and usually are not) created through the exercise of legal authority, the
rule that validates them cannot be power-conferring. The only alternative, then, is to treat
the rule of recognition as a duty-imposing rule. The rule of recognition, on this account,
imposes a duty on officials to apply rules that bear certain characteristics.15
3.1. Two preliminary points

Let us start with two preliminary points. First, it is not the case that power-
conferring norms and duty-imposing norms are the only norm-types that are
acknowledged as such in Hart’s theory of law. Hart was a sophisticated reader of
Hohfeld,16 and he was therefore well aware of the existence of liberty-conferring
laws and disability-imposing laws. In The Concept of Law, for example, he explicitly
invoked Hohfeld while referring to the disability-imposing laws that restrict the
legal powers of legislatures.17
Second, although the Rule of Recognition does impose duties, the final sentence
in this latest quotation from Shapiro again subtly misstates the nature of those
duties. What the officials are duty-bound to do under their Rule of Recognition is
to hold that norms which bear certain characteristics belong to their system as
binding laws that should be given effect.
3.2. A complex array

Duty-imposing though the Rule of Recognition is, it also confers powers. It is a
complex array of duty-imposing norms and power-conferring norms. Let us first
ponder this complex structure of the Rule of Recognition,18 before we go on to
assess Shapiro’s contention that the Rule of Recognition is not power-conferring.
The duty-imposing character of the Rule of Recognition ties in directly with its
power-conferring character. A legal system’s Rule of Recognition lays obligations
on the officials in the system to treat certain sources of laws as binding, and it
bestows powers on the officials to engage in authoritative acts of law-identification
that can fulfill those obligations. Whereas norms of adjudication empower officials
to ascertain authoritatively whether any violations of the prevailing laws have
occurred, the Rule of Recognition empowers them to ascertain authoritatively the
existence and contents of the laws themselves. (Of course, a law-ascertaining
determination is essential for any violation-detecting determination. Consequently,
when the former takes place, it often is an element of the latter. Nonetheless, the
two types of determinations can be distinguished analytically. Admittedly, distin-
guishing them in practice will in some contexts be tricky.) Precisely because the
15 Shapiro (2009: 239–40), footnotes and paragraph breaks omitted.

16 See, most notably, Hart (1982: 162–93).
17 Hart (1994: 289).
18 I have argued at length elsewhere for the proposition that the Rule of Recognition is both power-
conferring and duty-imposing. See Kramer (2004: 104–5). I here summarize the gist of my argument.
law-ascertaining endeavors of the officials are legally dispositive—that is, precisely
because the officials legally bind citizens and other officials with their findings, and
because they thereby alter people’s legal positions—their engaging in those en-
deavors of law-ascertainment consists in their exercising of legal powers vested in
them by their Rule of Recognition.
Were the Rule of Recognition not duty-imposing, the officials in a legal regime
would be legally at liberty to identify any norms at all as the laws of their system.
Were the Rule of Recognition not power-conferring, the officials would be unable
to identify the law in a legally binding fashion and would thus be unable to carry
out their duties as officials. They would not be able to undertake definitively the
process of law-ascertainment which they are duty-bound to perform in accordance
with the requirements which their Rule of Recognition imposes. In short, only the
hybrid composition of each Rule of Recognition provides both the structured
constrainingness and the dynamic operability of a legal system.
3.3. No collapse into rules of change

Let us now scrutinize Shapiro’s suggestion that the attribution of a power-conferring
role to the Rule of Recognition is a conflation of the Rule of Recognition with rules
of change. His suggestion mischaracterizes the way in which the Rule of Recognition
is power-conferring. The correct formulation of the strand of the British Rule of
Recognition concerning Parliamentary enactments is not “The Queen in Parliament
has the power to create British law,” but is instead “Every legal-governmental official
in the United Kingdom is authorized (as well as duty-bound) to hold that any norms
enacted by the Queen in Parliament as statutes are laws which belong as such to the
UK’s system of governance and which are therefore to be given effect by legal-
governmental officials in the UK.”19 Likewise, a somewhat simplified version of the
strand in the English Rule of Recognition that pertains to customary laws is correctly
formulated as follows: “Every legal-governmental official in England is authorized
(as well as duty-bound) to hold that customary norms endowed with certain
properties—such as longstandingness, reasonableness, and prevalent operativeness
within this jurisdiction—are laws which belong as such to the English system of
governance and which are therefore to be given effect by legal-governmental officials
in England.” As these formulations make clear, the powers directly conferred by the
Rule of Recognition are powers of law-ascertainment rather than powers of law-
alteration. In other words, once those powers are correctly encapsulated, we can see
that Shapiro has erred in maintaining that the ascription of a power-conferring role
to the Rule of Recognition would efface the distinction between two types of
secondary norms.
19 This formulation prescinds from various complexities such as the fact that certain Parliamentary
enactments are applicable to the whole of the United Kingdom whereas other such enactments are
applicable to only some of the four main components of the UK. Providing a full and precise account
of any of the strands in a real-world Rule of Recognition is no easy task, as Kent Greenawalt revealed a
quarter of a century ago in his classic article on the American Rule of Recognition.
4. Rules and regularities
Quite puzzlingly, Shapiro in Legality repeatedly accuses Hart of the category

mistake of conflating social rules with the behavioral and attitudinal regularities
of social practices. He first sounds this theme on the second page of his chapter
about Hart, when he writes that “for Hart, social rules are social practices.”20 He
expands on this theme several pages later: “Hart tries to demystify the creation of
social rules by reducing them to social practices. In other words, social practices
generate rules because these rules are nothing but social practices. Thus, when men
regularly take off their hats in sacred places because they take the internal point
of view toward this sort of behavior, the behavioral regularity is itself a rule.”21
After repeating this characterization of Hart’s account of rules a couple of further
times, Shapiro points out the obvious erroneousness of any reduction of rules to
regularities:
Social rules cannot be reduced to social practices[,] because rules and practices belong to
different metaphysical categories. Rules are abstract objects. They are like games, numbers,
plots, propositions, and concepts—they are objects of thought, not entities that exist within
space and time. . . . Practices, on the other hand, are concrete events. They take place within
the natural world and causally interact with other physical events.22
Shapiro is of course correct in declaring that rules cannot be reduced to the
regularities of practices, but his ascription of a reductionist project to Hart is
extraordinarily ungenerous. Once again, he provides no citations while insisting
that Hart espoused a blunderingly mistaken view. We should not accept such an
insistence in the absence of any strong evidence that supports it.
On the one hand, despite the absence of any citations by Shapiro, there are some
statements in The Concept of Law that might be seized upon by an overweeningly
uncharitable reader as supportive of the accusations of a category mistake. For
example, Hart wrote that “if a social rule is to exist[,] some at least must look upon
the behaviour in question as a general standard to be followed by the group as a
whole.” He likewise asserted: “How many of the group must . . . treat the regular
mode of behaviour as a standard of criticism, and how often and for how long they
must do so to warrant the statement that the group has a rule, are not definite
matters” (1994: 56). In a broadly similar vein, he declared that “to refer to the
internal aspect of rules seen from [the] internal point of view” is “to bring into the
account the way in which the group regards its own behaviour” (1994: 90).
On the other hand, for two main reasons, these quoted statements—and some
other similar pronouncements by Hart—furnish no support for Shapiro’s allega-
tions that Hart committed a crude metaphysical blunder. First, both in philosoph-
ical discourse and in ordinary English, terms such as “standard” and “norm” are
20 (2011: 80), emphasis in original.

21 (2011: 95), emphasis in original.
22 Shapiro (2011: 103), emphases in original.
equivocal. Each of those terms can refer either (i) to an abstract normative entity
toward which some practice is oriented or (ii) to the behavioral and attitudinal
regularities that make up the practice. Hart plainly sometimes used the term
“standard” in the latter sense, but there are no grounds whatsoever for the claim
that he collapsed the former sense into the latter. Likewise, although he wrote both
about the adoption of an internal perspective on rules and about the adoption of an
internal perspective on the behavioral/attitudinal regularities that are oriented
toward those rules, he nowhere suggested that the rules are nothing more than
the regularities. Rather, he was simply reflecting a duality in ordinary ways of
discussing normative attitudes. Quite unobjectionably and idiomatically, one can
sometimes speak about the harboring of such attitudes toward the norms that are
immanent in our practices, and one can at other times speak about the harboring of
such attitudes toward the behavioral and attitudinal regularities of those practices.
That duality in those idiomatic ways of discussing these matters will not strike
anybody as muddled, unless one’s interlocutor or reader is relentlessly determined
to find confusion where there is none. Exactly because the regularities of a com-
munity’s practices are so closely connected to the norms that guide and structure
those regularities, either the norms or the regularities can aptly be regarded as the
objects of normative attitudes. Likewise, we can properly speak either about
deviations from the norms or about deviations from the regularities of the practices
that embody those norms. To accept as much is hardly to commit the egregious
error of presuming that the norms are nothing but the regularities. Despite all the
italics in Shapiro’s accusations, there is no sign that Hart committed that egregious
error. (As should be evident, incidentally, I take social practices to be complexes of
particulars and universals—namely, complexes of behavioral/attitudinal regularities
and the norms immanent in those regularities. The relationship between a practice
and its norms is constitutive and mereological.)
Second, far from committing the error just mentioned, Hart frequently evinced
his awareness of the distinction between the regularities of social practices and the
norms toward which those practices are oriented. For example, he adverted to “the
regularities of observable behaviour in which conformity with the rules partly
consists” (1994: 89). Had Hart taken the position which Shapiro tries to foist
upon him, he would not have referred here to conformity with the rules; he would
instead have written about the regularities of behavior in which the rules themselves
partly consist. Given that he did refer here to conformity with the rules, we can
rightly infer that he did not take the rules to be reducible to the behavioral
regularities and associated attitudes of which social practices are composed.
In a similar vein, Hart wrote that “[t]he social practices which underlie . . . legis-
lative authority will be, in all essentials, the same as those which underlie the simple,
direct rules of conduct, like that concerning baring the head in church” (1994: 58).
Instead of submitting that social rules simply are the regularities of social practices,
Hart here correctly maintained that such rules are underlain by those regularities.
Though The Concept of Law is not a treatise on metaphysics, Hart’s firm grasp of
the metaphysics of social rules is evidenced in pronouncements such as this quoted
statement.
One further example will suffice. Hart declared that “our social group has . . . rules
which, like that concerning baring the head in church, makes [sic] a specific kind
of behaviour standard” (1994: 57). Had Hart espoused the position which
Shapiro imputes to him, he would not have written that the rules make certain
patterns of behavior standard. Rather, he would have written that the rules are
certain standard patterns of behavior. Once again, that is, we find in Hart’s text
a solid awareness of the distinction between social rules and the regularities
of social practices.
In short, there is no basis for Shapiro’s allegations that Hart muddled the
distinction between rules as universals and behavioral/attitudinal regularities as
particulars. Shapiro’s allegations are so tendentious as to be baffling.
5. Individualized reasons and social norms

Shapiro asserts that Hart “might admit that social rules are not reducible to social
practices, but claim that social practices necessarily generate social rules in their
wake.” He retorts: “The problem with this version of [Hart’s theory] is that the
metaphysical relation it claims exists simply does not obtain: social practices do not
necessarily generate social rules” (2011: 103). Shapiro’s mistake here is subtle but
far-reaching.
5.1. The wrong conditional

Shapiro in effect maintains that Hart might be drawn toward the conditional
proposition “If there is a social practice at some time t in some community C,
then there are some social rules in C at t.” In fact, the conditionality should run
in the opposite direction: “If there are some social rules in C at t, then there is at
least one social practice in C at t.” That is, the claim toward which Hart should be
drawn is that the existence of social rules presupposes the existence of social
practices—rather than the claim that the existence of a social practice entails
the existence of social rules. I shall henceforth designate the former claim as the
“Presuppositions of Rules Conditional” and the latter claim as the “Entailments of
Practices Conditional.”
Why should Hart be drawn toward the Presuppositions of Rules Conditional
rather than toward the Entailments of Practices Conditional? As Shapiro maintains,
one of the problems that must be addressed by any philosophical theory of law
(at least implicitly, if not explicitly) is what he designates as the “Possibility
Puzzle.”23 He perceives that puzzle as a classic chicken-and-egg problem, which
he formulates as follows:
23 The initial main discussion of this problem occurs in Shapiro (2011: 37–40). Shapiro then
broaches the problem recurrently throughout the first half of his book.
Egg : Some body has power to create legal norms only if an existing norm confers that power.
Chicken: A norm conferring power to create legal norms exists only if some body with power
to do so created it.24
As Shapiro indicates, Hart in effect resolved the Possibility Puzzle “by rejecting the
Chicken Principle. . . . And by rejecting the Chicken Principle, Hart stops the
infinite regress: while legal authority must always be conferred by norms, legal
norms may [ab initio] be created without authority” (2011: 95–6).
When Shapiro characterizes Hart’s rejection of the Chicken Principle, he usually
implies that Hart embraced the Entailments of Practices Conditional. For instance,
he attributes to Hart the following view: “To create a social rule, those involved need
not be authorities; rather, they need only engage in a social practice” (2011: 95).
Occasionally, however, Shapiro instead ascribes to Hart the Presuppositions
of Rules Conditional: “Hart’s solution to the Possibility Puzzle is a scrupulously
positivistic one. Legal systems, on his view, are necessarily founded on social
practices” (2011: 97). This latter ascription is correct, for the task of establishing
the possibility of the existence of legal norms requires a theorist to delineate the
presuppositions of their existence. In so doing, Hart of course had to maintain that
social practices which generate social norms (such as the Rule of Recognition) are
possible. He did not have to contend—and wisely did not contend—that all social
practices generate social norms.
5.2. Norm-guided behavior versus habitual behavior

Of course, to address the Possibility Puzzle satisfactorily and thereby propound a
full-blown theory of law, Hart had to specify in some detail the distinction between
social practices which generate norms and social practices which do not. Shapiro
(2011: 104–5) is correct to state as much. What is remarkable, however, is any
suggestion that Hart did not draw just such a distinction. He in fact devoted many
pages of The Concept of Law to differentiating between rule-guided behavior and
merely habitual behavior (1994: 9–11, 55–60). Merely habitual social practices do
not give rise to social norms, whereas social practices marked by the widespread
adoption of a critical reflective attitude among the participants in the practices do
give rise to social norms.
Shapiro is of course aware of the pages in The Concept of Law to which I have just
referred, but he believes that Hart’s distinction is untenable. More specifically, he
believes that some social practices marked by the widespread adoption of a critical
reflective attitude among the participants do not give rise to any social norms. He
begins his argument on this point by adducing a sporting example that was
originally propounded by Geoffrey Warnock and later by Raz.25 However, because
I have rebutted that example elsewhere at some length (1999: 251–3), and because
Shapiro adds nothing to it that would shield it from my rebuttal, I will move
straightaway to his own examples.
24 Shapiro (2011: 40). 25 Shapiro (2011: 103–4).

A lengthy quotation is warranted here:

In fact, many (if not most) of our social practices [marked by a critical reflective attitude
among the participants] fail to generate social rules. Among the professional class in the
United States, for example, it is now generally accepted that people ought not to smoke even
when no one else is affected. Smokers are routinely criticized by nonsmokers. Smoking, they
say, is “stupid,” is a “dirty habit,” and sets a “bad example.” Moreover, these nonsmokers are
not criticized by other nonsmokers for engaging in such criticism. Yet there is no social rule
against smoking alone or with other smokers. Members of this group simply believe that
there are excellent reasons not to smoke, reasons that exist independently of the practice and
which the practice is supposed to promote. Similarly, there are no social rules requiring that
people save for their retirements, that acquaintances say hello to one another, that drivers
lock their cars at night, and that people comb their hair, despite the fact that nearly everyone
accepts these practices from the internal point of view.26
From his armchair, Shapiro makes some large empirical claims in this passage.
Let us grant that his armchair sociology is accurate and that individuals in the
professional class in the United States act as officiously or discourteously toward
one another as he suggests. What will be contested here is his philosophical
interpretation of those empirical claims.
5.2.1. The example of smoking

Shapiro’s description of the situation in the United States with regard to smoking is
somewhat equivocal. If a social norm against smoking in the absence of any non-
smokers (as well as in the presence of non-smokers, of course) prevails among
professional people in the United States, then all three elements of the Hartian
critical reflective attitude have to be widely operative. That is, (1) most professional
people do not smoke even in the absence of any non-smokers; (2) most professional
people criticize other professional people who smoke, on the ground that they are
deviating from a shared standard of behavior to which they are urged to conform;
and (3) most professional people accept that the criticism and the pressure for
conformity to the shared standard are appropriate responses to deviations from that
standard. Let us suppose that the first of these elements of the critical reflective
attitude is indeed present in the contemporary United States. What about the
second element?
According to Shapiro, criticism of smokers among American professionals is
focused on three main properties of the habit of smoking: its stupidity, its dirtiness,
and its tendency to set a bad example. The first two of these three lines of criticism
appear to be directed not against deviations from a shared standard but instead
against individualized imprudence. Though everyone has a prudential interest in
his or her own abstention from smoking, nobody has such an interest in anyone
else’s abstention (if we leave aside the interests of dependents, which seem to lie
outside the scope of the criticism recounted by Shapiro). Consequently, the
26 Shapiro (2011: 104).

negative remarks by professionals about the stupidity and dirtiness of a smoker’s
habit pertain solely to the smoker’s foolishness in neglecting her own interests,
rather than to any departures by the smoker from some shared standard in which
everyone has a stake. Heretofore, then, Shapiro appears to be correct in submitting
that there is no anti-smoking social norm among contemporary American profes-
sionals. However, he appears to be correct on that point precisely because the
second element of the Hartian critical reflective attitude appears heretofore to be
missing. In other words, with regard to the remonstrations that are focused on the
stupidity and dirtiness of smoking, Shapiro’s denial of the existence of any social
norms is entirely consistent with Hart’s analysis of social norms.
When we come to the third objectionable property of smoking—its tendency to
set a bad example—things are rather different. Here the criticism of smokers by
non-smokers does appear to be oriented toward a shared standard from which the
smokers are taken to be deviating. We need not pin down here whether that
standard should be understood as a concrete norm against setting a bad example
by smoking or instead as a more general norm against setting a bad example in one’s
everyday behavior. The point is that objections which concentrate on one’s setting
of a bad example are not adverting to individualized imprudence. Rather, they are
adverting to one’s departure from a shared standard concerning one’s responsi-
bilities to other people.
Hence, when we look at the third line of criticism mentioned by Shapiro, we
find that he is incorrect in contending that no relevant social norm obtains. His
stance is inconsistent with the stance that would have been taken by Hart on the
matter, but the inconsistency arises because Hart would have been led by his
analysis of social norms toward a correct position rather than toward Shapiro’s
mistaken position.
5.2.2. Shapiro’s other examples

Three of Shapiro’s four other examples—saving for one’s retirement, locking the
doors of one’s car at night, and combing one’s hair—appear to involve individual-
ized imprudence rather than departures from a shared standard.27 Given as much,
no social norm is being invoked when someone is criticized for failing to save for
her retirement or for failing to lock the doors of her car at night or for not combing
her hair. However, the absence of a social norm is fully consistent with Hart’s
analysis of social norms, because the second element of the critical reflective
attitude is not operative when the target of people’s criticism is individualized
imprudence. In such circumstances, the criticism does not assert or imply that the
imprudent person has deviated from a shared standard of conduct to the detriment
of others.
27 Here and in the preceding subsection I have used the verbs “appear” and “seem” (and their
cognates) quite extensively because Shapiro’s examples are presented very laconically. Those examples
could perhaps be fleshed out in ways that would lead me to change my concrete assessments of them—
though the general nature of my rejoinder to Shapiro would remain unchanged, of course.
The one remaining example in Shapiro’s list is that of greetings between

acquaintances. That example does appear to involve a social norm, since the matter
at issue pertains to the basic civility that is owed by friends or colleagues to one
another. When someone is reproached for failing to say “hello” in a context where
such a greeting would be expected as an elementary courtesy, the reproach is
directed at her deviation from a shared standard that prescribes how people should
interact with one another. Given that most people in the society accept the
appropriateness of such a reproach, all three elements of the critical reflective
attitude are present. Hence, a social norm is being invoked. Shapiro errs in trying
to assimilate this example to his other examples.
Of course, if someone persistently declines to greet acquaintances, they might
nonetheless forbear from upbraiding him. Perhaps they believe him to be autistic
and therefore beyond the reach of the relevant social norm, or perhaps they believe
that he is incorrigibly boorish and that expostulating with him about quite a minor
matter of social etiquette would thus be pointless and frustrating. A social norm
that requires greetings between acquaintances would continue to obtain in such
circumstances, since most people would continue to be disposed to uphold that
norm as a shared standard of conduct in most contexts. They might stoically
acquiesce in the lapses of the inveterate boor, but they will be disposed to object
to other significant lapses.
In the situation recounted by Shapiro, people are disposed to censure deviations
from the practice of greeting acquaintances. Shapiro explicitly states (in the long
passage quoted above) that “everyone accepts these practices [including the practice
of greeting acquaintances] from the internal point of view.” Given that everyone
adopts the internal point of view vis-à-vis the practice of greeting acquaintances,
everyone exhibits all three elements of the critical reflective attitude in relation to
that practice. Hence, everyone is disposed to censure deviations from the norms of
that practice and to do so because such deviations are perceived as departures from a
shared standard or set of standards. Shapiro apparently believes that the censure is
instead addressed to instances of individualized imprudence; however, such a view
of the matter is quite implausible. When requirements of elementary civility are the
point of reference for criticism of people’s conduct, the criticism is invoking social
norms rather than merely considerations of individualized imprudence.
6. Simulative statements versus internal statements
Shapiro thinks that Hart’s theory not only fails to solve the Possibility Puzzle but
also violates Hume’s Law: namely, the principle that there is no valid argument in
which all the premises are non-normative (and logically consistent) and in which
the conclusion is a substantive normative proposition.28 The problem here, as
Shapiro presents it, derives from Hart’s contention that the endeavors through
28 This formulation glosses over some complexities that are fully explored in Kramer (2009: 6–9).
which the officials in any jurisdiction ascertain the existence and contents of laws
are norm-generating social practices. How does each official derive normative
conclusions from non-normative facts about other officials’ conduct and attitudes?
As Shapiro suggests, Hart maintained that each official derives such conclusions by
adopting the internal point of view—the critical reflective attitude—toward the
other officials’ practices. Shapiro recounts Hart’s position as follows:
Thus, one may take the internal point of view toward the social practice of rule recognition
and, in so doing, treat it as a standard for guidance and evaluation. The normative judgments
that are formed through this practical engagement with social practice can then be used
to derive other normative judgments about legal rights, obligations, and validity. Legal
reasoning in a positivistic framework does not, therefore, violate Hume’s Law, insofar as
legal judgments are derived only from other, similarly normative judgments.29
Shapiro insists that Hart did not succeed in escaping from contraventions of
Hume’s Law, because people can reason about the law and reach legal conclusions
even if they have not adopted the internal perspective toward the practices of law-
ascertainment in their jurisdiction. They can derive normative inferences about the
law while occupying an uncommitted perspective. He writes:
The [uncommitted observer] not only can talk the talk; he can think the thought. He too
can “think like a lawyer.” Legal reasoning, we might say, is a remarkably open process. Even
those who judge the law morally illegitimate, or reject it for self-interested reasons, can figure
out what the law demands of them. Indeed, it would be bizarre if the only people who could
understand the law were those who accepted it. The law claims the right to demand
compliance from everyone, even those who reject its demands.30
According to Shapiro, Hart was unable to explain how an uncommitted observer
could manage in this fashion to draw normative conclusions from the non-norma-
tive facts of officials’ conduct and attitudes. The uncommitted observer does not
ascribe normative force to the conduct and attitudes, but is able to arrive at
conclusions about legal norms and obligations. Hart could not credibly deny that
the drawing of such conclusions by such an observer is possible, yet within the
confines of his theory the drawing of such conclusions by such an observer is a
contravention of Hume’s Law. So Shapiro contends.
Shapiro is aware that Hart himself in his later work invoked a distinction with
which he was able to explain why an uncommitted observer can reach conclusions
about legal norms and obligations. That is, Hart invoked Raz’s distinction between
committed statements and detached statements.31 (Because I have elsewhere re-
elaborated and slightly modified Raz’s distinction in terms of internal statements
versus simulative statements—Kramer [1999: 165–70]—I shall henceforth use the
latter pair of phrases except when I am directly referring to Raz’s writings.)
29 Shapiro (2011: 111–12). 30 Shapiro (2011: 112).

31 Hart (1982: 153–5). For Raz’s principal discussions of the distinction, see Raz (1979: 153–7);
(1980: 234–8); (1999: 171–7).
Someone who occupies the simulative perspective does not merely attribute
normative beliefs to officials and some citizens; in addition, he articulates such
beliefs as if they were his own. He speaks or writes as if from an internal point of
view without actually occupying that point of view. To some degree, a person
engaged in simulation is quite closely akin to a theatrical actor who recites lines in a
play. If an actor plays Iago in a production of Othello, for example, he will recite
Iago’s lines and generally do his best to convey Iago’s thoughts and emotions—his
point of view—to the audience. However, even a superb performance of the role
does not perforce count at all as an endorsement of Iago’s villainy. When a skillful
actor strives to fathom Iago’s outlook and thereby manages to deliver a splendid
performance, he nonetheless might disapprove strongly (and very likely does
disapprove strongly) of the vile sentiments which he is feigning to express.
Yet, in a crucial respect, the analogy between the theatrical actor’s performance
and the simulative speaker’s discourse breaks down. Though some productions of
plays require or permit the improvisation of numerous lines, most do not; a
production typically involves the recitation of lines that have been written before-
hand by a playwright. In any normal staging of Othello, for example, the man who
plays Iago will recite some or all of the lines that were written for that character by
Shakespeare. He will add few or no lines himself. By contrast, the speaker or writer
who adopts a simulative perspective will frequently elaborate the implications of
the set of beliefs which he is articulating. He may well draw inferences or develop
arguments or undertake extrapolations that have not thitherto occurred to anyone
else. Furthermore, even when his discourse is a straightforward recapitulation of
thoughts and sentiments that have previously been expressed by other people, he
typically does not simply repeat lines verbatim that have already been written. In
sum, a simulative utterance is both illuminatingly similar to a thespian performance
and importantly different from it. Each of those two modes of communication
consists in giving voice to a point of view that is not one’s own, but—even if
we allow for the creativity and ingenuity that are involved in a deft theatrical
performance—a simulative perspective leaves much more latitude for innovation
than does the recitation of lines.
By adverting to the as-if role of the simulative perspective, Hart could account
for the ability of an uncommitted observer to articulate normative conclusions
about the law in this or that jurisdiction. To attack Hart, then, Shapiro assails Raz’s
distinction between committed statements and detached statements. He presents
his summary and critique of that distinction as follows:
According to Raz, committed and detached statements express the same normative propos-
ition but have different truth conditions. Committed statements have normative truth
conditions, whereas detached statements have exclusively descriptive truth conditions . . .
Hart seized on Raz’s theory of normative statements and argued that the bad man can
describe the law in normative terms and engage in legal reasoning because legal statements
can be detached. Thus, when the [uncommitted observer] ascribes legal validity to a rule, he
is not expressing his acceptance of the rule of recognition, but rather the point of view of
those who accept the rule of recognition. This detached statement is true just in case there is
judicial acceptance of a rule of recognition that validates the rule in question, regardless of
whether the utterer accepts that rule of recognition himself. Unfortunately, Raz’s distinction
between committed and detached statements is a mysterious one . . . [T]he distinction
presupposes an unorthodox semantic theory. According to Raz, both committed and
detached statements express the same proposition but have different truth conditions.
The usual semantic assumption is that propositions are individuated according to their
truth conditions.32
Shapiro’s contentions are not reconcilable either with the natural way of under-
standing Raz’s distinction or with most of what Raz himself says about that
distinction. The committed/detached dichotomy is naturally understood (by a
philosopher) as a matter of pragmatics rather than as a matter of semantics. That
is, it is naturally understood as a difference between the purposes for which people
utter various statements rather than as a difference between the meanings which
people’s statements bear. Given that Raz’s dichotomy is a difference of the former
sort rather than of the latter, it does not pertain to any distinction between truth-
conditions. If the content of a simulative utterance is the same as that of an internal
utterance, then the truth-conditions for the utterances are likewise the same.
This natural understanding of Raz’s dichotomy is borne out by his own most
straightforward discussion of the matter. Contrary to what Shapiro asserts, Raz in
The Concept of a Legal System makes clear that the truth-conditions for a detached
statement that bears the content C are the same as the truth-conditions for a
committed statement that bears the content C. Raz declares that “a detached
statement normally made by the use of a certain sentence is true if and only if
the committed statement normally made by the use of the same sentence is true—
given the non-normative facts of this world—if all the ultimate rules of the legal
system referred to are binding and if there are no other binding normative consider-
ations.”33 Quite bewildering, in light of this quotation, is Shapiro’s assertion (in the
long passage excerpted above) that Raz maintains that “detached statements have
exclusively descriptive truth conditions.”
In sum, the simulative/internal distinction—the distinction that would enable
Hart to avoid any contraventions of Hume’s Law—is a matter of pragmatics rather
than a matter of semantics. If a sentence S is uttered simulatively, the truth-
conditions for that simulative statement are the same as the truth-conditions for
an internal statement that consists in the utterance of S. (Indeed, this point falls out
of the minimalist account of truth which I have advocated elsewhere.34) Hence,
pace Shapiro, there is nothing mysterious about the simulative/internal dichotomy.
That dichotomy is available to be invoked pertinently by anyone who seeks to
defend Hart against Shapiro’s onslaughts.
32 Shapiro (2011: 415 n. 44).

33 Raz (1980: 237–8). Much more ambiguous is a sketchy remark in Raz (1999: 177). That remark
could be construed in line with what Shapiro says about Raz, but it could equally be construed in line
with what I say about him. Given that I am mainly interested here in the simulative/internal distinction
itself rather than in the exegesis of Raz’s texts, I shall leave aside the puzzlingly ambiguous pronounce-
ment to which I have just referred.
34 Kramer (2007: 71–82); (2009: 261–75, 287–8).
7. The nature of legal concepts
In opposition to Raz, Hart contended that the concepts of obligation and authority
in legal contexts are not equivalent to those concepts in moral contexts. Shapiro, in
agreement with Raz, insists that Hart’s position on that matter is untenable: “Hart’s
attempt at preserving the distinction between legal and moral thought stakes out a
middle ground that is unstable and unsupportable. For once it is admitted that legal
concepts are normative, it becomes difficult to deny, as Hart did, that they are
moral as well.”35 Having elsewhere lengthily defended Hart’s general position
against Raz,36 I will here much more briefly redefend that general position. In so
doing, however, I will not be endorsing Hart’s own way of arguing for his stance.
As my previous discussion of this matter has made clear, Hart’s attempt to vindicate
his position through a strange analysis of the deontic auxiliary “ought” is to be
discountenanced.37 Raz and Shapiro are right to take exception to Hart’s analysis.
Nonetheless, although Hart did not argue very adeptly in favor of his claims about
the nature of basic legal concepts, those claims are correct.
7.1. A conflation of two dimensions of morality

In Legality, the main complaint by Shapiro is that Hart’s position implies that the
authoritative pronouncements of adjudicative and administrative officials are
amoral in character. Shapiro (2011: 111, 115) retorts: “The claims of law are far
too serious to accept the possibility that they are amoral in nature. . . . For once we
focus on the role that legal judgments and claims play in social life, it becomes hard
to deny that they are constituted not only by normative concepts and terms, but by
moral ones as well.” Shapiro’s chief argument in support of his retort should be
quoted in full:
But once we recognize the social role that legal concepts are typically made to play, it
becomes hard to resist the conclusion that these concepts must be moral as well. For
concepts such as AUTHORITY and OBLIGATION are characteristically employed in
contexts where moral notions are uniquely apt, namely, where the conduct being guided
and evaluated is contrary to our self-interest, requires considerable sacrifice, significantly
affects the lives of others, and subjects the actor to punishment and censure for failure to
comply. Thus, claims of legal authority and obligation are not merely used to guide conduct;
they are used to make demands that materially constrain freedom. They are not used simply
to assess behavior; they are employed to assess guilt, blame the guilty, and ground coercive
and punitive responses. In other words, we are not talking here about demands to go to jail
in Monopoly, stand up when an elder enters the room, or avoid splitting infinitives. In legal
contexts, we require people to pay their taxes, join the army, pass difficult licensing exams
35 Shapiro (2011: 111).

36 Kramer (1999: ch. 4). I obviously cannot here advance afresh all the arguments which I have put
forward on this matter in Chapter 4 of In Defense of Legal Positivism. I hope that at some point Shapiro
will familiarize himself with those arguments.
37 Kramer (1999: 84).
before practicing a profession, and testify in a criminal trial, under the threat of jail or heavy
fines. Only moral concepts have the heft to make such serious claims.38
Shapiro has here conflated two distinctions which I have elsewhere sought to
disentangle: the distinction between morality and prudence, and the distinction
between the moral and the non-moral.39 Very roughly, the former contrast is
between (1) reasons-for-action that are focused primarily or exclusively on the
interests of other people rather than primarily or exclusively on one’s own interests
and (2) reasons-for-action that are focused primarily or exclusively on one’s own
interests rather than primarily or exclusively on the interests of other people. The
moral/non-moral contrast is quite different. It distinguishes between (3) phenom-
ena that are appropriately subject to moral assessment and (4) phenomena that are
not appropriately subject to moral assessment. This moral/non-moral contrast can
be further explicated in part as a distinction between (3A) actions or decisions that
bear non-trivially on the interests of other people and (4A) actions or decisions that
do not bear more than trivially on the interests of other people.
When morality is differentiated from non-moral phenomena along the lines
just indicated, it is evident that the actions and decisions undertaken by legal-
governmental officials fall on the moral side of the divide. Nobody could sensibly
think otherwise. As Shapiro emphasizes in the long passage just quoted, legal
requirements and judgments impinge very heavily on the interests of the myriad
people who are affected by them. Manifestly, such requirements and judgments are
appropriately open to moral assessment. Hence, if anyone is faced with the question
whether legal concepts are moral or non-moral, the only tenable answer is that they
are moral.
However, that question is not the question which Hart was addressing when he
impugned Raz’s insistence that the concept of obligation is the same in law as in
morality. Hart was focusing on the moral/prudential dichotomy rather than on the
moral/non-moral dichotomy. He was maintaining that the normativity of the
public interaction among adjudicators and administrators in a given regime can
credibly be prudential rather than moral. The reasons-for-action implicitly or
explicitly adduced by the officials in their exchanges with one another can credibly
be centered on their shared interests in sustaining an oppressive system of govern-
ance. Moreover, the pronouncements of the officials in their interaction with
ordinary citizens can credibly be imperative rather than truly prescriptive; that is,
the officials can be issuing orders to the citizens instead of presenting them with any
punishment-independent reasons for conformity to the law. So Hart maintained,
and so I have elsewhere contended at length in support of his position. His stance
on this matter has nothing to do with the misguided view that the concept of
obligations invoked by legal-governmental officials is somehow non-moral (or
“amoral”). On the contrary, Hart fully recognized that any invocation of that
concept by such officials is pertinently open to moral appraisal. Instead of ever
38 (2011:114), footnote omitted.

39 Kramer (1999: 63–77, 123–5, 189–91, 200–4).
absurdly suggesting otherwise, Hart simply held that the officials need not be
invoking or purporting to invoke non-prudential reasons for conformity to the
law when they give effect to legal obligations in their interaction with citizens. They
can be asserting sheer imperatives. In so doing, of course, they render themselves
susceptible to moral condemnation.
7.2. An alternative conception of legal obligations

If the concept of legal obligations is not equivalent to the concept of moral
obligations, how are legal obligations to be understood? I have answered that
question in detail in the fourth chapter of my 1999 book In Defense of Legal
Positivism. Here the pithiest sketch of an answer will have to suffice.
Legal obligations are the requirements imposed by any legal norms that render
certain courses of conduct mandatory. If I am under a legal duty to do X in
circumstances C, I will be conducting myself in violation of some legal norm(s) if
I do not do X in circumstances C. By failing or refusing to comply with what is
required of me under the terms of some legal norm(s), I commit a legal wrong;
I contravene the terms of the norm(s) by acting at odds with what those terms
render mandatory. This straightforward explication of the notion of legal duties
does not assert or presuppose that every legal duty is a moral duty, nor does it
presuppose or assert that every legal duty is implicitly or explicitly presented as a
moral duty in the authoritative utterances of legal officials.
Shapiro himself as a legal positivist is keen to deny that every legal duty is a moral
duty, of course. However, like Raz, he maintains that every legal duty is implicitly
or explicitly presented as a moral duty in the authoritative utterances of officials. As
he declares, “from the legal point of view the law’s activities [give rise to obligatory
moral reasons for compliance with legal mandates].”40 Whether or not any of the
officials in a legal system sincerely believe that the requirements imposed by the
system are morally binding upon citizens, their authoritative pronouncements
concerning the effectuation of those requirements presuppose or assert just such
bindingness. So Raz and Shapiro contend.
My account of legal obligations, laconically summarized in the penultimate
paragraph above, is squarely at odds with Raz’s and Shapiro’s contentions. It of
course allows that the authoritative obligation-enforcing pronouncements of officials
in countless legal-governmental systems (including many evil legal-governmental
systems) do assert or presuppose the moral bindingness of the obligations that are
being enforced, but it leaves room for other legal-governmental systems in which
such pronouncements are sheer imperatives rather than prescriptions. In systems
of the latter sort, the normativity of the discourses that take place among adjudi-
cative and administrative officials in their public roles is prudential rather than
moral.
40 Shapiro (2009: 259).

Though I cannot here repeat the several long arguments from In Defense of Legal
Positivism in support of my account of legal obligations, what should be re-empha-
sized is that the points of contention between Hart and Raz are not centered at all
on the moral/non-moral dichotomy. Neither Hart nor any other competent legal
philosopher has failed to recognize that, insofar as “moral” is contrasted with “non-
moral” (or “amoral”), law is inherently a moral phenomenon. Insofar as “moral” is
contrasted with “non-moral,” legal concepts are inherently moral. Everybody recog-
nizes as much.
Instead of being focused on the moral/non-moral division, the Hartian challenge
to Raz is focused on the moral/prudential division. Is every legal system such that
the reasons for conformity to the law explicitly or implicitly invoked by officials in
the course of giving effect to legal mandates are moral, or can there credibly be a
legal system in which the reasons for conformity to the law invoked by officials in
the course of giving effect to such mandates are punishment-focused prudential
reasons (and in which the mandates themselves are therefore correctly construable
as imperatives rather than as prescriptions)? While maintaining that a legal system
of the latter type is indeed credibly possible, Hart and his defenders have hardly
maintained that the operations of such a system are non-moral.
8. Disagreement over the Rule of Recognition

In his determination to undermine Hart’s model of the Rule of Recognition,
Shapiro joins forces with Dworkin: “[A]s Ronald Dworkin has famously argued,
[Hart’s] account of the criteria of legal validity is seriously flawed. For if it were
correct, widespread disagreements about such criteria would be highly problematic.
Since the criteria of legality are supposedly fixed by consensus, any pervasive
disagreement about their content would indicate the absence of consensus, and
hence the absence of a fact of the matter over which disagreement could be had.”41
Like Dworkin, Shapiro contends that disagreements over the criteria of legal
validity are pervasive. He refers to “the prevalence of disagreements about the
criteria of legality,” and to “pervasive, well-known and sincere disagreements
about the ultimate criteria of legal validity.”42 He reiterates his endorsement of
Dworkin’s critique of Hart: “As Dworkin pointed out, widespread disagreements
about the content of the rule of recognition are inconsistent with the consensus that
supposedly generates its content. Thus, if legal participants are neither hopelessly
confused about legal practice nor opportunistic liars, the criteria of legal validity
cannot be determined by judicial agreement about those very criteria.”43
This final main section of my chapter will first criticize an instance of hyperbolic
distortion in Shapiro’s characterization of Hart’s position. It will next submit that,
under Hart’s model of the Rule of Recognition, standards for legal validity can exist
41 Shapiro (2009: 248), footnote omitted, emphasis in original.

42 Shapiro (2009: 249 n. 50, 260).
43 Shapiro (2009: 260).
even though there is widespread disagreement about them. It will then very briefly
query Shapiro’s response to Dworkin’s critique. Finally, it will present my own
(largely Hartian) responses to Dworkin’s critique.
8.1. An instance of hyperbole

Occasionally, Shapiro indulges in distortive hyperbole when delineating Hart’s
account of the Rule of Recognition. Most notably, he contrasts his own Planning
Theory with Hart’s account along the following lines: “Like Hart’s rule of recogni-
tion, this [planning] norm is always at least partially constituted by official conver-
gence on certain standards of conduct; but in contrast to Hart’s account, total
convergence is not necessary.”44
This claim about total convergence is not borne out by anything in Hart’s
writings and is squarely at odds with what Hart remarked in Chapters 6 and 7 of
The Concept of Law and elsewhere. Moreover, it is not readily reconcilable with
what Shapiro himself affirms when he is outlining Hart’s theory of law: “On Hart’s
theory, the rules of recognition, change, and adjudication derive their content solely
from consensus. The rule of recognition in the United States, for example, validates
rules enacted by Congress and signed by the President that regulate interstate
commerce because most judges/officials take the internal point of view toward
such a test.”45 As is indicated by Shapiro’s use of the plurative quantifier “most,”
total convergence is not necessary for fixing the content of a criterion of legal
validity in the Rule of Recognition.
8.2. Criteria in the presence of disagreement

Hart was an Inclusive Legal Positivist, and he therefore accepted that a Rule of
Recognition can deem various correct principles of morality to be standards for
legal validity; under an Inclusive Rule of Recognition, consistency with each such
principle is a necessary condition for legal validity. If a Rule of Recognition is
genuinely Inclusive, then the referential intentions of the officials are such that the
principles which they have embraced as criteria of legal validity are correct prin-
ciples of morality rather than the principles which the officials believe to be correct
principles of morality.46 (Obviously, those two sets of principles can coincide.
Typically they will at least overlap.) Given as much, correct principles of morality
are validative criteria of legality in the relevant jurisdiction even though the officials
may well disagree frequently with one another in identifying those principles. It is
therefore not always the case that, within Hart’s model of the Rule of Recognition,
the absence of a broad consensus on certain criteria of legal validity is inconsistent
with the existence of those criteria as such.
44 Shapiro (2009: 267).

45 Shapiro (2009: 248), emphasis added.
46 I discuss this matter in chs. 2 and 4 of Kramer (2004).
8.3. A query about Shapiro’s response to Dworkin

When seeking to come to grips with Dworkin’s critique of positivism, Shapiro
assures his readers that his own Planning Theory resolves the difficulties that
supposedly bedevil Hart’s account of the Rule of Recognition:
The planning theory . . . does not privilege the views of the system’s original constitutional
designers. Because legal systems always contain mechanisms for revision, the constitutional
designers change as the structure of the system changes. The designers of the present
American system include not only the Framers and ratifiers of the Constitution of 1787,
but also the numerous agents that have changed the complexion of the system over the past
two hundred years. Moreover, the constitutional designers who are relevant for determining
interpretive methodology are those singled out by the present consensus in the legal
community. This present consensus determines which past consensus to heed. The idea,
once again, is that it is irrational for a group to treat a set of agents as designers whose role is
to resolve second-order uncertainty, and at the same time not to privilege their attitudes
about appropriate goals, values, and trust when trying to figure out how to interpret their
instructions.47
In two respects, this quoted passage is dubious.
First, having highlighted the quarrels among jurists over matters of interpretive
methodology, Shapiro then helps himself to the assumption that there is a present
consensus on the question of whose attitudes (concerning goals, values, and trust)
are determinative of interpretive methodology. Such an assumption is fanciful.
Second, the passage just quoted is not easily reconciled with what Shapiro says in
the very next paragraph of his essay: “The advantages of the planning theory,
I believe, are considerable. Chief among them is that, insofar as official consensus
is not necessary for the determination of interpretive methodology, the planning
theory is able to account for the possibility of disagreements about the ultimate
criteria of validity” (2009: 265). Though this second quoted passage is consistent
with the claim that a consensus on matters of interpretive methodology is not
necessary for the existence of criteria of legal validity, Shapiro has maintained in the
first of these two quoted passages that what is indeed necessary is a consensus on the
identity of the group whose attitudes are determinative of interpretive method-
ology. Untenable, then, is the second passage’s sweeping assertion that—according
to Shapiro’s Planning Theory—“official consensus is not necessary for the deter-
mination of interpretive methodology.”
8.4. A broadly Hartian response to Dworkin

A Hartian response to Dworkin should begin by mentioning a point which I have
substantiated at length elsewhere: namely, the fact that Dworkin himself acknow-
ledges that any viable legal system must partake of a substantial degree of settledness
47 Shapiro (2009: 265), emphases in original.

and regularity.48 He is well advised to acknowledge as much, since his account of

law would otherwise be fanciful. If in a certain country the implications of a very
large proportion of the norms promulgated by the prevailing system of governance
are indeterminate not only in controversial appellate cases but also in the system’s
everyday workings, then no full-fledged legal regime exists within that country.
Though any jural regime is bound to include a number of aspects that are unsettled
or open-ended, the sheer operativeness of any such regime as a jural system
presupposes a considerable measure of predictability and routineness. If intractable
controversy is typical rather than exceptional—that is, if the normative consequences
of people’s multitudinous actions under the ostensible laws of their country are
ordinarily (rather than occasionally) “up in the air” and truly murky—then “lawless-
ness” is the correct designation for such a state of affairs.
We should now re-evaluate the significance of Dworkin’s highlighting of overt
theoretical disagreements in hard cases. Let us concede arguendo that the criterial
divergences which become exposed in such cases are also present (subterraneously)
in easy cases. How damaging to legal positivism is such a concession, if it be a
concession? The short answer to this question is that Dworkin’s attentiveness to
hard cases should be welcomed as salutary rather than resisted as troublesome. His
work, set within its proper limits, can serve to refine the insights of positivism rather
than to undermine them.
Although some other critics of Dworkin have likewise sought to defuse his anti-
positivist barrages by accommodating them within positivism, this maneuver has
not been sufficiently connected to an emphasis on the routineness of the law’s
quotidian functioning. Once we apprehend the extent to which the ordinary
workings of a legal system must be regularized if the system is to be viable as a
regime of law—an extent which varies from system to system, but which is always
considerable—we can see that Dworkin’s attack on the Rule of Recognition is
untroubling for a Hartian positivist. Criterial divergences among judges and other
officials may indeed exist in straightforward cases as well as in difficult cases, but
they exist against a background of extensive commonality. The sorts of criterial
divergences which Dworkin recounts in his hard cases are important but fairly
narrow. For example, when judges disagree over the question whether the ordinary
meanings of statutory language or the underlying intentions of a legislature should
be deemed decisive, they are hardly disagreeing about the broader question whether
the legislature’s enactments lay down authoritative standards. Similarly, when
judges disagree about the extent to which various considerations of policy can
properly be invoked to distinguish precedents, they are hardly disagreeing about the
broader question whether precedents that cannot plausibly be distinguished are
generally binding.
Of course, the firm agreement on the most wide-ranging criteria comprised in
the Rule of Recognition would be hollow if judges continually clashed with one
another about the proper applications of those criteria. A consensus on abstract
48 See Kramer (1999: 140–2). My response to Dworkin in this subsection is a substantially
abridged and modified version of an array of arguments put forward in Kramer (1999: 135–46).
precepts is consistent with chaos at a practical level. However, the chief reason for
my drawing attention to the officials’ consensus on the more general standards in
the Rule of Recognition is that such a consensus is almost certainly necessary
(though not sufficient) for officials’ agreement at the level that matters most: the
level of concrete outcomes, the “bottom line.”49
Though the regularity that is practically essential for the very existence of a legal
system as such does pertain to the rationales for specific decisions, it pertains even
more importantly to those decisions themselves. If most jural officials in a regime R
disagreed with one another most of the time about the concrete legal implications
of people’s actions, their responses to those actions would be erratic and chaotic
rather than norm-governed. R would not be a regime of law, where behavior is
generally subsumed under the regulating and guiding sway of norms.
As has been remarked, Dworkin’s attempt to refute Hart’s model of the Rule of
Recognition is centered on theoretical disagreements in hard cases. Such disagree-
ments, Dworkin presumes, are lurkingly present in easy cases where they fail to
surface. We now can see, however, that Dworkin’s ruminations on theoretical
disagreements do no harm at all to the positivist conception of law. Though criterial
divergences may always be present in any jurisdiction, they cannot go beyond the
point where they would bring about substantial indeterminacy and erraticism in the
law at the level of concrete results. At any rate, they cannot go beyond that point if
the officials in any given jurisdiction are to maintain a functional legal system.
Widespread though the criterial divergences among the officials may be, the
practical impact of those divergences—their tendency to produce a welter of
discordant conclusions about the specific jural consequences of people’s behavior—
must be quite limited. If a legal system is to endure as such, the rivalry among judicial
perspectives will be cabined by the need for most officials to agree on the “bottom
line” in most circumstances.
Consider a situation in which the unimaginative judges in some jurisdiction
maintain that the plain wording of statutes is always dispositive in cases to which
the statutes are applicable, while the more imaginative judges maintain that the
underlying purposes of the legislature can take priority over the explicit wording of
statutes in exceptional cases. Dworkinians would probably claim that the imagina-
tive judges orient themselves toward the deeper intentions of the legislature in
routine cases as well as in knotty cases. What gives that claim its credibility is
the fact that it does not overlook or gainsay the routineness of the routine cases.
It acknowledges that the imaginative and unimaginative judges agree heartily on
the appropriate results in easy cases; more precisely, in regard to any such case, it
assimilates the specific substance of the legislature’s underlying intentions (on
which the imaginative judges rely) and the specific substance of ordinary linguistic
meanings (on which the unimaginative judges rely). By presuming that the general
intentions of the legislature for the implementation of a statute in any straightfor-
ward case will center on the normal import of the wording in the statute’s
49 Of course, I am not writing here about logical necessity. Rather, I am writing about what is
necessary in any credible circumstances. I am putting aside bare logical possibilities.

provisions, the Dworkinian view accepts that the imaginative judges and un-
imaginative judges concur firmly with one another in easy cases at the level of the
“bottom line.” The criterial divergences that separate the imaginative from the
unimaginative do not prevent a judicial consensus on the apposite concrete out-
comes of easy cases. Any abiding theoretical disagreements are structured in ways
that enable unanimity concerning the disposition of myriad cases. No colorable
theory of law could deny as much. Only because the Dworkinians’ approach
acknowledges the consensus among judges on the proper outcomes of most actual
and credibly possible cases, do Dworkinian claims about the constant presence of
criterial divergences enjoy any plausibility.
Theoretical disagreements among judges in any viable legal system are thus
circumscribed not only by the most wide-ranging standards in their Rule of
Recognition, but also by the pressures for regularity in the detailed effectuation
of the law. Moreover, these two forms of circumscription are closely linked. As was
remarked earlier, a state of virtual unanimity among judges and administrative
officials on the most general criteria in their Rule of Recognition is almost certainly
necessary for a very substantial measure of agreement among them on the concrete
applications of the law. Hence, given that a legal system is not sustainable as such
unless its officials are indeed in accord with one another to a considerable extent
about the law’s specific implications, it is likewise not sustainable unless the officials
are unanimous or virtually unanimous in their acceptance of the most wide-ranging
precepts in their Rule of Recognition. Criterial divergences among officials obtain
against the background of the officials’ unanimity or virtual unanimity on the
paramount criteria in their Rule of Recognition; and the intensity and range of the
divergences will be limited as far as is necessary for the preservation of a large degree
of regularity in the day-to-day administration of the law. Such will be true, that is, if
the system of governance containing the divergences is to endure as a functional
legal system.
Accordingly, Dworkin’s alertness to theoretical disagreements in hard cases can
be welcomed by legal positivists. On the one hand, his pointing out of such
disagreements is undoubtedly salutary, for it serves to counter the notion that the
officials who run an operative legal regime must be in harmony with one another
about all or nearly all the standards that compose their Rule of Recognition. Within
many legal systems, and certainly within the American and English legal systems,
there is quite a bit of room for disaccordance at the criterial level. On the other
hand, despite the importance of the criterial divergences (and despite their prom-
inence in law-school casebooks, which consist almost entirely of appellate opin-
ions), they are perfectly compatible with the healthy existence of a Rule of
Recognition. So long as the disunity remains within the confines that have been
sketched here, it does not impair the vitality of a legal system. Judges who differ
with one another about some of the criteria in their Rule of Recognition can and do
concur with one another about the most significant criteria therein and about the
practical legal consequences of most instances of conduct.
9. Conclusion
Save for some glancing remarks in section 8, this chapter has not grappled directly
with Shapiro’s Planning Theory of Law. Instead, it has sought to reveal that the
Planning Theory is at best superfluous. Though Shapiro endeavors to expose some
major weaknesses in Hart’s jurisprudential theory, nearly all of his objections are
misconceived and are often based on significant misrepresentations of Hart’s
ideas.50 Hart’s theory does need some refinements and modifications, but they
can be incorporated into his account without any wholesale departures therefrom.
When his account is recapitulated accurately, it proves to withstand Shapiro’s
broadsides. Legality is a thought-provoking book, and its author is a man of
admirable perspicacity, but its distortions and unfounded criticisms of Hart’s
jurisprudential insights are a setback for the philosophy of law.
References
Greenawalt, K. (1987). “The Rule of Recognition and the Constitution,” Michigan Law
Review, 85: 621–71.
Hart, H. L. A. (1965). “Lon L. Fuller, The Morality of Law,” Harvard Law Review, 78:
1281–96.
Hart, H. L. A. (1982). Essays on Bentham. Oxford: Oxford University Press.
Hart, H. L. A. (1983). “Kelsen’s Doctrine of the Unity of Law,” in Essays in Jurisprudence
and Philosophy. Oxford: Oxford University Press, 309–42.
Hart, H. L. A. (1994). The Concept of Law. Second edition. Oxford: Oxford University Press.
Kramer, M. (1999). In Defense of Legal Positivism. Oxford: Oxford University Press.
Kramer, M. (2003). The Quality of Freedom. Oxford: Oxford University Press.
Kramer, M. (2004). Where Law and Morality Meet. Oxford: Oxford University Press.
Kramer, M. (2007). Objectivity and the Rule of Law. Cambridge: Cambridge University
Press.
Kramer, M. (2009). Moral Realism as a Moral Doctrine. Oxford: Blackwell.
Raz, J. (1979). The Authority of Law. Oxford: Oxford University Press.
Raz, J. (1980). The Concept of a Legal System. Second edition. Oxford: Oxford University
Press.
Raz, J. (1999). Practical Reason and Norms. Oxford: Oxford University Press.
Shapiro, S. (2009). “What is the Rule of Recognition (and Does it Exist)?,” in M. Adler and
K. Himma (eds.), The Rule of Recognition and the U.S. Constitution. Oxford: Oxford
50 Shapiro also misrepresents the writings of John Austin. For example, without any citations he
declares: “Austin was fully aware that his theory of rules did not appear to fit power-conferring rules
very well” (2011: 62). In the only text by Austin to which Shapiro ever refers, The Province of
Jurisprudence Determined, there is in fact no inkling whatsoever of the problem posed by power-
conferring rules.
3
Law’s Authority is not a Claim
to Preemption*
Kenneth M. Ehrenberg
In the past, it was thought that if there was a general obligation to obey the law, it
would have to be content independent (applicable to law regardless of what it
demands), universal (applicable to everyone who is subject to it), and categorical
(comprehensively applicable to all duty-imposing laws).1 Given the difficulty in
establishing such an obligation, many more recent theorists have either given up
one or more of these facets in order to make it easier for law to succeed in
obligating, or they have refocused on the nature of the authority the law must
claim and when we might have good reason to accede to that claim. Joseph Raz, for
example, tells us that, by its nature, the law must claim moral authority where that
authority consists in the capacity to provide preemptory or exclusionary reasons for
action.2 More generally, moral or legal obligations are understood as consisting in
exclusionary reasons, which are first order reasons to comply with the content of the
directive or obligation, coupled with second order reasons to exclude certain other
reasons we might have from other sources. Among other things, this would mean
that the law is telling us to replace our own reasons against a directed action with
the reason that the action should be done simply because it is the law.
I am bothered with the idea that law tells us to replace our reasons. Of course, the
law is frequently telling us what to do. But this characterization of the law as claiming
to replace our reasons strikes me as making the law out to be more demanding than it
* For discussion, replies, and comments, particular thanks go to Guyora Binder, Mark Murphy,
Henry Richardson, Ekow Yankah, Stefan Sciaraffa, David Velleman, Danny Priel, Neil Williams, Ken
Shockley, David Braun, Matt Bedke, Arie Rosen, James Specyal, Cindy Phillips, as well as numerous
participants in the McMaster University Conference on the Nature of Law, the New Voices in Legal
Theory Roundtable, the Georgetown Law and Philosophy Discussion Group, the World Congress of
Philosophy of Law and Social Philosophy, and my “Topics in Legal Philosophy: Authority” graduate
seminar at the University at Buffalo.
1 Kramer (2005: 179–80). It is usually generally (and still) agreed that any obligation to obey the
law is only prima facie (or perhaps more correctly pro tanto, see Hurley (1989: 261); Edmundson
(2004: 215–16), and can be outweighed by more pressing concerns. That otherwise authoritative
directives can be defeated by pressing concerns should be distinguished from instances in which the law
itself allows pressing concerns to trump those directives as a kind of exception.
2 Raz (1979: 30).
52 Kenneth M. Ehrenberg
actually is. Even if, as Raz claims, the law is only asking us to put aside reasons to act
contrary to the directive in question3 (implying it is unconcerned with our reasons for
compliance), it attributes to law the claim that we ought never to act on contrary
reasons. I don’t believe this is an accurate characterization of the demands law
places upon us, although it may be an accurate characterization of our response to
its demands when we accede to its authority. Raz’s conception of authority and
preemptive reasons nicely captures our attitudes when we do accede to law’s demands
and obey simply because it is the law (and not out of fear of punishment or social
disapprobation). So the question becomes: can we understand the theoretical treat-
ment of law’s claims in a weaker way without damaging the strengths of Raz’s theory?
In this chapter I show that the law does not claim to preempt our reasons, although
we allow it do so when we accept its authority. While Raz holds the law’s claim to
preemptive authority to be a part of his service conception, I show that the rest of the
service conception does not require attributing this particular claim to law and that it
is enough for us to allow it to preempt conflicting reasons when we agree to its
demands. Although not fully developed here, the result of this will be that authorita-
tive directives can be understood as simple commands, albeit commands that are
generally to be legitimized, in accord with Raz’s theory, whenever compliance helps
the subject to conform better with the balance of reasons that already apply to her.4
This might seem a bit of a nit-picky pot shot at Raz’s theory. But the importance
of this clarification lies in a better theoretical reflection of our relationship to the law
and its putative practical authority. Under Raz’s current explanation, the law’s
ability to manipulate reasons that apply to us weighs heavily on the subject and
amounts to a demand that we might be very reluctant to agree to. Many replies and
criticisms of Raz stem at least partially from this reluctance.5 But rather than focus,
as others do, on seeing that reluctance as a reason to reject or alter Raz’s notion of
preemption, I want to focus on its role in Raz’s notion of law’s claim. If the law is
not understood to be making such a stark demand, then it will be more palatable to
allow for that manipulation of our reasons. Hence this attack on a small part of
Raz’s theory is actually meant as a way to deflect more serious criticism. Addition-
ally, if we can interpret law’s demands upon us in this slightly weaker way, then the
task of legitimizing the authority of legal directives might become just a bit less
arduous than Raz’s theory otherwise seems to entail.6
The argument proceeds in three parts. In the first part I present an overview of
Raz’s theory, showing what I take to be a misstep in attributing the claim of
preemption to the law and that Raz’s own exposition does not require that
3 Raz (2009: 140). 4 Raz (2006: 1018).

5 Hurd (1991); Alexander (1990); Moore (1989); Perry (1989); Regan (1989); Regan (1990);
Green (1988). Alexander (1990) is notable for presenting an argument that is somewhat similar in
denying that the law provides exclusionary reasons but leaving open the possibility that we should still
treat legal directives as exclusionary reasons. My argument differs in concluding that authoritative legal
directives can actually be exclusionary reasons when we accede to them, but that law does not claim
them to be.
6 For some of Raz’s reasons for making the legitimization of authoritative directives difficult see Raz
(1986: 57).
Law’s Authority is not a Claim to Preemption 53
attribution. This is only a half-argument in that it calls into question Raz’s reasons
for attributing the claim of preemption to law, but does not itself constitute a
reason for rejecting that attribution. In the second and third parts, I present two
independent but related reasons for rejecting the attribution of a claim of preemp-
tion to the law. One is the choice-of-evils (necessity) defense to a criminal
accusation. If the law allows (especially novel) claims of necessity in defense to a
criminal accusation, then it is cognizing individual reasons for contrary action as
trumping authoritative directives—even potentially when those individual reasons
were considered and rejected by a legislature. The other is the theoretical claim that
the law has gaps (a claim made by Raz and many others). If the law has gaps then
individuals must always use their own reasons when determining how to act in any
potentially gappy situation. If those reasons are the same ones possibly excluded by
a vague or conflicted authoritative directive, then the law cannot be understood to
be claiming to preempt them.
Granted, each of these two independent reasons is limited in itself. The first only
applies in legal systems that allow for choice of evil defenses. The second will only
be applicable against any theorist who accepts that the law necessarily has gaps. But
together, they cover quite a few bases. The first especially is useful against any
argument that the law must claim preemption as a conceptual matter. Since we
have an example of a legal system that cannot be understood to be making that
claim (and it is from our own systems), it would appear that the claim cannot be a
necessary characteristic of law.
1. Raz’s service conception and the half argument

from the failure of closure
Raz’s service conception of authority was novel for understanding that, in order to
be justifiable, authority had to serve the subject rather than vice versa.7 On the
other hand, legitimate authority is still understood as a right to rule, which the law,
by its nature, must claim.8
The theory consists of two theses, each of which independently implies a third.
The “dependence thesis” holds that, to be legitimate, authorities must base their
directives on reasons that already apply to the subjects of those directives.9 The
“normal justification thesis” (NJT) holds that authority is normally justified where
the subject does better at conforming to reason by following the directive than by
following her own understanding of the balance of reasons with regard to the
matter directed.10 (To these Raz has also added an “independence condition,” that
the matter being directed by the authority is not one on which it is better to decide
for oneself even at the risk of acting contrary to reason.11)
7 Raz (1986: 57). 8 Raz (1985: 6).

9 Raz (1986: 47). 10 Raz (2006: 1014); Raz (1986: 53).
11 Raz (2006: 1014); Raz (1989: 1180) (replying to Green (1989) at 810).
The dependence and the normal justification theses are said to imply a third
thesis: the “pre-emptive thesis,” that the fact of the authoritative directive is not
simply one reason to be added to the balancing of reasons in determining how to
act, but “should exclude and take the place of some of them.”12 Specifically, the
authoritative directive preempts “the reasons [against the directive] that the author-
ity was meant to consider in issuing its directives. . . .”13 The dependence thesis
implies the preemptive thesis in that the authority is already considering the reasons
that apply to the subject in issuing its directive. Not to see the directive as
preemptive would be to count the reasons behind the directed action twice: once
in the balancing the subject would have been doing on her own, and again in
importing the reason behind the authoritative directive.14 The normal justification
thesis implies the preemptive thesis in that, where the NJT applies, its success
requires that the subject replace the background reasons against the directed action.
The whole point of the NJT is that the subject does better at conforming to reason
by following the authoritative directive than by acting on her own. If the authori-
tative directive does not preempt the background reasons leading to her own
estimation of how to act, then the authority cannot do its job and get her to
conform better to reason.15
Notice, however, that these theses imply only that the subject must treat the
authoritative directive as preemptive, not that the law must claim to impose that
preemption. Most of the time, the language Raz uses in explaining the preemptive
thesis focuses on the subject’s perspective in treating the directive as preemptive.16
When we accede to the legitimacy of the authority and accept its right to issue
directives that are binding upon us, we agree that we are bound to obey its
commands. That bindingness may be understood in terms of reasons we have to
exclude our reasons that militate against compliance with the directive. But this
does not mean that the law is claiming to offer those exclusions, only to be justified
in seeking to control behavior.17 Hence, law’s claim to authority, the claim of the
right to rule, can be understood as a claim of the right to have its directives
followed. However, this is a claim to control behavior and need not be a claim to
penetrate that behavior to the reasons behind them. This point bears some
emphasis: There are many ways to control the behavior of others. While admittedly
most of them involve manipulating the subjects’ reasons for action, we can also
imagine dystopian scenarios in which subjects’ behavior is controlled by mind
control, and less dystopian scenarios in which behavior is controlled simply by
limiting the physical options open to the subject (which controls behavior through
12 Raz (1986: 46); Raz (2006: 1019).

13 Raz (2006: 1019).
14 Raz (1986: 58).
15 Raz (1986: 58); Raz (2006: 1019).
16 “The only proper way to acknowledge the arbitrator’s authority is to take it to be a reason for
action which replaces the reasons on the basis of which he was meant to decide” Raz (1985: 10)
(emphasis added); Raz (1986: 46) (discussing the preemptive thesis in terms of the subject not
“adding” the directive to the other reasons she already has); Raz (1990: 192–3).
17 This could be seen as a reason to agree with others who have said that the right to rule and the
duty to obey come apart. See e.g. Edmundson (1998), passim.

limitations placed on the choices open to the subject rather than through direct
manipulation of the reasons the subject has for making one choice or another).
While a directive or command is best understood as controlling behavior via the
manipulation of reasons (the command, when authoritative, serves as a reason), the
claim of authority is properly understood as a claim of the right to control behavior.
Since that control can take forms other than the manipulation of reasons, we
cannot immediately infer that a claim of the right to control behavior is a claim
of a right to manipulate reasons. Of course, the subject is the ultimate determinant
of her behavior (except in the mind control scenario) and so will view a directive she
considers legitimate to be a reason not to act on contrary reasons.
While Raz’s presentation of the preemptive thesis and his explanations of how
authoritative directives consist in providing exclusive reasons are generally framed
from the subject’s point of view, he occasionally suggests that the law is demanding
that we treat its directives as preemptive.18 It is this that I wish to deny.
The argument to the conclusion that the law is claiming preemption might seem
a simple matter of deductive reasoning: Authority consists in providing preemptive
reasons. Law must claim to have authority. Therefore, law must claim to provide
preemptive reasons. If I accept the two premises, it might seem that I am bound to
accept the conclusion.
However, this is mistaken because the argument rests upon a principle of closure
under entailment for claiming which is not supportable.19 If I claim that the glass
contains water, and water consists of hydrogen and oxygen, this does not mean that
I claim that the glass contains hydrogen and oxygen. Even if we can say that the law
necessarily claims authority, this does not mean that it must claim to offer
preemptive reasons, even if authority operates on us by providing preemptive
reasons.
Put another way, the nature of authority might consist in preemptive reasons
for the subject of authoritative directives, but this does not mean that one who
claims to have authority must demand that the subject preempt her (contrary)
reasons with the reason represented by the authoritative directive. The nature of
authority to the subject might be opaque to the claimant. Alternatively, even if the
authority is aware that the subject receives the directive as a reason to preempt her
contrary reasons, the authority need not be making that demand on the subject
simply because the authority is more concerned with action than with reasoning.
(This is a point Raz himself stresses.20) The legitimate authoritative directive must
be based on reasons that apply to the subject. But since the authority’s primary
concern is compliance, getting the subject to behave in conformity with the balance
of reasons that already apply, the authority is not generally concerned with the
subject’s reasoning after the directive has been issued. Yet the preemptive thesis
18 Raz (1990: 150–1). 19 Soames (1987).

20 In replying to H. L. A. Hart’s claim that the function of an authoritative directive is to supplant
the will of the subject Hart (1982: 253), quoted by Raz (1986: 39), Raz argued that authorities are not
generally concerned with thought or deliberation, and are rather usually concerned only with action:
Raz (1985: 7); Raz (1986: 39). See also Shapiro (2002: 406–7).
concerns the subject’s reasoning about how to behave in response to the authorita-
tive directive.21 Insofar as the preemptive thesis holds that subjects must take the
legitimately authoritative directive as a preempting reason, the thesis sets forth a
norm of reasoning, even if the preemption itself consists in simply refraining from
acting on the excluded reasons.
The argument from the failure of closure under entailment for claiming might
not satisfy those who criticize Raz for the notion that a legal system can make claims
at all. Of course, if we say that the law can make claims only as a metaphor or
“metaphysical fiction,”22 we would not be able to use the failure of closure under
entailment for claiming as an argument against seeing the law as claiming to have
preemptive authority. If there isn’t really any claiming going on, then any opacity in
the preemptive nature of authority to the claimant is irrelevant since there isn’t
really a claimant.23 But if we do see it as a metaphor for the “demands” (understood
in terms of our supposed obligations) made by a putative norm (or normative
system), that would still be to focus only on the subject’s perspective and how she
treats the authoritative directive. To the extent we can distinguish between the
demand made by a normative system and how the recipient of that demand treats it
when accepting it and acting upon it, what constitutes proper acquiescence to that
demand when accepting it can differ from what precisely is being demanded.
One place to see this is in Raz’s argument that the preemptive thesis is required
in order to prevent the double counting of reasons. If we have a rule and the rule is
seen as a reason for the action along with the reasons behind the rule, then those
reasons are being double counted.24 Preemption is therefore needed to prevent the
double counting of reasons by the person who performs the action, not necessarily
by the authority. As far as the authority is concerned, it can pile on the reasons
offered in the expectation that the subject will ultimately depend on one of them,
even if some of them are in reality duplicates of others, and even if the subject ends
up acting for some other reason than that he was directed to do so. Hence, it is fine
for a parent to tell the child not to touch the stove because he will get burned,
which will hurt, and because “I’m the mommy.” The child can then choose which
of these reasons seems more salient to him. Similarly, if one is confronted with a
legal authority who happens to be generous in providing reasons, perhaps a police
officer who tells you to detour because there is construction ahead, one can choose
to act to avoid the construction, or to avoid expected sanction, or to comply with
the justified authority. (In reality, our action is over-determined by these reasons
21 Raz’s “preemptive” reasons differ from Hart’s “peremptive” reasons in that Hart’s were explained
as reasons to exclude other reasons from deliberation. Hart (1982: 253). Raz’s preemptive or exclusion-
ary reasons are reasons not to act on the excluded contrary reasons. Raz (1986: 39). Yet they are still
reasons that go to the subjects’ process of balancing reasons. See Shapiro (2002: 406–7), rejecting this
distinction.
22 Himma (2001: 279).
23 Raz does believe that the notion of the law being able to make claims upon us must be more
robust than simply reducing such claims to those made by officials. This is necessary in order to
understand the demands of customary laws, which have force without official enactment. Raz
(1979: 29).
24 Raz (1985: 23).
and it is unlikely that a subject simply chooses one on which to act, instead acting
because the balance of all these reasons is in favor of compliance.) In complying
with the legitimate authority for the sake of that authority, we are using the
authoritative directive to preempt other reasons that apply to us.25 This does not
entail that the authority is asking us to do so. The authority does not generally care
what reasons we follow so long as we comply (which is why the directive preempts
only contrary reasons).
One objection that might be raised notes that the failure of closure argument
ignores the distinction between the semantic content of claims and their pragmat-
ics.26 This is the notion that the meaning of a given utterance depends partially
upon the use to which the utterance is put and can come apart from the semantic
content of that utterance. This objector might say that “The law claims to preempt
contrary reasons” is an accurate account of what law claims but not how it makes
that claim. If these come apart then it would be possible to say that the failure of
closure holds for the contextual pragmatics of the claim but not the semantic
content, “and sometimes the semantic content of a sentence is not itself asserted, or
even included in what the speaker is committed to.”27
There are a number of difficulties with this objection stemming again from the
fact that the law is not a person and its claim is not a verbal act. For one, the
distinction between semantics and pragmatics is about utterances (understood as
“specific events, the intentional acts of speakers at times and places”28), and while it
might make sense to attribute to law the claim to have or to be a practical authority,
that does not mean we understand it to utter that claim. Since law is not the kind of
thing that can make an utterance, it does not make sense to say that we can
distinguish the semantics and pragmatics of its claims.29
On the other hand, one might note that where there is no speaker, it doesn’t
make sense to point out the failure of closure either; if the claiming is itself only
metaphorical and not reducible to a person’s utterance, then perhaps the law
“claims” everything that is entailed by its “claim.” Not only can’t the law’s claims
be reduced to particular official utterances (since there can be law without official
enactment), they also cannot be reduced to the wording of particular laws. Raz says
that the law’s claim to have or to be a practical authority is a conceptual feature of
law wherever it is encountered.30 In that, it is a property of legal systems rather than
25 Raz (1985: 23) (showing that we must choose between the authority and the underlying reasons
to determine our actions).

26 I am indebted to Matt Bedke for this objection.
27 Soames (2008: 404). Notice that this qualification actually helps my argument since once we
focus on the pragmatics rather than the semantic content of the claim, we are in a position to say that
the law’s claim is not supposed to include the claim of preemption when understood as the semantic
content of its claim to have practical authority.
28 Korta and Perry (2011).
29 To be more precise, we might understand the law to perform utterances in the wording of written
statutes and court decisions, and in what comes out of the mouths of officials. While for any of these
there might be a distinction between the semantic content and pragmatics of the utterance, the claim to
authority is attributed to the law as an institutional whole (and as a conceptual truth) and not to its
individual parts. To that extent, this particular claim cannot be seen as an utterance.
30 Raz (1986: 76–7).
individual laws. (We might go so far as to say that it is an emergent property of legal
systems in their purported supremacy and comprehensiveness31 that supervenes
upon the individual legal norms.) Maybe we can do without the “claiming”
language, but probably only by using other terminology that still has some
metaphorical aspect: the “demands” the law places upon us, the norms it “purports”
to impose upon us, etc. We are subjects of those norms and we are anthropomorph-
izing the law a bit in order to impute to it a kind of agency in its impositions of
norms. Perhaps the most precise way to speak about this would be to say that
“society purports to impose these obligations upon us through the institution of
law,” since we might be more comfortable attributing that bit of agency to a
collection of people than to an abstract social institution. But whether we attribute
the claim of authority to law, to society, or to government, we cannot simply leap
from the claim of authority to the claim of preemption.
More importantly, however, it is precisely with the semantics of the claim that
we are concerned. In sections 2 and 3 I will advance arguments that it cannot make
sense to attribute the claim of preemption to the law because it contradicts other
important facets of law. That contradiction goes to the semantic content of the
claim we are attributing to the law and not the way in which it is made.
This discussion underscores the importance of the perspective from which we
analyze the claims law makes. Another possible reply turns Raz’s reliance on the
subjects’ perspective into an attack on my basic position. The reply reminds us that
Raz is always looking at authority from the standpoint of the subject rather than the
claimant. Hence the attribution to the law of the claim to exclude the subjects’
contrary reasons is an attribution made from the subjects’ point of view (as is
underscored by my discussion of our anthropomorphizing the law by attributing
claims to it). Since I am saying the subject allows the law to preempt reasons by
acceding to the law’s claim to authority, when we look at the law’s claim from the
subject’s perspective it still makes sense to attribute the claim of preemption to the
law. That is, the claim to authority itself is really just attributed to the law by its
subjects, so if the subjects accede to the legitimacy of that authority by allowing it to
preempt their contrary reasons, then it still makes sense for them to attribute that
claim of preemption to the law.
This is a powerful reply and not to be rejected lightly. But I think that my
following arguments will show the weakness of this position. Even if we understand
the nature of law’s authority from the standpoint of the subject, our theory of that
authority should present a coherent picture of what law must be demanding of us
based on the way it operates. If other facets of the law (available to its subjects)
make a demand to preemption implausible as in tension with those other facets,
then it would be far better for a theory of its authority (analyzed from the
standpoint of its subjects) to avoid that attribution of a demand to preemption.32
31 Raz (1990: 150–1).

32 Of course many, if not all, legal systems frequently make conflicting demands upon their
subjects. But a theory of law’s authority is not a given legal demand that can come into conflict with
others. Rather, a theory of law’s authority such as Raz’s is supposed to be setting forth conceptual truths
While this section shows that Raz’s argument need not reach the conclusion that
the law claims to preempt our non-legal background countervailing reasons when it
makes its necessary claims to authority, I have not yet shown a good reason to deny
that it does so. In the next two sections I tackle that task.
2. Choice of evil defenses

That the law does not claim to exclude our non-legal reasons for action is at least
true in legal systems (such as ours) that cognize choice-of-evils (also called “neces-
sity”) as a justification defense to some crimes. However, if it is true in some legal
systems, then the claim of preemption is not a necessary component of law’s claim
to legitimate authority.
In order to understand how the choice-of-evils defense precludes law’s claim to
preemption, I need a distinction from the literature on criminal law. There is some
controversy over the boundary of this distinction, but it will not affect the use to
which I put it. The law generally allows both justification and excuse as defenses to
crimes. An excuse is offered in defense when the accused admits committing the
action that fits the external definition of the crime, but argues that she was not
responsible for the action because of circumstances that prevented her from
avoiding it. In general (and admitting there are exceptions to this way of character-
izing the distinction), if one were to view a situation that would later be grounds for
excuse from the crime, one would still have good reason to intervene to prevent the
action if possible. This includes as examples cases such as crimes of passion (jealous
lover situations), perhaps individuals assaulting or killing their habitual abusers
when not under direct threat by them, perhaps acts done under certain kinds of
insanity (to the extent that the applicability of the definition of the crime itself is
not vitiated).
Justification (of which I take choice-of-evils to be an example) is different. If one
were to view an action that conforms to the external elements of a crime in a
situation that would later be grounds for a justification defense, one would not have
a good reason to intervene to prevent the action. For example, if someone witnessed
a fire that threatened the life of someone trapped inside a building, one would be
justified in running into the adjacent fire extinguisher store and appropriating a
fire extinguisher to use in extinguishing the fire. If the only way to separate two
people locked in mortal combat was to render each of them unconscious, one
would be justified in doing so even though it has all of the external characteristics of
engaging in battery oneself to do so. In either of these cases, another person would
not be justified in stepping in to prevent the action that usually constitutes a crime.
This is important for underscoring the fact that, in justification situations, the
accused is asserting that his reasons for non-compliance were enough to trump the
about the nature of law itself. If we believe that the institution of law is not incoherent at its core, then
any successful theory of its authority must be in harmony with other common or core facets of its
operation.
authoritative directive. What’s more, in systems where the law countenances this
defense, the law is thereby allowing its authority to be trumped by the subjects’
countervailing reasons. Since law necessarily claims that all valid laws are authorita-
tive, then to say with Raz that the law claims preemption is to say that it claims we
ought never to act on contrary reasons. But in systems that specifically allow
subjects to act on contrary reasons, the law cannot be seen as claiming to exclude
those reasons and the claim to preemption cannot be understood as a conceptually
necessary feature of law.
The State of Illinois defined the choice-of-evils defense as follows:
Conduct which would otherwise be an offense is justifiable by reason of necessity if the
accused was without blame in occasioning or developing the situation and reasonably
believed such conduct was necessary to avoid a public or private injury greater than the
injury which might reasonably result from his own conduct.33
The decision for a court on a case of first impression for a novel use of the necessity
defense is then whether the accused’s belief was reasonable. The court is called
upon to decide, in effect, if it was reasonable for the accused to believe that the test
of the normal justification thesis failed in that instance and the accused would do
better to look to his own devices in balancing reasons. But if the law (in the person
of the court) is called upon to decide that, then either the law cannot be claiming
authority as legitimated by the NJT, or it must not be making claims to preemp-
tion. Accepting the latter option is more palatable if we wish to maintain the
advantages of the service conception.
Choice-of-evils involves using reasons that are usually moral, perhaps personal,
and generally non-legal (at least on first impression) to defend against conviction for
the commission of a crime. If the law recognizes that justification as a defense to a
given crime, then it is recognizing that one’s own reasons can trump, and are
therefore not excluded by, law’s supposed claim to exclusionary authority. If the law
holds out the possibility of novel justifications being successful defenses, then the
law carves out a space for trumping reasons that must be non-legal when used since
there is no description of the content of those reasons within the law. (This is why
novel applications of the defense are particularly important for our purposes—the
reasons behind a novel defense are not yet part of the law.) Hence, any classes of
reasons that are allowed to serve as novel justification defenses to criminal prosecu-
tion would not be excluded by law’s putative claims to provide exclusionary
reasons.
We should not conclude from this that law’s claims of preemption are simply
unsuccessful in these cases. Nor should we see this as a situation where we do not
accede to law’s preemptive claims. Rather, since the law itself is permitting (and
even encouraging) action for these kinds of reasons, the law cannot be claiming
to exclude them at all. Hence, legal authority cannot be claiming to exclude such
33 Ill. Rev. Stat. 1971, ch. 38, par. 7–13 as quoted in People v. Unger 66 Ill. 2d 333, 341 (1977), a
commonly taught case in criminal law textbooks, with thanks to Guyora Binder for pointing this and
the subsequent case out to me.
non-legal reasons. More importantly, if the nature of the law allows for these kinds
of justifications (admittedly, they may not be present in every legal system), then it
is difficult to see how legal authority itself rests on a claim to preclude reasons that
might be cited later in choice-of-evils defenses.
This argument from choice-of-evils is somewhat limited in that it depends
primarily on the availability of cases of first impression for the particular necessity
defense. Once the particular defense is presented successfully, then the class of
reasons characterized by that justification arguably becomes legal, perhaps as a type
of exception to the crime. Prior to the first acceptance of a particular claim of
choice-of-evils, however, the class of reasons serving as a justification is non-legal
unless they are found elsewhere in the law (in which case they would not really
be first impressions anyway). Thankfully, Illinois and many other jurisdictions
that countenance necessity defenses define them in a way that leaves them open to
novel uses.
A first objection to this analysis comes from Raz himself: such justifications are
not limits on legal authority.34 Rather, they are all to be interpreted as exceptions to
the crime against which one uses the justification to defend. Raz specifically
countenances necessity defenses as a doctrine “designed to allow exceptions to
legal requirements. . . . The point is that the law demands the right to define the
permissible exceptions.”35
There are several ways to resist this move, however. First of all, it strains the
imagination to say that choice-of-evils reasons are already exceptions built into
the definition of the crime in cases of first impression. That is, the first time the
particular justification is successfully used in court, it does not make sense to say
that it was already carved out of the crime before the precedent was set (especially
not for a legal positivist like Raz). Furthermore, there are two ways of interpreting
this reply, both of which still preclude seeing the law as claiming preemption. One
might take this claim to mean that the person still broke the rule represented by
the law, but that she was justified in doing so. Or one might take the claim to mean
that the justification is an exception that is not yet written into the rule; so that the
person with the justification does not break the rule at all. Putting aside the
Wittgensteinian point that this might be a distinction without a difference, both
still cede authority to the individual actor to discover the choice-of-evils exception
on her own, as circumstances arise. Both allow the individual to act for her own
reasons in contravention of the law as written and previously applied. And this
allowance is itself a part of the law. So the law cannot be telling us that we should
not act for those reasons.
While it is true that the law is reserving to itself the right to define the permissible
exceptions (which Raz sees as a reason to underscore its authority on the matter),
the fact that it does so after the fact in cases of first impression means that it cannot
be asking its subjects to exclude their contrary reasons before the exception is legally
cognized. While the effect of such legal recognition is to carve out that exception
34 I thank Jason Paget for pointing this out to me. 35 Raz (1986: 77).
retroactively, since the point of authority is to guide and coordinate behavior, the
otherwise authoritative directive is not performing that function for the subject
with the ultimately successful novel necessity defense.36 One might be tempted to
say that the court is simply recognizing that the directive was not legitimate for the
subject with a successful novel necessity defense. But that cannot be right because
by carving out the exception, the court is maintaining the authority of the directive.
(Under Raz’s theory a court can invalidate the legality of a directive, but it is not
clear how it could agree to the illegitimacy of the directive’s authoritativeness
without invalidating it since the law must claim authority. The legitimacy of a
directive’s authority is a moral, not a legal notion.) The result is that it makes more
sense to attribute to the law a simple directive, about which it can later recognize
retroactive exceptions, than the claim: “Exclude contrary reasons for action except
any YOU determine too important to exclude, about which we will later decide
whether to exempt.”
Another objection might remind us that any duty to obey the law is always pro
tanto. We can still see the obligation to obey as a moral obligation and interpret that
obligation as an exclusionary reason since moral obligations always yield to more
pressing moral obligations. Hence to claim that the legal obligation is a moral
obligation (which is implied by law’s necessary claim to legitimate authority) is not
say that it must trump other moral obligations. Since choice-of-evils defenses are
generally made on the basis of moral obligations, the law is simply making space for
the possibility that the obligations it claims to impose are trumped by more pressing
moral obligations. In a sense, these defenses are there simply to exculpate when the
pro tanto conditions are not met.
There are two lines of response to this objection, each dispositive in itself. The
first is that legal definitions of such justification defenses do not usually require
them to be based on moral obligation. This is apparent from the consequentialist
language used in Illinois’s definition, as well as the provision for private harms.
While using the word “evil” might make the justification sound moralistic, the test
is judged in terms of consequences. So, unless one is a strict consequentialist who
believes that the meaning of moral duty itself is to promote the best consequences,
one can imagine situations in which one can make use of a choice-of-evils defense
on the basis of providing for the best consequences without having to claim that
one fulfilled a moral duty in doing so. If it is possible, for example, to offer a
successful justification based on following an entirely prudential reason over a legal
duty (as suggested by the Illinois statute), then it is difficult to see how the legal
duty could still be claimed to be exclusionary since that prudential reason is
precisely one reason it would be claiming to exclude (by kind). That is, if one
can use some self-interested reasons to trump legal duties, and we don’t regard
those self-interested reasons as moral reasons, then non-moral reasons can be used
to trump legal obligations and therefore those obligations cannot be understood to
be a demand to exclude those non-moral reasons. Furthermore, as long as there is
36 This point will be developed further in section 3.

any possible legal system that allows justifications which are not based on more
pressing moral duties (certainly no stretch of the imagination considering it applies
to our own under some very common moral theories), then at least those systems
do not make comprehensive claims to provide exclusionary reasons. As long as there
are possible legal systems that do so, we cannot say that it is a conceptual truth
about the law that it must claim to provide exclusionary reasons.
A second line of defense against this objection reminds us that the law does not
present itself with these exceptions built in. When one’s actions fit the external
elements of the crime, one must affirmatively present the justification to defeat the
application of the law and avoid conviction. To say that the law does not claim to
be offering exclusionary reasons still allows the law to claim to be absolute, and even
for it to claim to impose a comprehensive and content-independent demand (so
long as that demand is not understood as a claim to provide exclusionary reasons).
Choice-of-evils defenses are offered as exculpating only after the fact of the action.
The law’s demands may be absolute; but it is enough to constitute our agreement or
acquiescence to those demands if we admit them only pro tanto. The presence of the
choice-of-evils defense is not understood to be guiding behavior in itself and on its
own. Indeed, we have to engage in the otherwise prohibited behavior, and risk the
failure of our contemplated justifications, in order to reach the point where they can
be invoked. Hence we should not see the pro tanto character of any moral obligation
to obey the law to be what is provided for by defenses of justification.
Another important objection comes from Raz’s language in explaining his service
conception of authority. “[A]uthoritative directives preempt those reasons against
the conduct they require that the authority was meant to take into account in deciding
to issue its directives.”37 Hence legal directives are not meant to exclude all non-
legal reasons for action, just those that weigh against the legal directive and that the
legal authority “was meant to take into account” when fashioning legal policy.
Choice-of-evils defenses, this objector might claim, capture exactly those reasons
that the legal policy makers could not have taken into account, allowing them to
exculpate non-compliance. Therefore legal directives can still purport to be offering
exclusionary reasons because they function as second order reasons to exclude only
first order reasons of the kind that legal policy makers meant to exclude. If one
comes up with a novel reason for action that can serve as a justification, then the
policy makers could not have included that reason in their deliberation and it could
not have been meant to have been excluded by the legal directive.
This is a strong objection but not an insurmountable one. First of all, for this to
be a correct interpretation of Raz, we would be weakening considerably the
possibility of a successfully binding authority. This comment of Raz’s is offered
in explaining his Normal Justification Thesis, that we are justified in acceding to
the claims of authority when we do better at complying with reasons that already
apply to us by following the authoritative directives than by deducing what the
balance of reasons require on our own. If the reasons that are to be excluded by the
37 Raz (2006: 1018), emphasis added.

authority are to be limited to those that they already were meant to have considered
in issuing the directive, then we would be in a position of always having to check
whether the reasons we are considering against complying with the directive are
ones that the authority was meant to have considered and rejected. This would
seem to lead us back out to considering the advisability of each action on its own
and rejecting the claims of authority (complying with the directives only when we
had our own reasons for doing so). It undermines the whole point of authority,
since checking whether they are of the type meant to be rejected by the authority
requires us to make an assessment of their advisability on our own.
Since Raz believes that the claims of authority must in principle be capable of
success, this cannot be a correct interpretation and the reasons that the authority is
purporting to exclude must be much wider in scope. A footnote on the page
following the text quoted above seems to support this broader interpretation:
[T]here are two kinds of reasons the preemption thesis affects: First, it preempts reasons
against the conduct required by the authoritative directive. Second, it preempts reasons that
do not necessarily bear on the pros and cons of behaving as the directive requires, but that do
militate against the desirability of issuing the directive. These may be that the matter should
be left to individual discretion, or that the directive will have undesirable side effects that
make it undesirable, and so on.38
It appears that the reasons the authority was “meant to consider” are grouped by
kind. It would do no good for Raz’s theory to say that every reason must be
considered and rejected by the authoritative legislator in minute detail. They are
considering reasons under general descriptions. Authoritative legislators are meant
to consider broadly described situations that would lead to successful justifications.
If they choose to reject them by not including them as an explicit exception to a
legal directive, then they are implicitly leaving it up to the courts to carve out
exceptions on a piecemeal basis (a facet to which we will return shortly).
A valuable concurrence in a 1979 Vermont Supreme Court opinion39 makes
precisely the Razian point,40 claiming that choice-of-evils defenses are not available
when the reasoning done by the defendant was considered and rejected by the
legislature. Justice Hill argued that the defendants had no recourse to the necessity
defense for their actions trespassing in protest at a nuclear power plant where the
legislature had specifically determined that the benefits of nuclear power outweigh
the risks.41 If the reasons relied on by the defendant were not already considered
and rejected by the legislature (but were still of the type meant for consideration in
their policy determinations), then he would be free to rely upon them in disobeying
the law. However, Hill was concurring in the result and taking the majority to task
for redoing the legislature’s weighing of reasons for and against nuclear power by
38 Raz (2006: 1018 n. 19).

39 State v. Warshow, 138 Vt. 22, 28 (1979), Hill, concurring.
40 “They exclude reliance on conflicting reasons, not all conflicting reasons, but those that the
lawmaker was meant to consider before issuing the directive.” Raz (2006: 1022).
41 There is a modicum of irony in the use of this case since substantially the same fact pattern was
used by Raz to support a kind of duty to obey the law. Raz (1984: 146).
deciding even to entertain the defense. Hence, we can imagine a majority response
to Hill that defends the court’s decision to entertain the defense and re-weigh those
reasons in deciding if the defendants acted reasonably. But that would be to say that
even the reasons supposedly considered and rejected by the policy maker in issuing
the authoritative directive can be used in a choice-of-evils defense. Hence, the law
cannot be said to require their exclusion as a conceptual matter.
Even though the majority did not find acting on those reasons to be reasonable,
entertaining the defense meant that the court deemed it proper for the defendants
to consider acting on those reasons even though the legislature had already rejected
them. While allowing the mere consideration of these reasons does nothing to call
Raz’s analysis into question,42 the possibility of the court finding that acting on
reasons explicitly rejected by the legislature could have been reasonable (which is
entailed by entertaining the defense) means the majority does not understand the
law to be claiming to exclude those rejected reasons.
It seems much more correct to say that the legislature chooses to remain silent or
vague on the reasons they considered possible exceptions precisely to leave open-
ended the possibility of novel justifications without undermining the authority of
the directive itself. If we consider the fire-extinguisher theft mentioned above, it is
very likely that the legislators imagined such a scenario and chose not to include it
in the wording of the law because there is already a provision for crafting necessity
defenses. A legislature’s silence is not merely the result of an inability to consider all
possible justifiable exceptions; it is a deliberate openness of the law stemming from
a likely legislative desire that violators take the directive seriously when deciding to
accept a risk that a novel affirmative defense will not be successful. Hence we
cannot simply say that the reasons undergirding any novel defense was not of the
type the legislature was meant to have considered. We can imagine them specifically
considering these reasons and still deciding not to include them as exceptions.
One wrinkle in this analysis is Raz’s observation that the supposedly excluded
reasons do not include reasons in favor of the action that the law commands.43
Hence, the law need not be saying to us that we must follow it simply because it is
the law in order to comply with it. It is permitted to follow the law for our own
reasons precisely because those personal reasons that still militate in favor of
conformity are not excluded (although in doing so, we are not accepting the legal
norms internally). Nevertheless, this point carries no weight against my claim that
novel choice-of-evil defenses preclude the conceptual claim that law must be
claiming to provide exclusionary reasons. Since the choice-of-evil defenses are
offered precisely in cases where agents do not comply with the law, and they are
offering their own reasons for non-compliance, the fact that Raz’s theory does not
have law exclude reasons in favor of compliance is immaterial to this argument.
Another twist on this objection comes from another qualification Raz makes.
Raz notes that “to fulfill its function, the legitimacy of an authority must be
42 See nn. 20 and 21.

43 Raz (1998: 17 n. 39) and accompanying text; Raz (2009: 144). This point was suggested to me
in conversation with David Velleman.
knowable to its subjects.”44 One might say that in a case where the subject has a
good novel choice-of-evils defense, the authority of the directive for that subject at
that time is not knowable since the subject in such situations would not be able to
access whether the directive is helping her to conform better to the balance of
reasons. But this cannot be right since it is clear that in deciding to act contrary to
the directive, the subject is deciding that the directive is not authoritative for her at
that time. So perhaps it is more correct to say that the lack of the directive’s
authority is knowable and the court is simply empowered to recognize the correct-
ness of the subject’s decision that the directive is not authoritative. However, as
noted above, the effect of the court’s decision to validate the novel choice-of-evils
defense is to include it as a new legal exception to the directive, not to recognize the
directive’s lack of authority for that subject at that time. It would make no sense
under Raz’s theory to contemplate a court denying the authoritativeness of a
directive without deeming it invalid as that would violate law’s claim that all its
directives are authoritative.45 And lest the objector reply that the court carving out
the exception is rendering the directive legally invalid as against that subject at that
time, consider that the court’s recognition of the exception does nothing to change
the directive’s membership in the set of valid laws for that jurisdiction. A successful
claim of necessity does not carry the recognition of any change in the directive’s
validity, only its applicability. That change in applicability was brought about by a
subject deciding correctly that certain reasons contrary to the directive were not to
be excluded by the directive. The best way to make sense of the legal provision for
such a subject-determined change is not to attribute the claim to exclude those
reasons to the law in the first place.
The most worrisome objection stems from Raz’s observation that a permission is
still an exercise of authority.46 In essence, where a right is granted in law, or an
exception is carved out (by legislature or by court), the law is refraining from
exercising its authority in that carved out area. But since it is the law that
determines the content and limits of those permissions and rights, the law is still
claiming preemption in those areas in which it chooses to issue directives. This is an
important reminder for enumerated exceptions and rights, but I do not think it can
win the day where novel justifications are concerned. The possibility of novel
justifications must be understood (at least within Razian positivism) to be excep-
tions without content until a decision is made by a legal body, an issue to which we
return in section 3. If the function of authority is to get individuals to comply better
with reason than they would do if left to their own devices, it does not make sense
for it to claim on the one hand to preempt our contrary reasons and at the same
time leave open catch-all exceptions that we must use our own devices to fill. One
might note that it is still the law that remains the final arbiter of whether the newly
raised justification is successful. This may be true, but at the moment of action
(before the law makes its determination), in choosing the lesser of two evils in a
novel situation, the subject is either rejecting guidance or acting under no guidance.
44 Raz (2006: 1025). 45 Raz (1986: 76–7). 46 Raz (1990: 151).

To say that the law claims to preempt our contrary reasons but leaves open catch-all
exceptions is to say that it provides no guidance whenever we find ourselves in a
lesser-of-two evils situation. It seems to me to make more sense to say that before an
exception is carved out, the law does provide guidance and allows us to reject that
guidance in such situations. That is to say that it is not claiming to preempt our
contrary reasons, since it is still allowing us to act upon them.
To be clear: choice-of-evils is not itself a reason to reject the service conception.
Raz could just respond that when we are in a choice-of-evils situation, we are clearly
in a situation where the authoritative directive isn’t passing the NJT (i.e. isn’t
justified). The law is, in essence, incorporating this realization into its own claim of
authority. Rather, choice-of-evils is a good reason to think that the law cannot be
claiming preemption since it is itself carving out a way in which the subject is legally
permitted to decide when the NJT fails. To be fair, Raz confronts the idea of such
justifications.47 But he doesn’t seem to appreciate the impact this facet of the law
has on his notions of what law must claim.
However, the strong objection that the availability of novel necessity defenses
amounts to a catch-all exception over which the law still exercises authority by
determining legality after the act raises the issue of legal gaps. Until a decision is
made by an authoritative body, there is not yet a fact of the matter whether a novel
justification is a defense to the crime for which it was offered. It might be thought
that those gaps can somehow provide room for law’s claim to preempt our contrary
reasons, so it is to those gaps we now turn.
3. Gaps
Raz and many other theorists who say that the law claims exclusionary authority
also say that the law is a matter of social fact, or even that the law is a social
convention. These claims are important in allowing us to see the truth conditions
for claims about what the law requires. For example, it is true that the speed limit
on the New York State Thruway is 65 m.p.h. For this to be the kind of thing that is
capable of having a truth value, there must be certain social conditions that must be
present to make it true. Many such theorists also say that because law is a social fact
the law has gaps: situations in which there is no fact about what it requires, permits,
or proscribes.48
The gappiness of the law would hinder the law’s ability to be successful in its
supposed claims to authority if those authority claims are interpreted as purporting
to exclude our non-legal reasons. Alternatively, the gappiness of the law precludes
the theoretical conclusion that the law is claiming to exclude our personal reasons.
47 Raz (2006: 1026): “legal systems typically allow some [countervailing reasons] to count and
sometimes to override legal requirements.”

48 Raz (1979: 70–7) (arguing for the existence of gaps in the law where the law is indeterminate or
there are unresolved conflicts).

If the law is gappy, then its claims are similarly circumscribed and it is not claiming
to exclude non-legal reasons at all.
To say that the law is gappy is to say that the domain to which it (now) applies is
somehow limited (although as Raz notes, it also claims the power to legislate on any
subject49). This is clear from an examination of the opposing claim that the law
does not have any gaps. Those who claim it does not have gaps say that every legal
proposition is already true or false; every action is already legally prohibited,
permitted, or required.50 If the law does not give a direction regarding an act,
then the act is legally permitted (in a system that includes a closure rule permitting
any action not expressly prohibited—other systems could theoretically prohibit any
action not expressly permitted). A new law simply changes the legal status of the
act. Under this view, the domain of law with regard to actions is unlimited and
everything is already within its ambit.
To deny this is to say that there are limits to what the law determines at the
moment. A given limitation is contingent in the sense that the law could expand to
include reasons heretofore ignored. Nevertheless, as long as there are gaps, there are
situations to which the law does not apply. This is not necessarily the result of new
situations. Rather, gaps tend to arise as a result of vague or borderline cases the
application of the law to which has not yet been made firm, or as the result of
conflicts in the law or in its application. In the time-worn example of “no vehicles
in the park,” those who believe the law is gappy claim that there is a gap before the
appropriate legal official makes a decision about whether this rule applies to
bicycles. This is because there is no legal fact of the matter of whether the rule
applies to bicycles before the correct legal official makes a decision, thereby
declaring that the law does or does not apply to bicycles.
Most who claim that the law has gaps do not claim that this is a contingent
feature of the law, but rather it is in the nature of the law that it cannot speak to
every possible situation since it rests on social facts and/or human conventions. If
there is not a fact or convention in place yet for a given set of circumstances, then
the law cannot yet cover those circumstances. Hence it is in the very nature of the
law that it speaks to a limited set of issues. Even if the system has a closure principle,
that principle only applies where the law is silent, not where it speaks in vague or
contradictory terms.
Regarding those actions about which legal propositions would be indeterminate,
i.e. within those gaps, the law cannot even in principle provide any direction. If it is
not providing direction, it certainly is not claiming to exclude any reasons that an
individual might have for or against actions within those gaps.
There is an easy objection here: if the law is guiding behavior, it is only doing so
when and where it is offering directives. If the nature of those putatively authorita-
tive directives is to claim to exclude individuals’ non-legal reasons for contrary
action, then that exclusion is only meant to apply where the directives apply. The
law might not exclude reasons in the gaps, but it still can exclude reasons in those
49 Raz (1990: 150–1). 50 See e.g. Dworkin (1991).

areas in which it does speak and that can still be the nature of legal authority.
More importantly, it can still claim to exclude reasons for contrary action wherever
it does apply.
The problem with this objection is that it would jeopardize the content-inde-
pendent criterion for legal authority. Law is supposed to be capable of obligating
you simply by the source of that obligation, rather than because of what it is telling
you to do.51 More to the point for Raz, it would jeopardize the function of
authority to guide behavior in compliance with right reason. If the law has gaps,
then the subject must always examine her situation and the contemplated behavior
against the content of possibly nearby legal norms in order to determine whether
she is in a gap. At the moment of action, it is always up to her to decide whether she
is in a gap, using her own reasoning to determine whether the law applies to her.
This undermines the point of authority under Raz’s service conception. Some
might see this as a reason to reject the claim that the law has gaps at all. However,
one could just as easily see this as a reason to reject the idea that the law is claiming
to provide exclusionary reasons.
This is not to say that every contemplated action might be in a gap. If the law is
clearly speaking to one’s situation and all the reasons are straightforward ones, then
one treats the unproblematic law’s legitimately authoritative guidance as excluding
any contrary reasons. But the fact that one might find oneself in a gap, and needs to
look to one’s own devices in making both the determination of whether one is in a
gap and how to act within, means that the law cannot be claiming to exclude
reasons (contrary or otherwise) since that need to look to one’s devices is a part of
the nature of law.
To be clear: the problem does not arise because of gaps themselves, the problem
arises from gaps in combination with the notion that the law claims to exclude
contrary reasons. If it is not claiming to exclude contrary reasons, then the
individual deciding whether she is in a gap is simply determining whether or not
the law is speaking to her. If the subject has to determine for herself whether she is
in a situation that legal reasons are supposedly excluding her other reasons, she can
only do this by weighing the reasons that exist for and against the behavior she is
contemplating in that situation. The characteristics of the situation that determine
whether or not she is in a gap in the law also provide the reasons relevant to the
contemplated action. Consider the novel decision about whether to ride one’s
Segway in the park upon confronting the sign indicating that vehicles are not
allowed. The very determination requires her to reject any claim to exclude reasons
that the law might be making. She must peer into the reasons behind the rule,
which is to say she must weigh the very reasons considered by the legislature, a
weighing that the NJT was to have helped us avoid.
51 While for Raz the legitimacy of any directive is dependent somewhat on its content in that it
must be based on dependent reasons, it is still the case that any obligation one has to a legitimate
directive flows from the fact that the authority is in a better position to get the subject to comply with
reason.
If, on the other hand, we do not see the law’s authority as consisting in a claim to
exclude non-legal reasons, but simply in the claim that it has a moral right to tell the
subject what to do (that is, the simple issuance of directives), then the subject has
only to determine whether she is covered by that directive. That will still involve
weighing the reasons that the legislature was to have considered, but doing so is no
longer quite so pernicious since she is admittedly in a gap. The law can still have
gaps as long as its authority is not claimed to be exclusionary. The authority can still
be exclusionary in that it is allowed to exclude reasons by the subject; but the claim
to authority is not itself a claim to exclude reasons. If she determines that the law
does apply to her situation, then she treats the legal directive as a preemptory reason
against contrary action. She allows it the status of a preemptory reason by acceding
to its authority over her and excluding those contrary reasons. But all the law has
done was to tell her what to do. It did not make any further claims about how she
should treat that directive vis-à-vis her pre-existing reasons against the action. That
part was all her.
One might seek to push this objection further based upon the realization that the
law can only exclude reasons when it provides guidance.52 If providing guidance is a
precondition for excluding reasons, then to say that the law provides no guidance in
the gaps entails that it is not claiming to exclude reasons in those gaps. Hence, so
this objection goes, this cannot be an argument against law not claiming to exclude
reasons since every gap theorist (including Raz) would immediately agree that the
law is not guiding (and hence not claiming to exclude reasons) in the gaps. Not
saying the law claims to exclude reasons in the gaps does not force the gap theorist
to agree that the law never claims to exclude reasons.
The reply to this requires us to return to what exactly the gap theorist claims and
what separates him from the theorist who denies that there are gaps. For the gap
theorist, there is no fact of the matter about what the law requires in the application
of vague or contradictory legal rules. It is not correct to say (with the anti-gap
theorist) simply that the closure principle leaves the matter up to the subject and
that is why the law is not providing guidance. Rather, since the law is a matter of
social fact, there is not yet a social fact about whether the term “vehicle” legally
applies to Segways. But this does not mean that we can assume that Segways are
legally allowed by the closure principle until the appropriate official says otherwise
(and applies that decision retroactively). That would be to say that there is already a
fact of the matter and that fact simply changes after the official’s decision, which
would be to deny the existence of the gap. Instead, the reason the law is not
providing guidance to the subject is that there is no fact of the matter about
whether the guidance offered by the law is aimed at that subject. A more precise
way of explaining the situation would be to say that the law offers guidance but that
there is no fact about whether it is offering YOU (the putative subject) that
guidance. In explaining gaps, above, I tried to suggest this by differentiating
between what the law offers and what it provides. In a gap, the law offers guidance
52 I am indebted to Stefan Sciaraffa for this articulation.

by issuing the general and vague or conflicted rule but never provides it to a subject
who cannot determine whether the rule applies to her.
Consider what happens in the two most common instances of gaps, vagueness
and conflict. In the case of vague laws, such as the vehicles in the park example, the
problem is that the law is offering guidance in such a general way that it is
impossible to determine whether that guidance is directed at you on your Segway
(because there is no fact of the matter about whether it is doing so). If you were a
legally sophisticated Segway rider who knew that no official had previously decided
whether the Segway was permitted and believed that hence there was no fact of the
matter of what the law requires you to do, you would know you are getting no
guidance from the law and hence are reduced to predicting what the official will do.
To make that prediction, you would probably need to peer into the reasons behind
the vague rule and decide if your reasons are likely to be officially interpreted as
covered by the rule. The less legally sophisticated Segway rider would try to
determine if the law is providing guidance by asking directly whether he comes
within the ambit of the vague term. That inquiry would also require him to peer
into the likely reasons behind the rule in a (futile) attempt to apply the vague term
where there is no fact about its application. (It is a futile attempt because there is
not yet any legal fact about whether the vague term applies and hence a direct
attempt to apply the term is not legally relevant.) So, even if we say that vague laws
are offering guidance, that vague offer cannot be seen as a claim to exclude the
reasons of those who cannot determine whether the vague term applies to their
situations. Since vagueness is unavoidably part of law’s nature, law cannot be
claiming to exclude reasons of some people to whom the vague law will eventually
be deemed to apply. Similarly, in the case of conflicting laws, we can say that the
law is offering guidance, but failing to provide it because it is offering conflicting
guidance. If we were to say that it is claiming to exclude reasons, then it would be
claiming to exclude reasons both for and against the contemplated action. In
offering conflicting guidance, it provides none. Since this is a structural feature of
law (wherever internally in conflict) the law cannot be said to claim to exclude
reasons as a conceptual matter. Since in both of these situations we can still say that
the law offers guidance in setting forth a rule, and since the gap theorist who holds
that the law claims to exclude reasons would say that the law is claiming to exclude
reasons wherever there is a directive, that theorist would have to say that the claim
to exclude reasons applies even in the case of gaps. The law is there offering
guidance but not successfully providing it. Hence if the law claims to exclude
reasons, that claim is patently false in any gaps. Rather than saddling the law with a
necessarily false claim (something Raz says we cannot do anyway), it is preferable
not to attribute this particular claim to it at all.
There is a deep way in which the choice-of-evils defense argument is actually a
particular example of the argument from gaps in the law. This was alluded to in the
need for the law to operate retroactively in recognizing novel necessity defenses. We
might understand the subject who is contemplating whether her subjective contrary
reasons rise to the level of a novel necessity defense to be considering whether she is
in a gap. After all, if those contrary reasons really are novel, then the law has not yet
spoken on whether or not they constitute an exception and there is not yet a fact of
the matter about whether they are an exception to the directive. By attributing to
law the claim to preempt those reasons even as it holds out the possibility of
exempting them from exclusion later on, we force the law into conflict with the
avowed function of authority: to guide and coordinate behavior. Of course all
retroactive operation of law (including any even slightly hard case decided by a
court) cannot serve to guide the behavior of those to whom it is retroactively
applied. But that does not create the same problem since we are usually concerned
with law’s behavior guidance overall when iterated and applied to others. Here,
however, by attributing the claim of preemption to law we render the law’s
demands incoherent to anyone in such a gap. Given that gaps are inherent in
law, the way to avoid seeing that incoherence in the fundamental nature of law itself
is to say that the claim to authority is not itself a claim to preemption of contrary
reasons. The subject allows the directive to preempt contrary reasons as a conse-
quence of acceding to its authority.
One might note that both gaps in the law and novel choice-of-evil defenses are
“penumbral” areas of law, where subjects are unsure of whether and how the law
applies to them. This gives rise to the possible reply that it is not generally a good
idea to reason straight from such penumbral areas to conclusions about the nature
of law, since they are exceptional circumstances.53 However this objection mistakes
the use to which these penumbral areas are being put. While it might be rare that
one finds oneself in such a gap or confronted with a possibly novel justification, the
very presence of such circumstances is not penumbral to law itself. That is, if the gap
theorists are right, then all legal systems necessarily have gaps. And we would
certainly not say that legal systems that leave open the possibility of novel choice-of-
evils defenses are outliers or penumbral examples of legal systems. It is the
(ubiquitous) presence of these elements within the law that calls into question
Raz’s attribution of the claim to preemption. We cannot say that it is a conceptual
truth about law that it claims to exclude contrary reasons when so many paradig-
matic legal systems allow for novel choice-of-evils defenses and when all legal
systems have gaps. While it is true that one will only rarely find oneself in such a
gap, it is what the subject must do in order to determine whether she is in a gap and
how to behave once she decides that she is that preclude the notion that the law
must be claiming to exclude contrary reasons.
4. Conclusion
The claim to authority is not a claim to provide reasons. It is a claim of a right to

direct action.54 Reasons are irrelevant and opaque to the claim of authority itself;
they are relevant to the assessment of that claim by the subject and to the decision
about how to exercise that authority by the claimant. Raz’s theory is primarily
53 I thank Arie Rosen for raising this objection. 54 Raz (2006: 1012).
directed to the subject of authority.55 In almost all of the expositions of Raz’s

theory, the focus is on the reasons for action and compliance of the subject. The
problems, theoretical and moral, that the service conception is supposed to solve,
are primarily problems from the subject’s point of view.56 The normal justification
thesis is that an authority is justified when the subject does better at conforming to
reason by submitting to that authority. The preemption thesis is that one complies
with authority by replacing one’s reasons with the reason that its directives issue
from a source with justified authority. It is in being the recipient of a justified claim
of a right to direct action that we accede to that claim by replacing our reasons with
its direction. But this does not entail that the authority is asking us to replace our
reasons; it is only asking us to act as it directs. Hence the authoritative directive is
not itself a claim to replace any reasons; it is only received as such when its authority
is acknowledged by the recipient.
References
Alexander, L. (1990). “Law and Exclusionary Reasons,” Philosophical Topics, 18(1): 5–22.
Dworkin, R. (1991). “On Gaps in the Law,” in P. Amselek and N. MacCormick (eds.),
Controversies about Law’s Ontology. Edinburgh: Edinburgh Univeristy Press.
Edmundson, W. A. (1998). “Legitimate Authority without Political Obligation,” Law and
Philosophy, 17(1): 43–60.
Edmundson, W. A. (2004). “State of the Art: The Duty to Obey the Law,” Legal Theory,
10: 215–59.
Green, L. (1988). The Authority of the State. Oxford and New York: Clarendon Press.
Green, L. (1989). “Law, Legitimacy, and Consent,” Southern California Law Review, 62:
795–826.
Hart, H. L. A. (1982). Essays on Bentham: Jurisprudence and Political Theory. Oxford:
Clarendon Press.
Himma, K. E. (2001). “Law’s Claim of Legitimate Authority,” in Jules Coleman (ed.),
Hart’s Postscript. New York: Oxford, 271–309.
Hurd, H. M. (1991). “Challenging Authority,” Yale Law Journal, 100: 1611–77.
Hurley, S. L. (1989). Natural Reasons: Personality and Polity. New York: Oxford University
Press.
Korta, K. and Perry, J. (2011). “Pragmatics,” in E. N. Zalta (ed.), The Stanford Encyclopedia
of Philosophy <http://plato.stanford.edu/archives/sum2011/entries/pragmatics/>.
55 “[A]uthority helps our rational capacity whose function is to secure conformity with reason. It
allows our rational capacity to achieve its purpose more successfully.” Raz (2006: 1012).
56 The theoretical problem is: “how could it be that the say-so of one person constitutes a reason, a
duty, for another?” The moral problem is: “how can it ever be that one has a duty to subject one’s will
and judgment to those of another?” Raz (2006: 1012). Both are problematic primarily when seen from
the subject’s point of view. The authority is more likely confronted with problems of how to generate
compliance and how to justify its legitimacy. Raz’s two problems are certainly relevant for those
concerns, but in a derivative way.
Kramer, M. H. (2005). “Legal and Moral Obligation,” in M. P. Golding and W. A.
Edmundson (eds.), The Blackwell Guide to the Philosophy of Law and Legal Theory.
Blackwell Philosophy Guides. Malden, MA: Blackwell Publishing, 179–90.
Moore, M. S. (1989). “Authority, Law, and Razian Reasons,” Southern California Law
Review, 62: 827–96.
Perry, S. R. (1989). “Second-Order Reasons, Uncertainty and Legal Theory,” Southern
California Law Review, 62: 913–94.
Raz, J. (1979). The Authority of Law: Essays on Law and Morality. Oxford and New York:
Clarendon Press.
Raz, J. (1984). “The Obligation to Obey: Revision and Tradition,” Notre Dame Journal of
Ethics and Public Policy, 1: 139.
Raz, J. (1985). “Authority and Justification,” Philosophy and Public Affairs, 14: 3–29.
Raz, J. (1986). The Morality of Freedom. Oxford and New York: Clarendon Press.
Raz, J. (1989). “Facing Up: A Reply,” Southern California Law Review, 62: 1153–236.
Raz, J. (1990). Practical Reason and Norms. 2nd edn; Princeton: Princeton University Press
(first published 1975).
Raz, J. (1998). “Postema on Law’s Autonomy and Public Practical Reasons: A Critical
Comment,” Legal Theory, 4(1): 1.
Raz, J. (2006). “The Problem of Authority: Revisiting the Service Conception,” Minnesota
Law Review, 90: 1003.
Raz, J. (2009). Between Authority and Interpretation: On the Theory of Law and Practical
Reason. Oxford and New York: Oxford University Press.
Regan, D. H. (1989). “Authority and Value: Reflections on Raz’s Morality of Freedom,”
Southern California Law Review, 62: 995–1095.
Regan, D. H. (1990). “Reasons, Authority, and the Meaning of ‘Obey’: Further Thoughts
on Raz and Obedience to Law,” Canadian Journal of Law and Jurisprudence, 3: 3–28.
Shapiro, Scott J. (2002). “Authority,” in J. L. Coleman and S. Shapiro (eds.), The Oxford
Handbook of Jurisprudence and Philosophy of Law. New York: Oxford Universty Press,
382–439.
Soames, S. (1987). “Direct Reference, Propositional Attitudes, and Semantic Content,”
Philosophical Topics, 15: 47–87.
Soames, S. (2008). “Interpreting Legal Texts: What is, and What is not, Special about the
Law,” in S. Soames (ed.), Philosophical Essays, vol. 1: Natural Language: What it Means
and How We Use It (Princeton: Princeton University Press), 403–23.
4
The Normative Fallacy Regarding
Law’s Authority*
Arie Rosen
1. Introduction
This chapter examines the question of law’s authority and its significance to
philosophical investigations into the idea of law. The main argument offered here
is that descriptive jurisprudence should concern itself with the socio-political
question of law’s de facto authority, rather than with the moral question of its
rightful authority. This, I argue, has both conceptual and methodological implica-
tions for legal philosophy.
The argument is divided into two parts. Section 2 is dedicated to the distinc-
tion between moral inquiries into law’s rightful authority and inquiries into the
socio-political phenomenon of its de facto authority. After discussing this funda-
mental methodological point, I move on to offer a critique of Raz’s theory of de
facto authority, which is currently one of the most influential philosophical
accounts of this phenomenon. My claim is that Raz’s account involves an
implicit, mistaken predication of the concept of de facto authority on his
particular theory of morally rightful authority. This predication is identified as
an instance of the normative fallacy—that is, an illicit move from statements in
moral theory to statements about social and political facts. Section 3 builds on
the conclusions drawn in section 2 regarding the relations between rightful and
de facto authority. It shows that our idea of law has been shaped and continues
to function in circumstances in which law has de facto authority. I therefore
argue that it is law’s de facto authority (and not its rightful authority) that
regulates our idea of law.
* I am grateful to Jeremy Waldron, Liam Murphy, and David Dyzenhaus for their advice and
comments on previous drafts of this chapter, and also to Ken Eherneberg, John Ferejohn, Ruth
Gavison, Noam Gur, Lisa Kerr, Roy Kreitner, Hillary Nye, Stefan Sciaraffa, Guy Sinclair, Kevin Toh,
and Michel Troper for their comments and suggestions. I would also like to thank the editors of this
volume and the participants at the 2011 conference on “The Nature of Law—Contemporary Perspec-
tives,” held at McMaster University.
76 Arie Rosen
Besides the importance of the analytical discussion of the concept of authority,
and the new critique it offers of Raz’s theory of de facto authority, this chapter
generally addresses the role of moral theory in descriptive legal philosophy. As we
shall see, in order to avoid the normative fallacy in our legal philosophy, we must
move away from an analysis of the idea of law that is based on the theorist’s own
moral theory, to an exploration of law’s de facto authority and the ideological and
practical reality of which it is part. This means that we should take a step back from
the moral theory of law’s rightful authority, and focus our philosophical attention
on the de facto authority of law, even if this de facto authority might be based on
widespread erroneous and misguided moral convictions.
2. Rightful authority and de facto authority

2.1. A methodological fallacy
I have already been using the terms “de facto authority” and “rightful authority” in
the introduction, and before going any further I would like to make clear exactly
what I mean by these terms. “Rightful authority” (also: “legitimate authority”)1
refers to a moral phenomenon. It describes a morally justified power to guide the
behavior of another, a power which also commands morally justified obedience.
“De facto authority” (also: “effective authority”) refers to a social relationship in
which one party is in a position to guide the practical reasoning2 of another, based
on the latter’s belief that obedience is legitimate.
An important account of the social phenomenon of de facto authority can be
found in the sociology of Max Weber. In his discussion of domination,3 Weber
notes that the motivation for submitting to relations of domination can be diverse,
and can range from uncritical habituation, to fear, to a belief in moral legitimacy.
“Legitimate domination” (that is, domination believed to be morally legitimate)
is distinguished from naked power and from domination secured by means of
coercive violence. Authority for Weber is defined as this legitimate domination,
i.e. as domination in which obedience is motivated by a sense of the moral
appropriateness of obedience.4 I will use the term “de facto authority” in the
same sense.
1 I prefer the term “rightful authority” to Raz’s “legitimate authority,” although the two describe the
same moral phenomenon. The term “legitimate authority” suggests an immediate ambiguity, an
ambiguity on the very point which I claim must be made clear. The term “legitimacy” itself has a
sociological sense as well as a moral one, and for that reason the term “legitimate authority” ends up
being less clear than it could be, and invites, I think, confusion and misunderstanding. Cf. Raz
(1979: 3).
2 Since our ultimate interest here is in law, I will focus on the question of practical authority, and
not on speculative authority regarding belief.

3 Domination (Herrschaft) is defined as a relationship in which there is sufficiently high probability
of obedience by one party to an issued command of another. Weber (1968: 61–2).

4 See Weber (1968: 212, 299, 943); Giddens (1971: 154–6). Cf. Habermas (1998: 30).
Normative Fallacy Regarding Law’s Authority 77
The dependence of the social phenomenon of de facto authority on a moral

belief is echoed in Raz’s account of the conceptual connection between de facto
authority and rightful authority. Raz explains that5—
(1) People with de facto authority are distinguished from people with naked
power by virtue of their claim to rightful authority.
(2) Consequently, any account of de facto authority presupposes an account of
rightful authority.
(3) Consequently, “the concept of legitimate [i.e. rightful] authority has explanatory
priority over that of a mere de facto authority. The latter presupposes the former
but not the other way around”.6
As long as we are concerned with social phenomena, it seems that we should prefer
Weber’s emphasis on the belief in rightful authority as the basis for relations
of de facto authority to Raz’s insistence on de facto authority’s “claim” to rightful
authority.7 Still, the relativizing, naturalizing moves of both accounts are similar
and appropriate. If we consider the phenomenon of de facto authority it is clear that
it is connected to a moral belief (or, in Raz’s case, a claim to moral appropriateness),
and not to morality proper.
Despite the connection between them, the phenomena of de facto authority and
rightful authority do not always overlap. The source of the potential disjunction is
in the possibility of divergence in opinions regarding rightful authority. Although
moral truth and moral beliefs can coincide, it is uncontroversial that they can come
apart as well. Many of us think that there is an objective moral standard of rightful
authority; but we equally acknowledge that people and even entire societies can
sometimes be wrong in their identification of this standard. When they are, we have
a social reality of de facto authority that exists despite the lack of rightful authority
on the moral level.
The potential disjunction between rightful and de facto authority explains the
plurality of forms de facto authority can take. Rightful authority is often under-
stood in unitary terms. Theories of rightful authority usually develop a particular
account of justified authority and negate or reject rivaling moral theories. However,
when it comes to de facto authority, it is quite possible for multiple moral
convictions regarding authority to coexist simultaneously. Since de facto authority
describes a social and cultural fact (rather than a moral reality), and since it is
possible for different people to hold different views about rightful authority (since,
5 Raz (1986: 25–6); Raz (2006: 1005–6). In summarizing Raz’s account, I do not follow his
terminology of “legitimate authority,” and I substitute this term with “rightful authority.” See n. 1.
6 Raz (2006: 1006).
7 An immediate question, which I would not like to discuss here, is how we should understand such
“claiming.” This sounds like a metaphor, as authorities—either abstract or concrete—do not, literally,
claim anything of the sort. One might say that such a claim is implicit, or is attributed to authorities by
others, but it is very rare to find such a claim actually being made. It therefore seems unhelpful to have
this “claim” at the center of our definition of de facto authority. This will become even more
problematic when Raz argues that law—an inanimate object—claims rightful authority, but we can
leave this problem aside for now. Cf. Dworkin (2002: 1666–7); Himma (2007: 22).
78 Arie Rosen
at the very least, some people can be wrong), there can be multiple grounds for
accepting de facto authority. The grounds on which de facto authority is justified
can change the characteristics and form of the relationship itself. The way in which
we justify de facto authorities will affect, for example, whether authoritative reasons
will have an exclusionary effect vis-à-vis other reasons. I will say more about this
in a short while. For now, however, I would just like to make clear the following
uncontroversial claim: in our contemporary social circumstances there are, in fact,
multiple grounds for the justification of de facto authority.
Consider for example the question of the rightful authority of government and
the many answers to this question that are available in our culture. According to
one of the most prevalent theories of rightful authority, the core feature which
makes for governmental authority is the ability to ensure order, security, and
cooperation.8 Some philosophers claim that the very presence of a single govern-
ment which is commonly followed is essential for the security of individuals within
society from internal and external violence, and that this gives us reasons to follow
the instructions of effective governments.9 In a wider variation, the government’s
ability to solve coordination problems or to bring about economic and social
development underlies its rightful authority.10 From yet another perspective,
submission to government can be seen as justified when it leads to the protection
of rights within society.11 These different views share a strategic move: they all
identify the establishment of government as in itself a legitimizing good, at least as
long as the person or institution in charge fulfills its main purpose (whatever that
may be).
As appealing as this strand of thought is, it is not the only—and sometimes not
even the main—ground for submission to government. Other theories of political
morality claim that only some special attribute of the particular judgment or will of
the governing party can legitimize domination. In this way, some philosophers
emphasize the formation of a particular will, either of the sovereign,12 the legisla-
ture,13 or the public14 as the crucial factor in the moral legitimation of government.
Submission to such wills can be justified on different grounds, most commonly
related to consent,15 authorization,16 or representation.17
A third group of theories includes views in political morality, according to which
rightful authority depends on the correctness (or probable correctness) of political
decisions. According to these views, the judgment embodied in authoritative
decisions is the source of the respect and obedience that we owe them. In this
way, political decisions are sometimes said to embody or approximate a standard of
8 Finnis (1980: 246).

9 E.g. Hobbes (1996); Schmitt (1996: 52).
10 Raz (2006); Finnis (1980: 244).
11 Kant (1983: 112); Finnis (1980: 246). See also Waldron (1996: 1562–6).
12 E.g. Hobbes (1996: 183–8).
13 E.g. Hamilton et al. (1999: nos. 78, 81). Cf. Waldron (1999).
14 E.g. Rousseau (1987: 139, bk. 1 ch. 6).
15 E.g. Locke (2008: ch. 8); Simmons (1999: 768–9).
16 E.g. Hobbes (1996: ch. 16).
17 E.g. Rousseau (1987). Cf. Kant (1996: 91).
reason achieved through public deliberation,18 or reflect the common good of the
society,19 or the common mores and ethics of a particular political community.20
This (probable) correlation is said to be the source of their authority.
As these examples show, de facto authority can be based on various conceptions
of rightful authority. It might be true that no government can ever be rightfully
(truly, morally) authoritative, or that only a government based on political partici-
pation can exercise rightful authority. These would be claims in moral theory, and
they do not directly concern us here. There is no doubt, however, that insofar as de
facto authority is concerned, multiple convictions exist. One does not have to be a
relativist or a nihilist in order to acknowledge this social reality. Different people
have, as a matter of social fact, different beliefs regarding the moral question of
rightful authority of government. This is clearest when it comes to people and
institutions in different times and places, but is also true within contemporary
pluralist societies: right or wrong, there are different common understandings of
what the grounds of rightful authority are.
This observation has important methodological implications. It should make us
wary of a false move that philosophers are prone to make, a move I call the
normative fallacy.21 In order to understand the fallacy and its origin, let us consider
the following two statements:
(a) The fact that this government enjoys de facto authority can be explained by
the citizenry’s convictions regarding its rightful authority.
(b) By finding the right moral theory of rightful authority we can find the source
of the de facto authority of governments.
While (a) sounds right and flows from our definition of “de facto authority” as
predicated on a belief in rightful authority, (b) expresses a methodological fallacy.
As we saw, forms of de facto authority can be radically different from each other,
and also radically different from the right theory of political morality. An illicit
strategy for theorists would be to engage in moral philosophy, come up with a
theory of rightful authority, and uncritically continue to give an account of de facto
authority that duplicates their moral theory into the factual realm. Strictly speaking,
doing so would be a mistake: one cannot make factual statements about convictions
and beliefs that uphold social structures based solely on one’s moral theory. Rather,
if we want to find out what forms of de facto authority are prevalent in a certain
society we should look at the actual practices of governance and justification that
people engage in.
It might be argued that moral philosophy cannot help but be attentive to the
actual understandings of concrete communities. This is probably true. However, it
18 E.g. Cohen (1986); Estlund (1997: 173).

19 Bentham (1988: 116–22); Mill (1992: 26–35); Lieberman (2008: 614–17); Postema (1986:
268).
20 E.g. Habermas (1998: 159). Cf. Dworkin (1986: 345–7).
21 I found that the term has been previously used in Campbell (1970), with a similar meaning
(“This fallacy consists of arguing from propositions which are themselves normative, or could count as
evidence only for normative propositions, to conclusions which contain factual assertions”).
80 Arie Rosen
is one thing to claim that every moral conception is rooted in a certain culture and a
certain tradition, and quite another to claim that any single moral theory reflects the
entirety of intuitions and convictions prevalent in our culture. As we saw, in our
culture there is a plurality of moral views and conceptions of rightful authority. No
single moral theory can consistently incorporate all of these different views, nor
should it be its goal to do so.
Consider the following example. Let us imagine that Professor K, a gifted
moral theorist, succeeded in finding the one true moral theory concerning the
authority of political institutions. Unfortunately, this theory is inconsistent with
the conceptions of institutional authority prevalent in K ’s society. What we have
here is a theory of rightful authority that does not tell us much about the de
facto authority of institutions. If K equates his theory of moral justification with
people’s beliefs regarding authority, he will end up claiming something false. He
will be deducing something about the actual moral beliefs of a community from
his moral theory. This would be an instance of the normative fallacy regarding
authority.
It might be that any theory is destined to deviate from reality. Even if K had been
a brilliant sociologist rather than a moral theorist, he still would have been dealing
in abstractions and generalizations, and his theory would not have been an accurate
record of the many divergent and idiosyncratic convictions people in his society
have. There might be an unbridgeable gap between theory and social reality in this
regard; but this was not the problem in K ’s case. As a piece of social theory K ’s
theory was not inaccurate or too general; it was simply a theory of the wrong thing:
it was theorizing rightful authority when it should have been theorizing de facto
authority.
2.2. Raz’s normative fallacy

The normative fallacy is a pitfall that constantly awaits theorists of social phenom-
ena. “Being legitimate,” or “justified” or “having a reason,” can be understood
either in a factual, sociological sense or in a moral, universal sense. It is always a
mistake to predicate accounts of any of these social phenomena on the accounts of
their moral counterparts. In the rest of this section, however, I would like to focus
on the normative fallacy regarding authority, and in particular on the way it features
in Raz’s important account of this concept.
Now, Raz is not our hypothetical Professor K, and he does not think that
everybody espouses his particular theory of rightful authority. Raz knows that
people might disagree or be confused regarding what rightful authority is (that is,
what morality actually tells us about the legitimacy of relations of domination).
Clearly, people disagree about this moral question, and when they “claim” rightful
authority they might have different ideas of what they are claiming. Think for
example of the difference between the authority associated with the Pope, the
authority associated with the British Parliament, and that associated with the
People’s Assembly in ancient Athens. The Pope, the Parliament, and the Assembly
may all “claim” to have “rightful authority,” but they would have very different
ideas of what this means. Raz denies none of this.22
In other places in his theory, however, Raz is less careful in attending to the
implications of this factual plurality. At times, he lets his moral theory of rightful
authority inform his social theory of de facto authority. This is the case in at least
two important arguments Raz makes: one regarding the relations between law’s de
facto authority and its positivity, and the other regarding the exclusionary nature of
factually authoritative reasons.
Raz’s theory of rightful authority is, appropriately, a moral one. It is a moral
theory since it is presented as answering a moral question, namely: how can
relations of deference and obedience be morally justified?23 Raz’s claim in this
regard is that “[t]he basis of legitimacy is relative success in getting people to
conform to right reason.”24 He therefore finds that the only circumstances under
which deference can be legitimate are those in which, by following the authoritative
utterance, the person subject to authority improves her conformity with reasons
that already apply to her.25 This is the basis for Raz’s famous three theses. He
concludes that people with rightful authority base their directives on reasons that
apply independently to the subjects of their authority (the “Dependence
Thesis”);26 that the normal way to justify authority is to show that by following
its directives people better conform to reason (the “Normal Justification Thesis”);27
and that all authoritative commands exclude the consideration of these prior
reasons, on which it is claimed that they are based (the “Pre-emption Thesis”).28
Raz argues for all these theses on moral grounds.29
It is extremely important that we see that the three features of Raz’s account only
make sense in cases in which domination is legitimized by reference to its ability to
lead to better conformity to reason. Therefore, they do not transfer to the descrip-
22 Raz (1986: 65–6). See also text accompanying n. 56.

23 Raz (1986: 38). I use the term “deference” as the attitude correlating to domination. When
someone (or in our case, an inanimate object such as law) is in a position to guide the behavior of
another, the latter is deferring to the former’s instructions.
24 Raz (1990: 13).
25 Raz (1979: 27). Raz thus believes that expertise and special capabilities of coordination can
justify domination. See Raz (1999: 63); Shapiro (2002: 405–6); Marmor (2005: 134–5).
26 See Raz (1986: 47): “[A]ll authoritative directives should be based on reasons which already
independently apply to the subjects of the directives and are relevant to their action in the circum-
stances covered by the directives.”
27 See Raz (1986: 58): “[T]he normal way to establish that a person has authority over another
person involves showing that the alleged subject is likely better to comply with reasons which apply to
him (other than the alleged authoritative directives) if he accepts the directives of the alleged authority
as authoritatively binding and tries to follow them, rather than by trying to follow the reasons which
apply to him directly.”
28 See Raz (1986: 46): “[T]he fact that an authority requires performance of an action is reason for
its performance which is not to be added to all other relevant reasons when assessing what to do, but
should exclude and take the place of some of them.”
29 Raz (1986: 38). “The explanation proceeds through normative theses of three kinds. One
concerns the type of argument required to justify a claim that a certain authority is legitimate. The
second states the general character of the considerations which should guide the actions of authorities.
The last concerns the way the existence of a binding authoritative directive affects the reasoning of the
subjects of the authority” (emphasis added).
82 Arie Rosen
tion of all cases of de facto authority. Some relations of de facto authority might
indeed be based on the sort of convictions that Raz’s theory of rightful authority
identifies. It might be, for example, a good account of the de facto authority of
appointed experts.30 But Raz’s theory of rightful authority corresponds, as we saw,
to only one of the many grounds which are available in our political tradition for
the justification of domination.31
As far as Raz is concerned, this is not very interesting: people might simply
be wrong, either in acknowledging others’ authority over them or in understanding
the foundations of this authority. This, however, would make his conclusions
regarding de facto authority extremely problematic. What can Raz’s theory of
rightful authority—even if correct in moral terms—tell us about the way people
who deny it behave and reason? Naturally, it cannot tell us much about “the de
facto authority of law,” if this authority is premised on a completely different
understanding of what justifies domination and deference.
Still, the three theses—which were developed on normative grounds—feature
in Raz’s descriptive account of de facto authority. We can see an example of this in
Raz’s argument in favor of legal positivism. Raz suggests that there is something in the
way law functions in our lives that makes it authoritative: it fits into our practical
inferences in a certain (authoritative) way, as a special reason for action.32 His analysis
brings him to the conclusion that all law must be positive law. His argument, as it
appears in Authority, Law, and Morality,33 can be summarized in the following way:34
(P1) All law has de facto authority;35
(P2) All de facto authority claims rightful authority;36
(C1) Law claims rightful authority, and such a claim is a part of the nature of law;37
(P3) Since law claims rightful authority it has to be able to possess rightful authority;38
(P4) In order for something to be able to possess rightful authority it must be able to play a
mediating role between people and reasons which apply to them;39
(P5) Only positive law can play such a mediating role40 (because only positive law can be
presented as someone’s view and can be ascertained without reference to the considerations
underlying its content);41
30 Let’s say that I am contemplating whether or not to take a certain medicine. I try to consider all
the various reasons for and against doing so. I go to an expert. She tells me that I must take the
medicine. It is implied that her statement is based on reasons that independently apply to me. As my
reason for going to the expert was for her to give me a reason to act which would replace all these
underlying reasons I had, it also makes sense that her statement would be taken by me as a reason for
action, and would exclude previous medical considerations I had regarding this practical question. Cf.
Marmor (2005: 135).
31 See text accompanying nn. 8–20. Moreover, some argue that Raz’s theory of rightful authority is
marginal among conceptions of rightful political authority in this tradition. See e.g. Dworkin (2002:
1675); Shapiro (2002: 401–2); Perry (2007).
32 Raz (1979: 10).
33 Raz (1994).
34 I am substituting, again, “legitimate authority” with “rightful authority.” See n. 1.
35 Raz (1994: 199). 36 Raz (1994: 199). 37 Raz (1994: 199).
38 Raz (1994: 199, 203). 39 Raz (1994: 202–4).
40 Raz (1994: 205–9). 41 Raz (1994: 202).
(C2) Only positive law can claim rightful authority,42 and therefore, it follows from C1
that—
(C3) Only positive law is law.43
This argument came, of course, under criticism by non-positivists,44 and there is no
need to rehash and assess these attacks here. To these critiques one should add the
fact that Raz’s theory of de facto authority, as it is employed in this argument,
involves a methodological fallacy. The first premise of his argument (P1) is about
law’s de facto authority. This is the only assumption Raz is willing to make
regarding law in this argument. Still, this argument involves assertions taken
from Raz’s own particular theory of rightful authority. In P3 and in P4, Raz makes
an illicit move by assuming that since law has de facto authority the “claim” law
makes (or the belief people hold about it) is that it is playing a mediating role
between people and reasons that already apply to them.
As we have seen, whatever it is that de facto authority “claims” need not
necessarily correspond to Raz’s account of rightful authority. Even assuming that
Raz’s moral theory is sound, law’s de facto authority might be based on misguided
convictions. The claimed rightful authority can be based, perhaps erroneously, on
consent, sanction by popular will, or respect to institutions. However, it is clear that
in the latter part of this argument (P3 to C2)—an argument regarding the de facto
authority of law—“rightful authority” is used in the universal-normative sense à la
Raz’s moral theory. When Raz says “rightful authority” he means “morally justified
judgment-based authority.” We know this since other kinds of de facto authority
are not committed to playing a mediating role between people and reasons.
Perhaps unwittingly, Raz plugs his moral theory of rightful authority into his
argument in social theory regarding the social phenomenon of law’s de facto
authority. Raz’s argument therefore fails to prove that there is something in the
concept of de facto authority that necessitates a positivist concept of law. The fact
that people tend to follow the law, or that people believe that they are right in
ascribing rightful authority to law, does not win the day for legal positivism. In
circumstances in which the rightful authority “claimed” by law is not judgment-
based, P4 is false and C2 and C3 are non-sequiturs.
This is not only a refutation of Raz’s argument in favor of legal positivism, but
also a proof against those who would like to interpret his general theory of authority
in a manner that saves it from committing the normative fallacy. In fact, we can see
in this later argument that the Razian inference 1 to 3 regarding the “explanatory
priority” of rightful authority45 is not as innocuous as it might seem. This
conclusion is supported by another instance of the normative fallacy, which occurs
in Raz’s argument for the exclusionary nature of authoritative reasons.
According to Raz, authoritative utterances serve as reasons for actions. They are
special in that they also serve as reasons for the exclusion of other reasons.46 This
42 Raz (1994: 213). 43 Raz (1994: 213).

44 E.g. Dworkin (2002). 45 See text accompanying nn. 5–6.
46 See Raz (1999: 34–48, 183); Raz (1986: 46).
84 Arie Rosen
exclusionary quality is explained in Raz’s later work in terms of his theory of rightful
authority:
[Authoritative directives] exclude reliance on conflicting reasons, not all conflicting reasons,
but those that the lawmaker was meant to consider before issuing the directive. . . . Think
about it: authority improves our conformity with reason by overriding what we would do
without it, when doing so would not conform with reason.47
It is now clear, however, that there can be no argument in favor of the exclusionary
nature of de facto authority from Raz’s theory of rightful authority. Raz employs
here, in his argument regarding de facto authority, two of the three theses de-
veloped in his moral theory of rightful authority: the Preemption Thesis and the
Dependence Thesis.
Here is Raz’s Preemption Thesis:
the fact that an authority requires performance of an action is reason for its performance
which is not to be added to all other relevant reasons when assessing what to do, but should
exclude and take the place of some of them.48
Raz explains—and this is important—that the Preemption Thesis stems from the
Dependence Thesis:
The argument for the pre-emption thesis proceeds from another, which I shall call the
dependence thesis. It says: all authoritative directives should be based on reasons which
already independently apply to the subjects of the directives and are relevant to their action
in the circumstances covered by the directives.49
As we saw, the Preemption Thesis and the Dependence Thesis are not a necessary
part of all forms of de facto authority (since not all forms of de facto authority are
based on the conception of moral legitimacy that Raz espouses). It would therefore
be wrong to establish a connection between de facto authority and exclusionary
reasons based on the Preemption Thesis.
In order to see this we can consider an instance of de facto authority that is not
based on Raz’s theory of rightful authority. Take Socrates for example. Socrates
knew that he had no reason, apart from the law, for drinking hemlock. The law was
not claiming to mediate between Socrates and reasons that independently apply to
him, and Socrates’ practical reasoning was not premised on such claims. Socrates
ended up drinking the hemlock and dying. He might have been wrong in acknow-
ledging the authority of law in his case (Raz, you, and I might believe that he was
wrong in this way), but he drank the hemlock nonetheless, and Plato reported his
practical reasoning in Crito. His practical reasoning was working in a certain way,
based on conceptions of political morality that are neither Raz’s nor mine, but were
definitely his. Based on this understanding of the authority of law, its authority is
not based on a claim to play a mediating role between people and reasons, but on a
47 Raz (2006: 1022). 48 Raz (1986: 46). 49 Raz (1986: 47).

belief in a quasi-contract between Socrates and the laws, a notion that was related in
Socrates’ mind to consent, gratitude, and social need.50
This insight should lead us to reject Raz’s exclusionary account of de facto
authority. It must make us wonder what reasons were supposed to be excluded
from Socrates’ reasoning. Raz does not argue that authoritative statements exclude
all reason, but only the dependent reasons that the person in authority should have
considered. It makes absolutely no sense to exclude considerations of a kind that the
person with authority could not consider, and Raz does not claim that we exclude all
reasons contrary to authoritative directives. So, then, what reasons were supposed
to be excluded by Socrates? All reasons against drinking hemlock? That would be
going further than Raz. Only some reasons? But then, which? The problem is that if
de facto authority does not “claim” to play a mediating role between people and
dependent reasons (no Normal Justification Thesis and no Dependence Thesis)
then there is no reason to exclude reasons (no Preemption Thesis); and this is
exactly what happened in Socrates’ case. Since in this case law’s authority was not
based on its claim to play a mediating role between people and reasons there was
simply no group of dependent reasons that he should have excluded. Of course,
some reasons can triumph over other reasons in other ways besides exclusion, that
is, by outweighing them. Wouldn’t it do more justice to Socrates’ process of
practical reasoning to say that law’s authority gave Socrates (what he took to be)
very weighty moral reasons for drinking hemlock, reasons that outweighed the
other reasons he had against doing so? Isn’t that what was actually going on? This
was, after all, Socrates’ own understanding of things.51
My point, of course, is general, and goes beyond Socrates’ example. Raz is
completely right in suggesting that the consequences and function of de facto
authority depend on the grounds people believe they have to follow the authority’s
directive. If one espouses Raz’s theory of rightful authority, her practical reasoning
will follow Raz’s exclusionary logic. However, the de facto authority of institutions
can be effectively justified in different ways as well, based on ideas of representation
and consent, will-formation, legality, et cetera. When this is the case, it makes no
sense for people to treat authoritative reasons as exclusionary. Raz might contest the
claim that these other types of de facto authority are based on valid moral theories,
but his contestation would be at the moral-normative level. Raz cannot contest the
availability of radically different convictions in political morality and their influence
on our culture and tradition of thinking about political institutions. He cannot
deny the fact of the existence of these forms of de facto authority.52 Nor can he
50 Plato (1892: 151–5, paras. 50–3).

51 Plato (1892: 156, para. 54) (“Listen, then, Socrates, to us who have brought you up. Think not
of life and children first, and of justice afterwards, but of justice first, that you may be justified before
the prince of the world below”).
52 This point is different from the point recently made by Marmor regarding rightful authority:
Marmor (2011). Marmor argues that the moral legitimacy of domination can vary depending on the
details of its institutional exercise, so that different institutional practices of domination should be
legitimized in different ways. This strikes me as a cogent claim in moral theory, which shows that
rightful authority can come in different shapes and sizes. To this multiplicity of ways in which
domination can be morally legitimate, de facto authority adds another level of variance. This level of
86 Arie Rosen
deny that different models of de facto authority—informed by different convictions
about rightful authority—lead to different modes of practical reasoning and distinct
social relationships. Different understandings of the grounds of rightful authority
lead to different models of de facto authority. This makes Raz’s theory inadequate
for explaining many if not most instances of de facto political authority.53
In his earlier work, Raz puts forward a descriptive argument in favor of the
exclusionary nature of authority.54 Although I am not discussing here this alterna-
tive line of argument, I think that my inquiry reveals how much Raz’s descriptive-
social theory is enmeshed with his normative-moral work. If de facto authority is
not premised on its ability to play a mediating role between people and reasons,
what reasons is it supposed to exclude? Since the rulers do not claim to be basing
their decisions on reasons that already apply to people, and since no one is treating
the rulers as playing this mediating role between people and reasons, there is no
limited set of reasons that people should exclude. It seems that Raz’s description of
the exclusionary nature of de facto authority makes perfect sense given his model of
rightful authority, but cannot apply to cases in which other conceptions of rightful
authority are espoused.
2.3. A note on semantic externalism

Before I move on, I should respond to an important objection.55 This objection
sees something dangerous in the move I make from the plurality of conceptions
that people have regarding rightful authority to a certain plurality of models of de
facto authority. The objection might go something like this:
Although people clearly have different personal understandings of a certain concept (particu-
lar “conceptions”), they do nonetheless share a single concept—in our case, a concept of de
facto authority. Since my account insists on a plurality of models of de facto authority which
is the result of different beliefs people have regarding rightful authority, it might seem that
Raz and I are simply engaged in different projects. While Raz is interested in a concept we all
share (that is, in the meaning we as a socio-linguistic community attribute to “de facto
authority”), I am concerned with the variety of individual understandings (or misunder-
standings) of this shared concept.
The objection builds on a distinction between individual understandings and
concepts present in a certain culture. Indeed, when Raz considers the possibility
multiplicity is due to the fact that there is no necessary connection between the de facto authority of an
institution and its rightful authority.
53 It is important to note that Raz is by no means alone here. Many thinkers develop a unitary
conception of legitimacy and build a political theory around it. It is in this way that Hobbes equates
authority with domination based on authorization of an agent by a principal; that legal positivists
sometimes equate authority with legitimate domination based on legality; and that Arendt equates
authority with domination rooted in something transcendent, in the legitimacy of tradition or the idea
of foundation. Cf. Hobbes (1996: ch. 16); Kelsen (1967: 56, 194); Arendt (2006: 177, 186–206);
Arendt (1968).
54 E.g. Raz (1999: 38–44, 74–5).
55 I am grateful to Joseph Raz for suggesting this point to me.
that his argument presupposes that people know and accept his specific theory of
rightful authority he writes:
To be sure such an assumption would not be justified. Nor is it made. All I am assuming is
that the service conception of authority is sound, i.e. that it correctly represents our concept
of [rightful] authority. It is not assumed that people believe that it does.56
For Raz, conceptual analysis is not the same as tracing the beliefs of people. He
believes that there is a deep structure of concepts that we share, even if we don’t
know it.
In fancier philosophical terms, the objection can be described as an accusation
that my critique of Raz (and my theory of de facto authority) is incompatible with
semantic externalism. Semantic externalism insists that the meaning of a concept is
not exhausted by the mental understandings of the people who commonly use it;
rather, meaning develops in a linguistic community in its interaction with a certain
reality, and is best seen as the product of socio-linguistic relations.57 This might
seem to conflict with my emphasis on a plurality of models of de facto authority.
My semantic externalist critic might claim that in our socio-linguistic community
there is only one concept of rightful authority, and, since all de facto authority
“claims” rightful authority, there is only one model of de facto authority as well.
Consequently, if Raz got the concept of rightful authority right (and let us assume,
arguendo, that he did), then he also got the concept of de facto authority right,
without any regard to what people actually believe.
The objection, however, is misconceived. Recognizing the fact of a plurality of
models of de facto authority does not entail granting any preference to individual
understandings over the existence of social and linguistic concepts. The point I wish
to make is thus completely consistent with semantic externalism. The important
thing to note is that we have two concepts here: the moral concept of rightful
authority and the factual concept of de facto authority. Nothing in what I say is
meant to challenge the singularity of either of these concepts, or to argue that their
investigation should be sensitive to the plurality of conceptions individuals have of
them.
If semantic externalists are right (and I think they are), then the fact of disagree-
ment between individuals in a certain community as to the content of the moral
concept of rightful authority cannot serve as the basis for arguing that there are
different concepts of rightful authority in our community. We may have one
concept, and some people are just wrong in their understanding of this concept.
Something similar can be said about the concept of de facto authority. People
might disagree about what de facto authority is, as, for example, Raz and I disagree.
But this does not mean that there are two different concepts of de facto authority,
one for Raz and one for me. Rather—and this is the claim of my critique—one of
us is wrong and the other is right about a concept that we share as members of a
socio-linguistic community, responding to a similar factual reality.
56 Raz (1994: 194, 204). 57 See Putnam (1975).

88 Arie Rosen
To say that our one concept of de facto authority must account for the factual
multiplicity of models of de facto authority actually present in our community is not
to argue for conceptual pluralism. If, for example, we were to debate the concept of
a chair, we would have been right in demanding that our single concept should be
able to accommodate all the different models of chairs that are in our world. Had a
chairs-philosopher come to us and claimed that the concept of a chair is that of a
blue metallic seat, we would have told him he’s wrong: the concept of a chair must
be wide enough to include brown and green models of chairs, as well as plastic and
wooden models. The same is true of de facto authority. As a concept that describes
a social phenomenon, the concept of de facto authority must be wide enough to
account for the multiplicity of models of de facto authority that we can actually
observe in our social circumstances. It must accommodate, for example, models
that are based on consent and participatory democratic procedures, just to name
two major models of de facto authority prevalent in our culture. Neither of these
models necessarily involves a claim that authoritative utterances play a mediating
role between people and reasons.
There is an undeniable fact of disagreement regarding what rightful authority is;
the question is what we should make of it. Semantic externalism tells us that one
wrong answer would be to say that this proves that there are many concepts of
rightful authority. However, granting that there is a single concept of rightful
authority, I still insist that the concept of de facto authority must accommodate a
multiplicity of relations that is factually prevalent in our social circumstances. If we
have a concept that purports to describe a social reality—as the concept of de facto
authority does—then this concept cannot exclude the multiplicity of forms that
relations of de facto authority actually take in reality; just like a concept of a chair
cannot disregard the multiplicity of forms that chairs take in our world. This is,
however, exactly what Raz is doing.
Here is another way to make the same point. I suggested earlier that the
phenomenon of de facto authority consists of a belief (in the legitimacy of the
directive) that gives a reason to engage in a certain behavior (deference). Crucial to
my point is that the belief here is part of the social reality. It is a fact that we must
describe, not a conception that we should evaluate. The plurality of beliefs matters,
not because they show us that there are multiple concepts of rightful authority, but
because this plurality is part of the social reality that our factual concept of de facto
authority purports to describe. We should not care, in terms of our descriptive
social theory, if people are wrong about what morality demands, as long as their
beliefs inform their practical reasoning and their behavior.
So, models of de facto authority, which can be based on different conceptions
(even wrong conceptions) of rightful authority, can be of different sorts. Some of
these models can entail exclusion of reasons while others do not, and some of them
(arguably) entail a certain understanding of law, while others do not. Raz’s
explication of the concept of de facto authority should account for this multiplicity
of moral conceptions as part of the social reality that it purports to describe. Since it
does not, I believe that this account is inferior to other possible elucidations of the
concept of de facto authority—namely, those that render the concept wide enough
to accommodate at least the primary forms that de facto authority takes in our
world.
The semantic externalist wants to say that the members of our socio-linguistic
community, who respond to the same social and moral reality as we do, have shared
concepts of de facto authority and rightful authority. I do not dispute this. My
claim is that our shared concept of de facto authority purports to describe the
multiplicity of forms that relationships of de facto authority actually take. Our
explication of the concept of de facto authority cannot hope to achieve this aim if it
is tied down to a single account of the moral concept of rightful authority—even if
it is the right one. Moral concepts do not determine factual concepts that purport
to describe social realities (at least not a priori). More generally we can say that
moral theory does not determine descriptive social theory; not if the latter wants to
avoid the normative fallacy. I do not believe that anyone should want to deny this,
especially a legal positivist like Raz.
3. De facto authority and the idea of law

3.1. The practical matrix of our idea of law
Law’s de facto authority is important for contemporary debates in legal philosophy
regarding the idea of law and the relations between law and morality. The
authoritative nature of law—as well as the contingent structures of institutional
authority that can be found in particular legal systems—seems to depend on moral
convictions we share, and which undoubtedly shape our legal practices and their
conceptualization. Now, I am not sure if it is appropriate to say that people are
dominated by law, or that law has authority over people. It might be argued that
this terminology is more suitable to relationships between people, rather than to
relationships between people and inanimate things such as law.58 But as a manner
of speaking, theorists often talk about law’s authority, and when they do they mean
by this something like the following: People defer to the content of law in a similar
way to the manner in which they sometimes defer to the commands of a person,
and this deference is held by them to be justified in an analogous way to the way
obeying others can be justified. To say that law has de facto authority can thus be
broken into the following two assertions: (1) people defer to law; and (2) they
believe this deference to be morally legitimate.
Raz suggests, as we have seen, that law necessarily has de facto authority, and
therefore has to be capable of having rightful authority.59 Raz’s argument, we now
know, involves a normative fallacy; but other than that it calls our attention to
something which strikes me as by and large right and important. Law normally has
de facto authority, and this is significant to the philosophical investigation of it.
This section discusses this significance.
58 Cf. Weber (1968: 14); Foucault (2000: 337). 59 Raz (1994: 199).
90 Arie Rosen
Now, it is one thing to say that law normally has de facto authority, and quite
another to say (like Raz) that law necessarily has de facto authority. The first
assertion is factual, and the other is conceptual. I wish to make only the factual
claim, and then argue for its significance in our investigation of the idea we have of
law. While to say that law necessarily has de facto authority raises a host of difficult
semantic and metaphysical questions, to say that, in the normal case, law has de
facto authority and that this is related to the way in which law is conceptualized in
our culture seems much less controversial. Although this might not reveal the
necessary and sufficient conditions for something to be called law, it can still show
us the way to answer some of the questions that plague our legal philosophy.
The actual de facto authority of law is important to legal philosophy because it
serves as the immediate context for the conceptualization of law. Ideas—all ideas—
do not subsist in isolation from material reality (and from other ideas). Ideas
themselves are the product of human activity, and they are shaped and maintained
in a social, ideological, and material context, which serves as their practical matrix.60
The idea of law is no different. No matter what our idea of law happens to be, it is
set in a practical matrix that necessarily constrains and influences the idea we have.
It is therefore significant that in the practical matrix in which the idea of law is set—
in which it emerged and in which it still subsists—law normally has de facto
authority.61
The claim that, in the practical matrix of our idea of law, law enjoys de facto
authority does not strike me as controversial. Let us examine both parts of this
double claim (regarding deference to law and its perceived legitimacy). The first
part—that people normally defer to law—is hardly controversial when it comes to
our social reality, and I would not like to spend too much time discussing it. If law
is anything it is something that guides our behavior, and especially our social and
economic interactions. Now, I would not like to claim that it is part of the nature of
law that it is deferred to. I would not even like to raise the more tenable claim that
law is always promulgated with the intention that it be obeyed. My only claim is
that in the normal case law is indeed deferred to. Sometimes, some may say, law is
not worthy of deference. Sometimes, others may say, law does not enjoy deference
in practice. Both of these cases might be grey areas in which people may debate
whether the phenomenon referred to is law or not. I am not getting into that.
It should be clear, however, that in the normal case, in the case that shapes and
60 I borrow here from the useful terminology of Ian Hacking. See Hacking (1999: 10). “ ‘Matrix’ . . .
derives from the word for ‘womb,’ but it has acquired a lot of other senses . . . The matrix in which the idea
of the woman refugee is formed is a complex of institutions, advocates, newspaper articles, lawyers, court
decisions, immigration proceedings. Not to mention the material infrastructure, barriers, passports,
uniforms, counters at airports, detention centers, courthouses, holiday camps for refugee children.” Cf.
Berger and Luckmann (1967).
61 Reference to the normal case is not meant to smuggle in conceptual assertions through the back
door. By “the normal case” I mean just that, the normal way in which law is treated by actual people in
actual historical and social settings. The normal case is always an empirical reality, and is always
important for understanding our concepts and ideas. See generally Wittgenstein (1958: }142) “[I]f rule
became exception and exception rule or if both became phenomena of roughly equal frequency—this
would make the normal language-games lose their point.” Cf. Leiter (2007: 168–9).
constrains our common understanding of what law is, law is deferred to by a certain
group of people.
The same is true of the second claim about law’s authority, regarding law’s
perceived relation to rightful authority. It is perhaps possible to imagine a law that is
deferred to on grounds other than belief in the moral legitimacy of such deference
(say, out of fear of sanction alone). If we do not like imagining, we can look at our
history or at the political realities of other cultures, and try to find instances in
which law was deferred to solely on grounds of fear from punishment. I am not sure
whether we would find any such examples—it might be that many people would
prefer to legitimize even the darkest regimes in order to avoid understanding
themselves as being motivated solely by fear—but for the sake of the argument
here I am willing to assume that we will find some such examples. It is in this sense
that de facto authority is not a necessary condition for something being called “law.”
Even if we were to find such examples, they do not constitute the matrix in
which our ideas of law emerged. In the historical context of the development of our
common understanding of law, as well as our philosophy of law and legal theory,
law has constantly been something that people deferred to, and their deference to
law has been widely believed to be morally legitimate. The reason for this is not
metaphysical or conceptual, but practical and evolutionary. Modes of government
that are perceived as illegitimate tend to be factually weaker and more susceptible
to rebellion, disobedience, revolution, and, ultimately, extinction. As Spinoza
observed:
For as long as men act only from fear, they are doing what they are most opposed to doing,
taking no account of the usefulness and the necessity of the action to be done, concerned
only not to incur capital or other punishment. Indeed, they inevitably rejoice at misfortune
or injury to their ruler even when this involves their own considerable misfortune, and they
wish ever ill on him, and bring this about when they can.62
As history shows us, the de facto authority of law has been, and still is, an important
part of the practical matrix in which our idea of law exists. This is evident in the
many justificatory stories we know to tell about the law—stories that were formed
in the course of millennia. Particular episodes of criticism, dissatisfaction, and
disillusion notwithstanding, the idea of law emerged as something that seems to
deserve our obedience and respect.
3.2. The idea of law and our moral convictions

The idea of law is not indifferent to the practical matrix in which it is set. There is
much to be said about this matrix and the way it influences our conceptualization
of law, but here I would like to point only to one such influence, namely, the
establishment of a connection between the ideas people have of law and ideas they
62 Spinoza (2001: 63). Spinoza also explains that an individual can be bound by fear, but can also be
bound “by love” (“ex amore obligatus”), the latter being a sounder basis for political stability (Spinoza
2001: 185–6). Cf. Hart (1994: 202); Weber (1968: 213).
92 Arie Rosen
have about political morality. If the practical matrix of the idea of law is one in
which deference to law is perceived to be legitimate, then the conceptualization of
law in this context must show it as an object deserving of deference. Clearly, our
practices of governing through law are the product of history, and they were shaped
in a manner that will make them acceptable to us.63 People are usually willing to be
governed only by a law that they can accept as legitimate. This, in turn, shapes and
regulates what law in a certain culture is understood to be.64
Consider the many ready-made stories we have for justifying deference to law. In
our contemporary culture alone, law is believed to be worthy of obedience because
it is just, or because it is needed for cooperation, or because it is an act of political
self-government, or because it is related to morality, or because we implicitly
consent to it, or because obedience to it is entailed by what we owe to our fellow
citizens, etc. There are even more stories ready to be told about the reasons we have
to submit to law’s demands if we look into the history of our political thought.
These stories, these schemes of legitimation, constrain and shape what we under-
stand law to be. An idea of law that cannot be justified in our culture will, most
likely, be unsuitable for government and, therefore, not part of our world at all.
It is important to remember that the argument that I am suggesting is not logical
or metaphysical, but contingent and contextual. The only claim I would like to
defend here is that in the actual practical matrix that regulates our conceptualization
of law, deference to law was normally related to some sort of justification that made
it worthy of obedience. In the reality in which our idea of law still functions, law is
related to moral ideas and to specific understandings of political morality.
Some legal positivists would contest this. They think that keeping law distinct
from the question of legitimate deference to law is crucial.65 In an important way,
I do not disagree, as I say nothing here about moral legitimacy. My claim is that the
idea of law is set in the ideological context of the perceived moral grounds for
deferring to law. Take, for example, natural law theories and the idea of law that
they advocate. According to the model of these theories, law makes sense not only
as “a concept of law” in an abstract, context-free theory. It makes sense because it
fits with certain political ideas and certain ideas about morality, which would make
law a viable part of social life as something that is believed to possess rightful
authority.66 The same is true of the legal positivist understanding of law. Legal
63 Cf. Fuller (1965: 1036–7): “[I]n judging what constitutes acceptable ‘manageability’ in a social
task we must take into account not only the degree of control that may be exercised, but the public
expectations that accompany that control. Furthermore, the question is not one of the brute quantum
of control that is possible, but of its mode and manner, or the point in the total situation at which
intentional direction makes itself felt.”
64 Berger and Luckmann elegantly explain the connection between perceived legitimacy and the
construction of meaning in a more general context, not related to law alone: “Legitimation not only
tells the individual why he should perform one action and not another; it also tells him why things are
what they are.” Berger and Luckmann (1967: 93). For a law-related example see Burgess (1992:
194–5) (describing the English Revolution of 1640 as centered on the question of what the law is, the
common law or the civil law that was used to make sense of the king’s authority).
65 See Section 3.3.
66 E.g. Finnis (1980: 260–91). Outside the narrowly defined camp of natural law, compare Fuller
(1965); Dworkin (1986: ch. 3); Waldron (2008: 19–36).

positivists can think that they are talking about law in the abstract, with no relation
to political or moral theory, but they would be wrong in thinking so. Their concept
of law is a perfect fit to certain political understandings, to certain moral theories, to
certain social realities. It is not a coincidence that positivist theories of law emerged
at an age of increasing legislative activity, as ideas of legal stability, predictability,
and self-government were gaining prominence. If it was not for this context, the
idea of law articulated by positivists would not have come into being at all.67
So what should we make of the claims of legal positivists that the concept or the
nature of law is unrelated to conceptions of legitimate deference to it? The best way
to understand such claims, I think, is to see them as trying to articulate ideas we
have of law outside their practical and justificatory context. De-contextualized ideas
about law are not wrong-as-such. They can reflect the phenomenology of law and
try to refine our intuitions about it. However, theories that are adamant in their
rejection of any connection between law and ideas in political morality have two
connected disadvantages. First, they do not take into account the existence of
something that is actually there (i.e. the connection between law and deference
which is perceived as legitimate). In this sense their articulation of ideas about law is
inferior to an articulation that would take this connection into consideration.
Second, some of these theories do not only disregard the context of the ideas
they articulate, but they deny the relevance of this context altogether. They are
therefore saying something that is wrong.
Again, all of this should be understood as a claim about the circumstances of the
conceptualization of law, and not about the necessary and sufficient conditions for
the use of the word “law.” We can take the concept of law out of this practical
matrix, disregard its perceived justification and say: “look at these bees: their nature
makes them participate in God’s eternal law, and our law lets us do the same;” or:
“consider this man: he makes law as if he is king but nobody recognizes him as ruler
or defers to his law;” or: “see these captive people in the camps, governed by their
oppressors: the commands of their tormentors are their law, in the same way that
the enactments of our parliament are our law.” These would be perfectly sensible
uses of the word “law.” They perhaps tell us something about the similarity
between certain situations and the circumstances of politics as we understand
them, accentuating certain features of human governance while disregarding others.
However, they might lead us away from appreciating what we understand as the
idea of law, that is, from the fullness of our common understanding of what law is.
This is so because they take some ideas we have about law and separate them from
the context that constrains them and in which they normally function.
To conclude this point we can say that by arguing that law necessarily has de
facto authority, or that law necessarily “claims” rightful authority, Raz and others
are pointing to something right and important, although in a vague way. I think
that the same intuition can be articulated more clearly by saying that the ideas we
have about law were shaped in a practical matrix in which law enjoyed de facto
67 Cf. Dyzenhaus and Taggart (2007: 162–3).

94 Arie Rosen
authority (that is, in which people deferred to law because they believed that such
deference is legitimate) and this is still the matrix in which these ideas about law
function. This matrix regulates the ideas we have of law and, among other things,
constitutes a connection between the ideas we have of law and ideas we have in
political morality.
3.3. Law’s authority between positivism and non-positivism

It is easy to see why tying legal philosophy to the theory of authority would be
attractive to a legal positivist like Raz. If we emphasize deference to law it seems to
make sense that law can only be positive law. If we go beyond positive law, what is
there to defer to at all? I think that this is the basic intuition expressed in Raz’s
argument in favor of legal positivism. However, as compelling as the argument
might seem, I think that it is wrong. Non-positivist conceptions of law are
compatible with deference to law just as much as positivist conceptions are. They
demand, however, that we defer not only to what people say but also to reason,
or to God’s law, or to morality, or to the values underlying the system. A non-
positivist law can have de facto authority. It is only that the object to which de facto
authority is attributed would be circumscribed in a different manner from the way
legal positivists would define “law.” In short: acknowledging deference to law does
not entail legal positivism.
Legal philosophers thinking through the relationship between law and morality
tend to follow one of two paths. Some philosophers, like Hans Kelsen and
H. L. A. Hart, maintain that the idea of law can be explored and explained without
addressing the morality or the justification of deference to law. Other philosophers,
like Lon Fuller, John Finnis, and Ronald Dworkin, think that the ideas we have of
law are very much related to moral phenomena. Let us consider the two positions,
now that we have established the significance of law’s de facto authority to legal
philosophy.
According to the positivist line Kelsen and Hart share, one should not confuse
the question of the evaluation of legal practices with the question of their nature.68
Using a terminology he might have avoided today, Kelsen famously distinguished
between evaluation and description of law:
[F]rom the standpoint of scientific cognition of positive law, its justification by a moral
order different from the legal order, is irrelevant, because the task of the science of law is not
to approve or disapprove its subject, but to know and describe it.69
Although one might argue that any description necessarily involves a dimension of
evaluation, still it seems that this position can be intelligible. Kelsen argued for a
division of labor between legal philosophy and moral philosophy: while legal
68 Hart (1994: 240). At an earlier stage, before the publication of The Concept of Law, it seemed
that Hart had a more moderate view on the matter. E.g. Hart (1958: 622). On this see Dyzenhaus
and Taggart (2007: 161).
69 Kelsen (1967: 68).
philosophy should provide a theory of the legal practices and the nature of law,
moral philosophy should deal with questions of political morality. Kelsen’s claim,
which for many years was considered a major tenet of legal positivism, was that the
nature of law can be ascertained without considering law’s rightful authority.
This theory is not wholly wrong, but it is not right either. No matter what your
philosophy of law is, whether you are a legal positivist or natural law advocate, a
non-positivist, a soft-positivist or whatever, your articulation of the idea of law
functions in a certain setting of practices and perceived political morality in which
deference to law can make sense. Between the different articulations of the concept
of law put forward by any of the major theories of legal philosophy, there is not a
single one that does not correlate to certain intuitions we have about political
morality. This includes, of course, legal positivism, whose conception of law
relates—however unwittingly—to political ideas we hold dear, such as self-govern-
ment, democracy, and liberty.
Distinguishing themselves from other strands of legal positivism, some later
positivists have come to criticize the de-contextualized theory of law of Kelsen
and Hart.70 Arguing for what law should be, rather than what law already is,
“normative positivists” stress the moral and political context in which the core ideas
of positive law were first articulated, and insist that these ideas regarding law cannot
be understood without the theories that render this law legitimate and desirable.71
These positivist theorists have revived an earlier strand of legal positivism, which
saw law as an element embedded in the context of political reality, and not as an
autonomous phenomenon to be separately examined and investigated.72 Focusing
on prescription rather than description, most normative positivists say little of what
law is in our world. But if they were to claim something about our idea of law, they
would have been going about it the wrong way. One cannot base a descriptive social
theory on a prescriptive normative one.
Natural law theories suggest that law must have rightful authority that would
make deference to it obligatory on moral grounds. Outside the more narrowly
defined natural law tradition, this relation between rightful authority and law
persisted in non-positivism. Dworkin suggests that theories of law should be
understood as interpretive theories of our legal practices. For him this means that
a theory of law should both fit our legal practices and offer the best possible
justification for them.73 His methodology led him to the conclusion that “[a]
conception of law must explain how what it takes to be law provides a general
justification for the exercise of coercive power by the state, a justification that holds
except in special cases when some competing argument is specially powerful.”74
70 E.g. Waldron (2002: 369–74); Postema (1986: 303–35); Campbell (1996).

71 Waldron (2002: 368–9). Waldron (2001). Cf. Hart and Sacks (1994: 109) “[T]he proposition
that settled law should be respected, until it is duly changed . . . is itself an ethical concept, resting on
the recognition that defiance of institutional settlements touches or may touch the very foundations of
civil order, and that without civil order, morality and justice in anybody’s view of them are impossible.”
72 See Waldron (2002: 369); Dyzenhaus (2004); Dyzenhaus and Taggart (2007).
73 Dworkin (1986: 66–7). 74 Dworkin (1986: 190).
96 Arie Rosen
Dworkin believes that the non-positivist idea of “law as integrity” provides the best
interpretation of our legal practices, according to these criteria.
Dworkin is not completely clear in explaining why theories of law must put legal
practices in the best possible moral light.75 The analogy he offers from art76 seems
to constitute a doubtful argument, and is open to easy attacks. One might rightfully
question a theory that claims that we should always do our best to justify our legal
practices.77 Indeed, Dworkin does not require only that our theory of law will be
able to make the multiplicity of our legal practices intelligible, but also that it will
be able to justify it. In his leap from understanding to justification Dworkin seems
to be moving away from the foundation his conceptual methodology has in
philosophies of language,78 science,79 and hermeneutics.80 For us, however, it is
clear that the intuition Dworkin expresses is pointing in an understandable direc-
tion: he is right to acknowledge that the idea we have of law is related to our legal
reality and to the moral convictions underpinning this reality. What he fails to see is
that this relation is to perceived moral justification rather than to a universal,
objective one. Law should be seen as justified not because this is the a priori
“point” of legal practices or because law is inherently related to (the one true)
morality, but because our common understanding of law was shaped in reference to
something that normally enjoyed de facto authority in our culture.
Non-positivism and the normative strand of legal positivism both offer theories
of law that include morally acceptable reasons for deferring to law. They are both,
however, insensitive to the possibility that our ideas of law are not regulated by the
context of their particular theories of political morality. Evaluating the conclusions
of these philosophers would fall outside the scope of this chapter. But we can see
that their methodology is flawed. Non-positivists argue that the ideas we have of
law are very much related to morality, legitimacy, and authority. But when they
speak of morality, legitimacy, and authority they speak of universal morality, moral
legitimacy, and rightful authority. This leads them to a dangerous move, analogous
to the one I have argued against in section 2. Making their way to ideas about law
through the exploration of moral theory, they presume—either implicitly or
explicitly—that our ideas of law have to be related to the specific theory of political
morality they have come up with. This presumption is unwarranted.
Looking into the de facto authority of law is different from inquiring into law’s
rightful authority. It produces a theory which aims at reporting the discourse of
morality rather than at engaging in the moral discussion itself. This must be the
right way for descriptive social theory, understood as a theory of social phenomena
that cannot be investigated and determined by moral philosophy. Investigating
law’s de facto authority means suspending our moral judgment and accepting
conceptions of political morality that are different from our own as relevant to
75 Cf. Schauer (1994: 504); Murphy (2005: 8); Marmor (2005: 39–43).
76 Dworkin (1986: 59–62).
77 E.g. Unger (1996: 46–50); Leiter (2007: 158–9, 165–6).
78 See Dworkin (1986: 53, 421 n. 3).
79 Dworkin (1986: 53, 421 n. 4).
80 Dworkin (1986: 55, 62, 419–20 n. 2).
our philosophical inquiry, as long as they are influential in the culture we wish to
investigate.
3.4. What law’s authority can and cannot do for legal philosophy
Here are some questions that I think that we should keep apart, along with the
answers I suggest to them:
Is our idea of law shaped and formed in the context of our moral convictions?
- Yes, of course.
Is our idea of law related to morality proper?
- Maybe, but this would only be the case if our contingent moral convictions
happen to correspond to the true precepts of morality.
Does answering the first question in the affirmative mean that there is a necessary
connection between the content of law and the content of our moral convictions?
- No. Normative legal positivists have shown us that it does not.
Does answering the first question in the affirmative mean that there is a necessary
connection between the content of law and the content of (true, objective, universal)
morality?
- No, definitely not.
Talking about the authority of law does not yield a neat philosophical argument for
any one conception of law. I have tried to show that there can be no direct
argument from the very idea of de facto authority of law in favor of a legal positivist
concept of law. This is so because the idea of legitimate deference can imply
deference to natural law, positive law, non-positivist law, etc. I have also tried to
show that there is no clear argument in favor of non-positivism based on law’s de
facto authority. First, there is the counterexample of normative legal positivism,
which puts forward a plausible theory of the rightful authority of a purely positive
law. Second, we are yet to see a non-positivist account of law that is methodologic-
ally committed to accounting for the multiplicity of conceptions of political
morality prevalent in our culture.
In other words, the fact that the idea of law is related to our moral convictions
does not, by itself, determine the relations between the content of law and morality
or, more generally, the question of what law is in our world. Still, our ideas of law
should be investigated with law’s de facto authority in mind. Insisting on this, while
making sure that we are not held captive by any single conception of law’s rightful
authority, would help us map and theorize the complex and sometimes contradict-
ory ideas we have about law. To be true to the project of ascertaining these ideas
means suspending our personal moral theory and accepting the multiplicity of
legitimizing schemes available in our culture for the justification of deference.
A methodology committed to tracing these different schemes and relating them
to ideas we have about law seems to be much sounder than those currently
employed in legal philosophy.
98 Arie Rosen
References
Arendt, H. (1968). “What Is Authority?,” in Between Past and Future: Eight Exercises in
Political Thought. New York: Penguin Books, 91–141.
Arendt, H. (2006). On Revolution. Revised edition with an Introduction by J. Schell.
London: Penguin Books.
Bentham, J. (1988). A Fragment on Government. New Authoritative Edition by J. H. Burns
and H. L. A. Hart. Cambridge: Cambridge University Press.
Berger, P. L. and Luckmann, T. (1967). The Social Construction of Reality: A Treatise in the
Sociology of Knowledge. New York: Anchor Books.
Burgess, G. (1992). The Politics of the Ancient Constitutions: An Introduction to English
Political Thought, 1603–1642. Pennsylvania: The Pennsylvania State University Press.
Campbell, T. D. (1996). The Legal Theory of Ethical Positivism. Aldershot: Ashgate
Dartmouth.
Campbell, T. D. (1970). “The Normative Fallacy,” Philosophical Quarterly, 20: 368–77.
Cohen, J. (1986). “An Epistemic Conception of Democracy,” Ethics, 97: 26–38.
Coleman, J. (2011). “The Architecture of Jurisprudence,” Yale Law Journal, 121: 2–80.
Dworkin, R. (2002). “Book Review: Thirty Years On,” Harvard Law Review, 115:
1655–87.
Dyzenhaus D. (2004). “The Genealogy of Legal Positivism,” Oxford Journal of Legal Studies,
24: 39–66.
Dyzenhaus, D. and Taggart, M. (2007). “Reasoned Decisions and Legal Theory,” in
D. E. Edlin (ed.), Common Law Theory. Cambridge: Cambridge University Press,
134–67.
Estlund, D. M. (1997). “Beyond Fairness and Deliberation: The Epistemic Dimension of
Democratic Authority,” in J. Bohman and W. Rehg (eds.), Deliberative Democracy: Essays
on Reason and Politics. Cambridge, MA: MIT Press, 173–204.
Finnis, J. (1980). Natural Law and Natural Rights. Oxford and New York: Oxford
University Press.
Foucault, M. (2000). “The Subject and Power,” in J. D. Faubion (ed.), Power. New York:
The New Press, 326–48.
Fuller, L. L. (1965). “Irrigation and Tyranny,” Stanford Law Review, 17: 1021–42.
Giddens, A. (1971). Capitalism and Modern Social Theory: An Analysis of the Writings of
Marx, Durkheim and Max Weber. Cambridge: Cambridge University Press.
Habermas, J. (1998). Between Facts and Norms: Contribution to a Discourse Theory of Law
and Democracy. Translated by W. Rehg. Cambridge, MA: MIT Press.
Hacking, I. (1999). The Social Construction of What? Cambridge, MA and London: Harvard
University Press.
Hamilton, A., Madison, J., and Jay, J. (1999). The Federalist Papers. New York: Penguin.
Hart, H. M. Jr. and Sacks, A. M. (1994). The Legal Process: Basic Problems in the Making and
Application of Law. Edited by W. N. Eskridge Jr. and P. P. Frickey. Westbury, NY: The
Foundation Press.
Hart, H. L. A. (1958). “Positivism and the Separation of Law and Morals,” Harvard Law
Review, 71: 593–629.
Hart, H. L. A. (1994). The Concept of Law. Second edition. New York: Oxford University
Press.
Himma, K. E. (2007). “Revisiting Raz: Inclusive Positivism and the Concept of Authority,”
APA Newsletter, 6(2): 20–7.
Hobbes, T. (1996). Leviathan. Revised student edition, edited by R. Tuck. Cambridge:
Kant, I. (1983). “To Perpetual Peace: A Philosophical Sketch,” in Perpetual Peace and Other
Essays. Translated by Ted Humphrey. Indianapolis and Cambridge: Hackett Publishing,
107–39.
Kant, I. (1996). The Metaphysics of Morals. Translated by Mary Gregor. Cambridge:
Kelsen, H. (1967). Pure Theory of Law. Translated by Max Knight. Berkeley and Los
Angeles: University of California Press.
Leiter, B. (2007). Naturalizing Jurisprudence: Essays on American Legal Realism and
Naturalism in Legal Philosophy. Oxford and New York: Oxford University Press.
Lieberman, D. (2008). “Bentham’s Democracy,” Oxford Journal of Legal Studies, 28:
605–26.
Locke, J. (2008). Second Treatise of Government. Thousand Oaks, CA: BN Publishing.
Marmor, A. (2005). Interpretation and Legal Theory. Second edition. Oxford and Portland:
Oxford University Press.
Marmor, A. (2011). “An Institutional Conception of Authority,” Philosophy and Public
Affairs, 39: 238–61.
Mill, J. (1992). “Essay on Government,” in Political Writings. Edited by Terence Ball.
Cambridge and New York: Cambridge University Press, 1–42.
Murphy, L. (2005). “Concepts of Law,” Australian Journal of Legal Philosophy, 30: 1–18.
Perry, S. R. (2007). “Two Problems of Political Authority,” APA Newsletter, 6(2): 31–7.
Plato (1892). “Crito,” in The Dialogues of Plato, vol. 2. Third edition. Translated by
B. Jowett. London: Oxford University Press, 137–56.
Postema, G. J. (1986). Bentham and the Common Law Tradition. Oxford: Clarendon Press.
Putnam, H. (1975). “The Meaning of Meaning,” in Philosophical Papers, vol. 2: Mind,
Language and Reality. Cambridge: Cambridge University Press, 215–71.
Raz, J. (1979). The Authority of Law: Essays in Law and Morality. Oxford: Clarendon Press.
Raz, J. (1986). The Morality of Freedom. Oxford and New York: Oxford University Press.
Raz, J. (1990). “Introduction,” in J. Raz (ed.), Authority. New York: New York University
Press, 1–19.
Raz, J. (1994). “Authority, Law, and Morality,” in Ethics in the Public Domain: Essays in the
Morality of Law and Politics. Oxford and New York: Oxford University Press, 194–222.
Raz, J. (1999). Practical Reason and Norms. Second edition. Oxford: Oxford University
Press.
Raz, J. (2006). “The Problem of Authority: Revisiting the Service Conception,” Minnesota
Law Review, 90: 1003–44.
Rousseau, J.-J. (1987). On the Social Contract, in The Basic Political Writings. Translated by
Donald A. Cress. Indianapolis: Hackett Publishing Company, 139–227.
Schauer, F. (1994). “Critical Notice of Roger Shiner’s ‘Norm and Nature: The Movements
of Legal Thought’,” Canadian Journal of Philosophy, 24: 495–509.
100 Arie Rosen
Schmitt, C. (1996). The Concept of the Political. Translated by G. Schwab. Chicago:
University of Chicago Press.
Shapiro, S. J. (2002). “Authority,” in J. Coleman and S. Shapiro (eds.), The Oxford
Handbook of Jurisprudence and Philosophy of Law. Oxford, Oxford University Press,
382–439.
Simmons, A. J. (1999). “Justification and Legitimacy,” Ethics, 109: 739–71.
Spinoza, B. (2001). Theological-Political Treatise. Second edition. Translated by Samuel
Shirley. Indianapolis and Cambridge: Hackett.
Unger, R. M. (1996). What Should Legal Analysis Become? London and New York: Verso.
Waldron, J. (1996). “Kant’s Legal Positivism,” Harvard Law Review, 109: 1535–66.
Waldron, J. (1999). Law and Disagreement. Oxford: Oxford University Press.
Waldron, J. (2001). “Normative (or Ethical) Positivism,” in J. Coleman (ed.), Hart’s
Postscript: Essays on the Postscript to the Concept of Law. Oxford and New York: Oxford
Waldron, J. (2002). “Legal and Political Philosophy,” in J. Coleman and S. Shapiro (eds.),
The Oxford Handbook of Jurisprudence and Philosophy of Law. Oxford and New York:
Waldron, J. (2008). “The Concept and the Rule of Law,” Georgia Law Review 43: 1–61.
Weber, M. (1968). Economy and Society: An Outline of Interpretive Sociology. Edited by
G. Roth and C. Wittich. Translated by E. Fischoff et al. Berkeley and Los Angeles:
University of California Press.
Wittgenstein, L. (1958). Philosophical Investigations. Second edition. Translated by
G. E. M. Anscombe. Oxford: Basil Blackwell.
5
The Problem about the Nature of Law
vis-à-vis Legal Rationality Revisited
Towards an Integrative Jurisprudence*
Imer B. Flores (UNAM)
[J]urists are still searching for their definition of law.

Immanuel Kant, Critique of Pure Reason.1
I
Reflecting about the nature of law from a contemporary perspective is the main aim
of this volume and in this chapter I will be revisiting questions such as: Does law
have a nature or even an essence? Can it be defined or explained? If so, which and of
what type are the characteristics that define or explain it? Can they be merely
descriptive, evaluative, morally evaluative, or some combination of these? Does the
concept of law correspond to a natural kind, or is it a criterial or an interpretive
concept? Is it a moral concept that necessarily evaluates, justifies, and even ascribes
moral importance or value to its proper object? What is legal rationality? What are
the levels, spheres, or types of legal rationality, and can we integrate them into a
complex legal rationality? What, if any, are the relationships between law and legal
* Revised version of a paper presented at McMaster University Philosophy of Law Conference “The
Nature of Law: Contemporary Perspectives,” May 13–15, 2011 in Hamilton, Ontario (Canada). I am
grateful to Chad McCracken for his interesting and not exasperating public commentary on the paper;
to Adam Riggio, Natalie Stoljar, and Wilfrid J. Waluchow for their challenging questions at the
McMaster presentation; and to colleagues, friends, and the public at the McMaster University
conference for their (morally) important or valuable support; to Edgar Aguilera, Enrique Cáceres,
Ana Conover, Mariana Treviño, and Juan Vega for helpful criticism of a previous draft; and, especially,
to Wil Waluchow for looking very carefully and critically at the different versions of this chapter. But
whatever errors remain are still mine.
1 This quote belongs to a footnote from Immanuel Kant’s Critique of Pure Reason, II. Method of
Trascendentalism, Chapter I “The Discipline of Pure Reason,” Section I “The Discipline of Pure
Reason in its Dogmatical Use,” and has different versions e.g. Kant (1896/1781): 58: “Jurists are
without a definition of right to the present day.”
102 Imer B. Flores
rationality? What are the lessons to be learned from the complex nature of legal
rationality and do they extend to law itself?
In section II, I begin my exploration of these important questions by following
Frederick Schauer’s2 lead in arguing that an exclusive focus on purportedly sufficient
and/or necessary features or properties of law leads to a false or impoverished picture
of its nature. We should instead, I argue, direct our attention to features or
properties that represent non-necessary but nevertheless important or valuable
aspects of the phenomenon we seek to understand. Actually, I intend to go one
step further and suggest that the concept of law—like the concept of
legal rationality—is, as Ronald Dworkin3 has argued, an interpretive concept that
designates a practice, which is not merely important or valuable but morally
important or valuable. It is a concept through which we evaluate, justify, and
even ascribe moral importance and value to what we identify as law. In section
III, I show how the views defended herein were anticipated by Oliver Wendell
Holmes Jr. who, by discussing the at least important or valuable as opposed to the
sufficient and/or necessary, successfully integrated both logic and experience into
law, along with a pluralistic methodological approach to legal rationality. In so
doing, Holmes was able to generate an enlightening theoretical and even practical
description and explanation of various legal phenomena.4 In section IV, I sketch an
analysis of the concept of legal rationality by pointing to the different levels,
spheres, or types of legal rationality, and consider whether all, or some combination
of them, are, if not necessary, at least important or valuable and even morally
important or valuable. And, finally, in section V, I present a general, very tentative
conclusion regarding the nature of law vis-à-vis legal rationality.
II
First, in May 2010 at a Conference in Girona (Spain) and, later, in May 2011 at
a Conference in Hamilton, Ontario (Canada), Frederick Schauer presented differ-
ent versions of a paper entitled “Necessity, Importance, and the Nature of Law”.5
In his paper and presentations, Schauer argued convincingly that sometimes
the nature of a phenomenon consists of those features or properties that are
important but not necessary. Furthermore, he claimed that there are features or
properties that are not necessary to law, but which overwhelmingly exist in actual
legal systems and are more concentrated in them than in other social structures and
institutions. In a nutshell, Schauer criticizes the exclusive focus on features or
properties that are held to be necessary—or essential—to law. On the contrary,
he advocates for the identification of important features or properties of law—
probabilistically concentrated in law but not necessary to it—that can also advance
our understanding of this social practice as it is lived and experienced. By way of
2 Schauer (2010). 3 Dworkin (1978, 1985, 1986, 2006, and 2011).

4 Holmes (1997/1897). 5 Schauer (2010).
The Nature of Law vis-à-vis Legal Rationality Revisited 103
analogy, Schauer cites the case of birds, which form a natural kind and necessarily
have feathers and a backbone, but do not necessarily fly. Despite this, he claims,
flying is an important feature of (almost all) birds, with the exception of penguins,
emus, and ostriches, to the extent that the only flying non-bird vertebrates are bats.
And so any helpful understanding of what it is to be a bird, must make some
reference to the capacity for flight.
I find Schauer’s argument not only breathtaking and challenging but also
inspiring for two main reasons: First, it moves us away from the strong ten-
dency—perhaps even the obsession—with the hunt for necessary and sufficient
conditions,6 usually associated with “natural kinds”7 but that arguably can be
applied to social artifacts and constructions, such as “law.”8 Second, it shifts the
target of our search away from the necessary—or even essential—properties to the
important or valuable but not necessary ones. These can be profitably used not only
to explain “natural kinds,” such as birds, but “social kinds” or even “functional
kinds”9 as well, including “law.”10
Let me make explicit that by focusing not on the necessary or essential, but on
the important or valuable, including the morally important or valuable, we can
rebuild a bridge reconnecting positivists with the rest of legal scholarship: on the
one hand, with non-positivists such as Ronald Dworkin; and, on the other hand,
with legal realists and members of the critical legal studies movement. The latter
group includes those who pursue sociological jurisprudence, as well as the legal
process school, starting with Oliver Wendell Holmes Jr. and ranging from Jerome
Frank and Karl N. Llewellyn to Duncan Kennedy and Roberto M. Unger and from
Benjamin N. Cardozo and Roscoe Pound to Lon L. Fuller, Henry Hart and Albert
Sachs, Harold Lasswell and Myres McDougal, among others. So, instead of one
side affirming fervently that something is necessary and the other side denying it
fervidly, or vice versa, both sides can argue—and might even agree—that some
feature of law is at least important or valuable, including morally important or
valuable, thus advancing our understanding of law rather than hindering it.
With the foregoing in mind, we will bracket discussion of whether or not certain
features of law are necessary, sufficient, or essential, and will instead consider
whether they are at least important or valuable, including some which may well
turn out to be morally important or valuable. Anyway, let me insinuate that analysis
from-the-important-to-the-necessary rather than from-the-necessary-to-the-important is
a more promising route towards advancing our understanding of law: if we begin
sifting the necessary and later proceed to the important we might leave something
vitally important—but apparently not necessary—behind, e.g. the fact that law
almost inevitably, and for very good reasons, resorts to coercion or sanction: “The
6 H. L. A. Hart—as pointed out by Schauer—appeared to be skeptical of recourse to necessary and
sufficient conditions for law, but did refer to them at least once: “There are two minimum conditions
necessary and sufficient for the existence of a legal system.” Hart (1961: 113) and (1994: 116).
7 Quine (1969: 114–38); Putnam (1975); and Kripke (1980).
8 Stavropoulos (1996).
9 Fodor (1975).
10 Moore (1992: 188–242); and Ehrenberg (2009: 91–113).
104 Imer B. Flores
fact that the sanction cannot be an account of law’s distinctive normativity does
not preclude it from being the best explanation of why most individuals comply
with the law’s reasons most of the time.”11 Further still, if we start with something
important but apparently not necessary, we might well end up with something that is
in fact necessary, or at least significantly and substantively quite important, such as
the justification of coercion or the institutionalization of sanction: “[L]egal argument
takes place on a plateau of rough consensus that if law exists it provides a justification
for the use of collective power against individual citizens or groups.”12
As the reader may have anticipated, I wish to defend a two-tier thesis: First, that
there are features in law and legal rationality which are important or valuable, some
morally important or valuable; and, second, that at least some of these features, after
due deliberation on their (moral) importance and value, may turn out to be
necessary as well—probably, not with respect to each and every part of law, but
with respect to law as a whole. For example, the idea that law provides a justification
for state coercion or represents the institutionalization of collective sanction can be
helpful in providing a tentative definition of law. It can do so by helping us draw
lines between law and other normative systems such as morals, religion, social rules
of etiquette, or technical rules, following the definitio fit per genus proximum et
differentiam specificam methodology. Moreover, even if we agree with Hart (1961:
13) that a “definition . . . is primarily a matter of drawing lines or distinguishing
between one kind of thing and another, which language marks off by a separate
word”), sometimes there are neither clear nor precise limits for drawing such lines
or distinguishing one object from the next, as Austin admitted: “But positive law
(or law, simply and strictly so called) is often confounded with objects to which it is
related by resemblance, and with objects to which it is related in the way of analogy:
with objects which are also signified, properly and improperly, by the large and vague
expression law.”13 Since precise boundaries among these different phenomena do
not seem to exist in reality ready to be discovered, but are traced by users of the
concepts through which they are conceived, law appears to be indeed nothing more
than a “family resemblance concept,”14 or perhaps a “cluster concept”15 ranging
over similar but not identical phenomena, phenomena which are subject to change
or variation from time to time and from place to place.16
Actually, the concept of law may be not only a “cluster” or “family resemblance”
concept but also an “essentially contested concept”17 as Dworkin has suggested.
That is, the concept of law may well be a contested concept, which essentially
11 Coleman (2001: 72, n. 12).

12 Dworkin (1986: 108–9).
13 Austin (1998: 9–10.)
14 Wittgenstein (1953).
15 Black (1954).
16 Rudolf Stammler (1925)—one of the most important German legal philosophers of the Weimar
Republic—made an analogous claim regarding not the concept but the content of (natural) law.
Cf. Fuller (1968: 91–110) and (1969: 33–94).
17 Gallie (1965: 167–83).
admits the possibility of competing conceptions or constructions.18 Because he

views the concept of law as an essentially contested one, Dworkin has been adamant
in his critique of semantic theories of law, which inevitably fail due to what he labels
as “the semantic sting.” In short, he challenges the thesis that the concept of law is
“criterial,” i.e. that the very meaning of the word “law” makes the existence of law
depend on satisfaction of certain necessary and sufficient “criteria;”19 and he further
challenges the thesis that the concept of law is a “natural kind concept.”20 Instead,
Dworkin argues, it is important to recognize that the concept of “law” is an
essentially contested “interpretive concept,” one that seeks to place legal practice
in its best moral light.21 In his own voice: “There is a better alternative: propos-
itions of law are not merely descriptive of legal history, in a straightforward way, nor
are they simply evaluative in some way divorced from legal history. They are
interpretive of legal history, which combines elements of both description and
evaluation but is different from both.”22
Regarding criterial concepts, Dworkin has accentuated that “People share some
concepts only when they agree on a definition—rough or precise—that sets out the
criteria for the correct application of the associated term or phrase.”23 For that
purpose, he brings to mind some examples to point out that there are not only
precise criterial concepts, such as an “equilateral triangle” (a triangle necessarily has
sides of equal length) and reasonably precise criterial concepts, such as “bachelor-
hood” (a bachelor is an unmarried, adult human male—but is the Pope a bach-
elor?), but also much less precise or moderately imprecise criterial concepts, such as
“marriage”: “[W]e call many different forms of legal and social arrangements found
in different societies marriages . . . It is a mistake to say, for example, as many now
do, that the essence of marriage is a union between a man and a woman so that gay
marriage is an oxymoron.”24
Regarding interpretive concepts, Dworkin has emphasized that “Some of our
concepts function differently still: they function as interpretive concepts that
encourage us to reflect on and contest what some practice we have constructed
requires.”25 In explaining what he means, Dworkin cites the case of boxing:
People in the boxing world share the concept of winning a round even though they often
disagree about who has won a particular round and about what concrete criteria should be
used in deciding that question. Each of them understands that the answers to these
questions turn on the best interpretation of the rules, conventions, expectations, and
18 Dworkin (1978: 103). In Law’s Empire, Dworkin not only confronts legal positivism with both
natural law theories and legal realism. He also develops three rival conceptions, which he labels
“conventionalism,” “legal pragmatism,” and (his own) “law as integrity” (1986: 33–7, and 94–6).
Elsewhere I have suggested not only that there are three main conceptions of law—“legal positivism,”
“legal realism,” and “legal idealism” (i.e. “natural law theory” or at least some form consistent with it)—
but also that these conceptions qua constructions can be integrated into a more comprehensive one, i.e.
an integrative jurisprudence. Flores (1997: 1001–36).
19 Dworkin (1986: 31–44); (2006: 9–12); and (2011: 158–9).
20 Dworkin (2006: 10) and (2011: 159).
21 Dworkin (1985: 146–8); (1986: 45–96); (2006: 10–12); and (2011: 160–3, 403–5).
22 Dworkin (1985: 147). 23 Dworkin (2006: 9). 24 Dworkin (2006: 9).
25 Dworkin (2006: 10).
106 Imer B. Flores
other phenomena of boxing and of how all these are best brought to bear in making that
decision on a particular occasion.26
He later affirms that the central concepts of political and personal morality function
as interpretive concepts as well.27 In Justice for Hedgehogs, Dworkin writes:
People participate in social practices in which they treat certain concepts as identifying a
value or disvalue but disagree about how that value should be characterized or identified.
The concept of justice and other moral concepts work in that way for us. We agree—
mainly—that these are values, but we do not agree about the precise character of these
values. We do not agree about what makes an act just or unjust, right or wrong, an invasion
of liberty or an act of tactlessness. Nor do we agree about what response, if any, would be
required or justified by a correct attribution of the concept. But we agree sufficiently about
what we take to be paradigm instances of the concept, and paradigm cases of appropriate
reactions to those instances, to permit us to argue, in a way intelligible to others who share
the concept with us, that a particular characterization of the value or disvalue best justifies
these shared paradigms.28
In a few words, Dworkin rejects the thesis that the concept of law is a criterial concept
that can be explicated through some neutral analysis that makes no assumption about
the importance and value of its object. On the contrary, he argues, the concept of law
is interpretive, identifying law as not merely important or valuable, but morally
important or valuable.29 Following Dworkin, I reject the idea that analysis of the
concept of law can be merely descriptive and morally neutral.30 Nor do I believe that
it can be in some way “evaluative” yet also morally neutral,31 since analysis of the
concept of law is, as Dworkin might say, a matter of interpretation.
I agree with Julie Dickson that “Jurisprudential theories must not merely tell us
truths, but must tell us truths which illuminate that which is most important about
and characteristic of the phenomena under investigation.”32 However, I disagree
with Dickson on whether there is a clear-cut distinction between directly and
indirectly evaluative propositions.33 Even though “X is good” is an example of
the former and “X is important” an instance of the latter, nothing precludes the
possibility of propositions taking the form “X is important because X is good”
which are nothing but interpretations of its moral importance or value. In that
sense, in my opinion, taking sides with Dworkin rather than with Dickson, once we
engage in an evaluative and even interpretive practice there may well be no point of
return. If we start discussing important or valuable things we usually end up
discussing morally important or morally valuable ones. Let me clarify, the fact
that we might end up with something morally important or valuable does not mean
26 Dworkin (2006: 10–11). 27 Dworkin (2006: 11).

28 Dworkin (2011: 160–1).
29 Dworkin (2011: 403–5). In a similar fashion, Dworkin suggests that democracy is not a criterial
concept but an interpretive one (2011: 347–50 and 379–99). See also Flores (2010: 84–90).
30 Hart (1994: 239–44); Waluchow (1994: 19–29); Coleman (2001: 175–9); and Marmor (2001:
153–9), (2006: 683–704).

31 Raz (1994: 195–209); and Dickson (2001: 15–25). 32 Dickson (2001: 25).
33 Dickson (2001: 55–7).
that everything we identify as important is—or has to be—(necessarily) morally

important or valuable or even morally necessary. It may instead be important or
valuable for some other reason, such as a theoretical one.
To reinforce this Dworkin-inspired conception of legal theorizing, let me point
briefly to Immanuel Kant’s Critique of Pure Reason and recall, apropos of defin-
itions, that for him “To define . . . means only to represent the complete concept of
a thing within its limits and its primary character.” In clarifying what he means,
Kant states that “Completeness means clearness and sufficiency of predicates; limits
mean precision, no more predicates being given than belong to the complete
concept; in its primary character means that the determination of these limits is
not derived from anything else, and therefore in need of any proof, because this
would render the so-called definition incapable of standing at the head of the
judgments regarding its object.” And, he goes so far as to suggest that “an empirical
concept cannot be defined, but can be explained only.”34 For that purpose, he
further cautions that “The German language has but one word Erklärung (literally
clearing up) for the terms exposition, explication, declaration, and definition; and we
must not therefore be too strict in our demands, when denying to the different
kinds of a philosophical clearing up the honourable name of definition.” He also
draws attention to important differences between mathematical and philosophical
definitions: “[P]hilosophical definitions are possible only as expositions of given
concepts, mathematical definitions as constructions of concepts, originally
framed by ourselves, the former therefore analytically (where completeness is
never apodictically certain), [and] the latter synthetically. Mathematical definitions
make the concept, philosophical definitions explain it only.”35 And he, finally,
concludes in a lengthy paragraph:
Hence it follows . . . [t]hat we must not try in philosophy to imitate mathematics by
beginning with definitions, except it be by way of experiment. For as they are meant to
be an analysis of given concepts, these concepts themselves, although as yet confused only,
must come first, and the incomplete exposition must precede the complete one, so that we
are able from some characteristics, known to us from an, as yet, incomplete analysis, to infer
many things before we come to a complete exposition, that is, the definition of the concept.
In philosophy, in fact, the definition in its complete clearness ought to conclude rather than
begin our work . . . 36
In sum, whereas in mathematics definitions are the concept—ad esse; in philosophy
definitions explain the concept—ad melius esse.37 If we assume that philosophical
definitions not only provide explanations of concepts such as “law” and “morality”
but also do not constitute the concepts themselves, such definitions cannot be
merely descriptive or simply evaluative. Rather they must be interpretive providing
tentative explanations, explications, and explorations, i.e. conceptions of a concept,
proceeding by way of experiment, to reflect on and contest the different interpret-
ations, to the extent that the worse ones, i.e. those debunked or falsified, must
34 Kant (1896/1781: 584). 35 Kant (1896/1781: 586).

36 Kant (1896/1781: 586–7). 37Kant (1896/1781: 587).
108 Imer B. Flores
surrender to the better ones. All this is—or at least can be—done in a quest for the
best (moral) interpretation and in so doing making its object the best it can be
or alternately providing the best (moral) justification for it, as Dworkin has argued
and maintained throughout all these years.38 In Dworkin’s own voice: “But for all
their abstraction, [theories of law] are constructive interpretations: they try to
show legal practice as a whole in its best light, to achieve equilibrium between
legal practice as they find it and the best justification of that practice.”39 In my
opinion, this is extremely significant and substantive, especially since law as a means
has or serves not one but several important or valuable ends, functions, interests,
principles, purposes, and values, including moral ones, as we will see in sections
III and IV.
III
From my perspective, it was Oliver Wendell Holmes Jr. who best captured the
complexity, not only of law but also of legal rationality, and who was the first to
fashion a sustained critique of legal formalism, with C. C. Langdell as his main target.
It is common knowledge among legal theorists that anti-formalism in general and
legal realism in particular benefited from his maxim: “The life of the law has not been
logic: it has been experience.”40 However, not everyone knows that the origin of
Holmes’s maxim lies in a publication prior to The Common Law in 1881. Contrary
to common belief, the maxim appeared for the first time in January 1880, in a “Book
Notice” to the second edition of A Selection of Cases of the Law of Contracts with a
Summary of the Topics covered by the Cases by C. C. Langdell:
Mr. Langdell’s ideal in the law, the end of all his striving, is the elegantia juris, or logical
integrity of the system as a system. He is perhaps the greatest living theologian. But as a
theologian he is less concerned with his postulates than to show that the conclusions from
them hang together . . . so entirely is he interested in the formal connection of things, or
logic, as distinguished from the feelings which make the content of logic, and which actually
shaped the substance of the law. The life of the law has not been logic: it has been
experience. The seed of every new growth within its sphere has been felt necessity. The
form of continuity has been kept up by reasonings purporting to reduce every thing to a
logical sequence; but that form is nothing but the evening dress which the new-comer puts
on to make itself presentable according to conventional requirements. The important
phenomenon is the man underneath it, not the coat; the justice and reasonableness of a
decision, not its consistency with previously held views.41
To assess the extent to which Holmes is attacking logic, I would like to start by
quoting a reference made by H. L. A. Hart in his famous essay “American
Jurisprudence through English Eyes: The Nightmare and the Noble Dream,”
38 Dworkin (1978: 134); (1985: 146–66, and 167–77); and (1986: 90–6). See also (Waluchow
1994: 15–19).
39 Dworkin (1986: 90). 40 Holmes (1991/1881: 1) 41 Holmes (1880: 234).
regarding the relationship of Holmes with Llewellyn and Frank, whose views
represented the “nightmare;”
Holmes certainly never went to these extremes [represented by Llewellyn and Frank].
Though he proclaimed that judges do and must legislate at certain points, he conceded
that a vast area of statutory law and many firmly established doctrines of the common
law . . . were sufficiently determinate to make it absurd to represent the judge as primarily a
law-maker. So for Holmes the judge’s law making function was “interstitial.” Holmes’s
theory was not a philosophy of “full steam ahead and damn the syllogisms.”42
Hart went on to add:
Perhaps the most misused quotation from any American jurist is Holmes’s observation of
1884 (sic) “[t]he life of the law has not been logic: it has been experience.” This in its context
was a protest against the rationalist superstition (as Holmes thought it) that the historical
development of the law by the courts could be explained as the unfolding of the conse-
quences logically contained in the law in its earlier phases. Judicial change and development
of the law were, Holmes insisted, the expression of judges’ “instinctive preferences and
inarticulate convictions” in response, as he said, to the “felt necessities” of his time. And his
protest was made to secure a conscious recognition by lawyers of the legislative powers of the
courts so that judicial change and readjustment of the law should be made after an explicit
weighing of what he termed “considerations of social advantage.”43
Although Hart tried to minimize Holmes’s frontal attack on logic per se, it is
nevertheless true that, for any anti-formalist, Holmes’s adage has become nothing
short of an anthem. Certainly Holmes’s critique—as Hart acknowledges—is part of
a wider “revolt against formalism” and “is taken as an example of a great reaction
against the excessive reliance on thought that is deductive, formal, abstract, or split
into firmly separated distinct disciplines. The revolt was born of a wish to cross
sterile, arbitrary, academic divisions and to substitute for formalism a vivid, realistic
attention to experience, life, growth, process, context, and function.”44
At this point it is imperative to consider more carefully the context in which
Holmes’s dictum is to be situated. Though Holmes is indeed critical of an exclusive
or excessive reliance on logic, at the expense of a wide variety of other factors that
come into play in judicial reasoning, his aim is not to abolish the use of logic
entirely. His aim is to make clear that logic, though necessary in legal reasoning, is
not sufficient and that other factors are, if not necessary, then at least important or
valuable as well. Let me call to your attention that in the preceding lines of The
Common Law Holmes explains: “The object of this book is to present a general view
of the Common Law. To accomplish the task, other tools are needed besides logic.
It is something to show that the consistency of a system requires a particular result,
but it is not all.”45 And, in the following lines, he adds:
42 Hart (1983a: 128).

43 Hart (1983a: 129–30).
44 Hart (1983a: 130). See also White (1949).
45 Holmes (1991/1881: 3).
110 Imer B. Flores
The felt necessities of the time, the prevalent moral and political theories, intuitions of
public policy, avowed or unconscious, even the prejudices which judges share with their
fellow-men have had a good deal more to do than the syllogism in determining the rules by
which men should be governed. The law embodies the story of a nation’s development
through many centuries, and it cannot be dealt with as if it contained only the axioms and
corollaries of a book of mathematics. In order to know what it is, we must know what it has
been, and what it tends to become.
In his seminal address at Boston University School of Law, on January 8, 1897,
which was published that same year by the Harvard Law Review as “The Path of the
Law,” Holmes—at that time Justice of the Supreme Judicial Court of Massachu-
setts—analyzed not only the limits of law but also the forces which determine its
content and its growth.46 In the spirit of the arguments made in section II, in what
follows I will bracket discussion of whether these various forces, including logic and
morality, are or are not necessary to legal practice. Instead, I will consider whether
they are at least highly important or valuable, including in some cases morally
important or valuable. In furtherance of this objective, I would like to begin by
drawing attention to the following Holmesian claims:
1. “[L]aw, if not a part of morality, is limited by it.”47 In expanding upon this
point, Holmes cautioned against two serious pitfalls: “the danger, both to specula-
tion and to practice, of confounding morality with law, and the trap which legal
language lays for us on that side of our way.”48
2. “The fallacy . . . that the only force at work in the development of law is
logic.” Hence, even though Holmes recognized an important place and role for
logic, he argued that it is not everything:
This mode of thinking is entirely natural. The training of lawyers is a training in logic. The
processes of analogy, discrimination, and deduction are those in which they are most at
home. The language of judicial decision is mainly the language of logic. And the logical
method and form flatter that longing for certainty and for repose which is in every human
mind. But certainty generally is illusion, and repose is not the destiny of man. Behind the
logical form lies a judgment as to the relative worth and importance of competing legislative
grounds, often an inarticulate and unconscious judgment, it is true, and yet the very root
and nerve of the whole proceeding. You can give any conclusion a logical form.49
3. “[J]udges themselves have failed adequately to recognize their duty of
weighing considerations of social advantage.”50 “[A] body of law is more rational
and more civilized when every rule it contains is referred articulately and definitely
to an end which it subserves, and when the grounds for desiring that end are stated
or are ready to be stated in words.”51 In addition, Holmes emphasized, “we are
called on to consider and weigh the ends of legislation, the means of attaining them,
and the cost. We learn that for everything we have to give up something else, and
we are taught to set the advantage we gain against the other advantage we lose, and
46 Holmes (1997/1897: 991–1009). 47 Holmes (1997/1897: 993).

48 Holmes (1997/1897: 997). 49 Holmes (1997/1897: 998).
50 Holmes (1997/1897: 999). 51 Holmes (1997/1897: 1000–1).
to know what we are going to elect.” Accordingly, he proposed “that every lawyer
ought to seek an understanding of economics”52 and went so far as to pronounce
that “For the rational study of the law the black-letter man may be the man of the
present, but the man of the future is the man of statistics and the master of
economics.”53
4. “At present, in very many cases, if we want to know why a rule of law has
taken its particular shape, and more or less if we want to know why it exists at all we
go to tradition . . . The rational study of law is still to a large extent the study of
history. History must be a part of the study.” Moreover, he also warned: “Every-
where the basis of principle is tradition, to such an extent that we even are in danger
of making the rôle of history more important than it is.”54
5. “There is another study which sometimes is undervalued by the practical
minded . . . I mean the study of what is called jurisprudence. Jurisprudence, as
I look at it, is simply law in its most generalized part.”55 This claim concerning
the centrality of jurisprudence to law was, of course, popularized later by Dworkin
in the course of developing his own distinctive theory, “law as integrity.”56 On one
side, Holmes advanced:
The way to gain a liberal view of your subject is not to read something else, but to get to the
bottom of the subject itself. The means of doing that are, in the first place, to follow the
existing body of dogma into its highest generalizations by the help of jurisprudence; next, to
discover from history how it has come to be what it is, and finally, so far as you can, to
consider the ends which the several rules seek to accomplish, the reasons why those ends are
desired, what is given up to gain them, and whether they are worth the price.57
On the other side, Dworkin clarified:
Law as integrity is . . . both the product of and the inspiration for comprehensive interpret-
ation of legal practice. The program it holds out to judges deciding hard cases is essentially,
not just contingently, interpretive; law as integrity asks them to continue interpreting the
same material that it claims to have successfully interpreted itself. It offers itself as continu-
ous with—the initial part of—the more detailed interpretation it recommends. . . .
Law as integrity, then, begins in the present and pursues the past only so far as and in a
way its contemporary focus dictates. It does not aim to recapture, even for present law, the
ideals or practical purposes of the politicians who first created it. It aims rather to justify
what they did (sometimes including . . . what they said) in an overall story worth telling now,
a story with a complex claim: that present practice can be organized by and justified in
principles sufficiently attractive to provide an honorable future.58
52 Holmes (1997/1897: 1005). 53 Holmes (1997/1897: 1001).

54 Holmes (1997/1897: 1003). 55 Holmes (1997/1897: 1005).
56 Dworkin (1986: 225–75). Ever since the appearance of Taking Rights Seriously, Dworkin has
been making the same—or at least a very similar—claim regarding the role of jurisprudence and its
interconnection with problems that certainly “lie beyond the ordinary techniques of the practicing
lawyers” (1978: 1). In short, for him: “Jurisprudence is the general part of adjudication, silent prologue
to any decision at law” (1986: 90).
57 Holmes (1997/1897: 1007).
58 Dworkin (1986: 226–8).
112 Imer B. Flores
Let me insist that Holmes considers that the wide variety of forces to which he
draws our attention, forces such as language and logic, history and tradition,
sociology and philosophy, economics and statistics, ethics and morals, all play, if
not a necessary role, at least an important or valuable role in legal reasoning and
hence in legal rationality.59 Following Dworkin, he might have added that a theory
of law must draw upon some combination of these factors in providing an
evaluation, perhaps even a justification, of legal practice. It is a real danger,
according to Holmes, to present any one of these factors as being more—or
alternately less—important or valuable than it really is. It is no less dangerous to
assume that one or more of them must be sufficient and/or necessary for law when
experience suggests that this is clearly not the case. In a similar vein, Holmes, in his
address “Law in Science and Science in Law,” offered up the following recollection:
I sometimes tell students that the law schools pursue an inspirational combined with a
logical method, that is, the postulates are taken for granted upon authority without inquiry
into their worth, and then logic is used as the only tool to develop the results. It is a necessary
method for the purpose of teaching dogma. But inasmuch as the real justification of a rule of
law, if there be one, is that it helps to bring about a social end which we desire, it is no less
necessary that those who make and develop the law should have those ends articulately in
their minds.60
It is clear that Holmes, in discussing the important or valuable but apparently
neither necessary nor sufficient, attempted to integrate not only logic and experi-
ence into law but also a pluralistic methodological approach in the core of legal
rationality. We are left, then, with two further questions: First, whether Holmes
was suggesting that the various factors to which he draws our attention—all, some
or some combination of them—are morally important or valuable aspects of legal
rationality; and, second, whether an adequate theory of law must provide an
account that evaluates, and perhaps even justifies, the practice of integrating logic
and experience in the exercise of legal reasoning.
Holmes was not alone in embracing the idea of integrating logic and experience
in law and judicial decision making. Consider Roscoe Pound’s Law Finding through
Experience and Reason (1960), where, in his opening remarks, he recalls that three
centuries before Sir Edward Coke, Chief Justice of the Court of Common Pleas
first, and of the King’s Bench later, argued that “Reason is the life of the law, nay
the common law itself is nothing else but reason” and concluded that “law is an
artificial reason”: “an artificial perfection of reason, gotten by long study, observa-
tion, and experience, and not of everyone’s natural reason; for nemo nascitur
artifex.” In the Centennial History of the Harvard Law School, in a part probably
written by Dean Pound, this observation was made:
59 Let me clarify that my claim is not that all of these forces are necessary to law but only that they
play at least an important or valuable role in law. I further claim that, after due deliberation, it may be
the case that they (all, some, or various combinations of them) turn out to be morally important or
valuable—perhaps even necessary—for an adequate explanation of the concept both of law and of legal
rationality.
60 Holmes (1920b: 238).
It has, however, become evident in recent years . . . that the scope of legal study must extend
beyond printed books, certainly beyond law books. Since LAW is not a water-tight
compartment of knowledge but a system of rules for the regulation of human life, the
truth of those rules must be tested by many facts outside the past proceedings of courts and
legislatures.61
The idea of integrating the different methodologies into a pluralistic legal method-
ology is also explicit in Benjamin N. Cardozo’s The Nature of the Judicial Process 62
and its sequel The Growth of the Law.63 In both books, Cardozo referred to four
forces of law and their corresponding methods of legal rationality. In the first
version:
The directive force of a principle may be exerted along the line of logical progression; this
I will call the rule of analogy or the method of philosophy; along the line of historical
development; this I will call the method of evolution; along the line of the customs of the
community; this I will call the method of tradition; along the lines of justice, morals and
social welfare, the mores of the day; and this I will call the method of sociology.64
And, in the second, with minor stylistic changes:
Our fourfold division separates the force of logic or analogy, which gives us the method
of philosophy; the force of history, which gives us the historical method, or the method of
evolution; the force of custom, which yields the method of tradition; and the force of justice,
morals and social welfare, the mores of the day, which its outlet or expression in the method
of sociology. No doubt there is ground for criticism when logic is represented as a method in
opposition to the others. In reality, it is a tool that cannot be ignored by any of them.65
To reinforce the integration not only of logic and experience to law but also of a
pluralistic methodological approach to legal rationality, let me call attention to the
idea of experimentation and to the possibility of reconciling the two competing
needs of fixity and flexibility by quoting a lengthy but admirable paragraph from
Munroe Smith:
In their effort to give the social sense of justice articulate expression in rules and in
principles, the method of the lawfinding experts has always been experimental. The rules
and principles of case law have never been treated as final truths, but as working hypotheses,
continually retested in those great laboratories of the law, the courts of justice. Every new
case is an experiment; and if the accepted rule which seems applicable yields a result which is
felt to be unjust, the rule is reconsidered. It may not be modified at once, for the attempt to
do absolute justice in every single case would make the development and maintenance of
general rules impossible; but if a rule continues to work injustice, it will eventually be
reformulated. The principles themselves are continually retested; for if the rules derived
from a principle do not work well, the principle itself must ultimately be re-examined.66
Holmes, in his dissent in Abrams v. United States, articulated a similar view in
respect of the US constitution, a view that, arguably, is equally applicable to other
areas of law as well. In Holmes’s view, the US constitution “is an experiment, as all
61 Griswold (1967). 62 Cardozo (1921). 63 Cardozo (1924).

64 Cardozo (1921: 30–1). 65 Cardozo (1924: 62). 66 Smith (1909: 21).
114 Imer B. Flores
life is an experiment.”67 And, one year later, in Missouri v. Holland, he emphasized
that we must realize not only that it is impossible to have foreseen completely
everything that will eventually occur. We must also recognize that words or more
precisely their meaning and usage vary throughout time to the extent that “The case
before us must be considered in the light of our whole experience and not merely of
what was said a hundred years ago.”68
In a similar vein, Pound advanced the need for fixity, i.e. certainty, security and
stability, on the one hand, and flexibility, i.e. change, growth, and progress, on the
other:
Law must be stable and yet it cannot stand still. Hence all thinking about law has struggled
to reconcile the conflicting demands of the need of stability and of the need of change. The
social interest in the general security has led men to seek some fixed basis for an absolute
ordering of human action whereby a firm and stable social order might be assured. But
continual changes in the circumstances of social life demand continual new adjustments to
the pressure of other social interests as well as to new modes of endangering security. Thus
the legal order must be flexible as well as stable. It must be overhauled continually and
refitted continually to the changes in the actual life which it is to govern. If we seek
principles, we must seek principles of change no less than principles of stability. Accordingly
the chief problem to which legal thinkers have addressed themselves has been how to
reconcile the idea of a fixed body of law, affording no scope for individual wilfulness,
with the idea of change and growth and making of new law.69
Similarly, Cardozo announced: “The law of our day faces a twofold need. The first
is the need of some restatement that will bring certainty and order out of the
wilderness of precedent. This is the task of legal science. The second is the need
of a philosophy that will mediate between the conflicting claims of stability and
progress, and supply a principle of growth.”70 And, finally, it was H. L. A. Hart
who recapitulated these insights when he famously wrote:
In fact all systems, in different ways, compromise between two social needs: the need for
certain rules which can, over great areas of conduct, safely be applied by private individuals
to themselves without fresh official guidance or weighing up of social issues, and the need to
leave open, for latter settlement by an informed, official choice, issues which can only be
properly appreciated and settled when they arise in a concrete case.71
The possibility of combining through experimentation the two competing needs
for fixity and flexibility,72 indicates not only that logic and experience are integrated
into law but also that at least an analytical-conceptual and a critical-functional
methodological approach are integrated into legal rationality.73 This idea can be
traced to Holmes’s assertion that our notions shrink and at the same time grow
more precise when we wash them with “cynical acid”74 and is implicit not only in
67 Holmes (1919: 630). 68 Holmes (1920a: 433).

69 Pound (1923: 1) and (1960: 23). 70 Cardozo (1924: 1).
71 Hart (1994: 130–1). 72 Waluchow (2007: 183). See also Flores (2009b: 40–6).
73 Jhering (1951: 678–89); Cohen (1935: 809–49); and Hart (1983b: 265–77).
74 Holmes (1997: 995).
Cardozo’s and Pound’s sociological jurisprudence, which interprets “legal insti-

tutions as social processes . . . responding to various sorts of social and personal
stimuli” but also in McDougal and Lasswell, at Yale, and Fuller, Hart and Sachs, at
Harvard, who “insisted on the importance of regarding the law as an instrument for
moving society toward certain larger goals and [ . . . ] tried to settle questions about
the legal process instrumentally, by asking which solutions best advanced these
goals.” Yet, as Dworkin pointed out, both approaches “ended by distorting juris-
prudential issues . . . by eliminating . . . issues of moral principle from their core.”75
In that sense, I am taking seriously Dworkin’s claim throughout all these years that
what a legal theory should provide is an interpretation of law that integrates—or
more precisely reintegrates—not only moral principle but also both morality and
philosophy into its core. In short, I am proposing that we pursue an integrative
jurisprudence.
IV
In this section, my aim is to provide an analysis of legal rationality that draws on the
insights of Holmes and the many others whose views are sketched in section III.
From my point of view, legal rationality comprises five different levels, spheres, or
types, which are, if not necessary, at least important or valuable. It is also my view
that all, some, or a combination of them may turn out to be morally important or
valuable and that, after due deliberation, we may even come to discover that they
are necessary or essential. It is worth mentioning that the same legal rationality is at
play both in adjudication or judicial decision making, and in any other legal
process, including the creation and modification of legislation.76 It is also true
that an adequate theory of law must provide not only a description or explanation
of law but also an evaluation, justification, and adscription of moral importance or
value to the legal rationality which is inherent in legal practice and as such in law
itself.
The five different levels, spheres or types of legal rationality are as follows:
1) Linguistic rationality: it is important or valuable that laws and legal decisions
are clear and precise to avoid the problems of ambiguity and vagueness (R1);
2) Legal-formal—or systematic—rationality: it is important or valuable that laws
and legal decisions are not only valid but also coherent, non-redundant, non-
contradictory, prospective or non-retroactive, and publicized to avoid prob-
lems of antinomies, redundancies, and gaps, while promoting the complete-
ness of law as a system (R2);
3) Teleological rationality: it is important or valuable that laws and legal decisions
are efficacious in serving as a means to an end and in consequence they
cannot establish something impossible or merely symbolic (R3);
75 Dworkin (1978: 4). 76 Flores (2005: 35–8); (2007: 264–6); (2009a: 106–9).
116 Imer B. Flores
4) Pragmatic rationality: it is important or valuable that laws and legal decisions
are not only efficacious, but also socially effective and economically efficient
in achieving certain goals, especially in the case of conflict between two or
more competing ones and their corresponding interpretations (R4); and
5) Ethical rationality: it is not merely important or valuable but morally import-
ant or valuable that laws and legal decisions are fair or just and, as a result, can
neither admit an injustice nor the violation of basic principles and rights,
including moral ones (R5).
It follows from (1) to (5) that our legal operators, including legislators and judges,
ought to know or at least keep in mind: the intricacies of our language and the use
or meaning of words (R1); the details of our existing legal system, its past, present,
and (possibilities for development in the) future (R2); the minutiae of our scheme
of ends, functions, interests, principles, purposes, and values (R3); the ins and outs
of their possible consequences and effects (R4); and, the implications of every single
principle, including moral ones that bears on whatever aspect of legal practice is of
current relevance (R5). In short, my threefold claim is: first, that these five different
levels, spheres, or types are at least important or valuable; second, that one of the
five different levels, spheres, or types is at least morally important or valuable, i.e.
R5; and, third, that it is necessary that the five different levels, spheres, or types be
integrated into a complex but still workable legal rationality, as I will now attempt
to show.
It is important or valuable that a legislator begins by using clear and precise
language to avoid problems related to ambiguities and vagueness (R1) and it is
equally important or valuable to carry on by inquiring about the coherency and
completeness of the legal system to avoid antinomies and gaps (R2). Similarly, it is
important or valuable that the legislator continues by drafting at least one end,
function, interest, principle, purpose, and value into law (R3). But it may be
necessary for the legislator to aim on many occasions to fulfill or honor more
than two ends, functions, interests, principles, purposes, and values, some or each
of which might be subject to more than one interpretation. In such cases it is
important or valuable that the legislator points out some sort of priorities (R4).
Finally, it is not merely important or valuable but morally important or valuable
that the legislator guarantees that an overall justified principle, including a moral
one, is embedded into the law or at least is not violated by it (R5).
By the same token, it is important or valuable that a judge begins by asking about
the clarity and precision of the language used by the drafter (R1); and, only when
the language is neither clear nor precise it is equally important or valuable that he or
she carry on by inquiring about the coherency and completeness of the legal system
(R2). Analogously, when the language and/or legal system appear to be incoherent
or incomplete, it is important or valuable that the judge goes on to decide in
accordance with a relevant end, function, interest, principle, purpose, or value (R3).
But in a case where there are two or more ends, functions, interests, principles,
purposes, or values, and/or two or more interpretations of them that are equally
accessible or available, it is important or valuable that the judge resolves any
resulting conflict by appealing to the better one or to the one more suitable for
advancing, to some degree at least, some of the ends in play in the case at hand (R4).
Finally, when the consequences or effects are illegitimate, it is not merely important
or valuable but morally important or valuable that the judge strive as best he or she
can to secure a result that accords with some overall legitimate principle, including a
moral one (R5).
Let me explain. The fact that it is important or valuable that our legal operators,
including legislators and judges, know and keep in mind all these levels, spheres, or
types of legal rationality and attempt, as best they can, to adhere to their demands,
does not necessarily mean that they accomplish them all the time. Experience
undoubtedly shows that legislatures fail to do so from time to time and that it is
such failures that sometimes open the door to the courts to intervene in an effort to
bring about greater legal rationality. But experience has also undoubtedly shown
that courts can and often do fail to respect the norms of legal rationality as well. In
my opinion, neither the legislator nor the judge qua legal agent is demonstrably
better or more capable than the other at successfully grounding their decisions in
acceptable premises. Quite the contrary: neither one seems better equipped and
situated than the other to fulfill the demands imposed by the five levels, spheres, or
types of legal rationality.
Elsewhere I have suggested that the term “compensation” regarding the eminent
domain or takings doctrine must be understood as requiring that compensation be
“fair” or “just.” Regardless of the agent or agency, i.e. whether it is the legislator, the
judge, or perhaps even a framer, the morally important or valuable thing is to make
explicit the reason or rationality behind a decision concerning fair or just compen-
sation. The identity of the agent who proclaims on such a question is not of the
essence. In that sense, either the legislator, judge, or framer can make the moral
conditions implicit in the term “compensation” clear by explicitly requiring it to be
“fair” or “just” if it is to be truly so. Actually, the framers of the United States
Constitution did just that by explicitly requiring “just compensation” in the Fifth
Amendment, while the framers of the Mexican Constitution failed by merely
requiring “compensation” in Article 27. However, in interpreting such a norm
and its purpose, Mexican judges have routinely relied on an understanding of
“compensation” that requires it necessarily to be fair or just. To be clear, the judge
in such cases is not, in fact, legislating or inventing new law. Rather he or she is
interpreting the principles objectively embedded in the term “compensation,”
which implies the objective criterion that what truly counts as compensation is
something that is necessarily “fair” or “just.”77
Legal rationality is ideally displayed in a progression from level R1 on to level R5.
This progression is not only workable in practice, but also brings into play a kind of
rational—and even predictable—order, as I shall now argue. Let’s begin by noting
that what the legal realists referred to as logic corresponds to the analytical-conceptual
methodological approach implicit in R1 and R2, whereas experience corresponds to
77 Flores (2009a: 99–100).

118 Imer B. Flores
the critical-functional approach implicit in R3, R4, and R5. Now consider the
following example:78 In an airport terminal (we could easily substitute a subway or
train station) we find a sign displaying the following disposition: “It is prohibited to
enter with dogs.” One day someone—call him James—intends to enter the terminal
with his pet grizzly bear—call him Ben—and is prevented from doing so by the
proper (administrative or executive) authority, say a policeman. Since the rationale
for extending the prohibition from dogs to grizzly bears like Ben cannot be found by
merely appealing to R1 or R2 it is important or valuable to look elsewhere for a
justification for widening the scope of the rule so as to cover James entering with Ben.
Let me clarify that, in my opinion, extending the prohibition from dogs to other
animals, like Ben, implies neither adding nor modifying the rule prohibiting dogs.
Rather it requires that the rule be interpreted in a manner such as to widen the
scope of its application to include animals like Ben. I have no difficulty in agreeing
that such widening of the rule is the product of a “creative judicial activity”
resulting from the interpretation of a pre-existing rule. What I wish to dispute,
however, is the claim that the exercise of such judicial creativity amounts either
to the legislative creation of a (new) rule or the quasi-legislative change of an
existing rule.79
In so arguing, I can reasonably be viewed as invoking, as important or valuable
(perhaps morally) among the rules of interpretation, what is commonly known as
the “golden rule.” According to this familiar canon of interpretation, a rule is not to
be interpreted as plain meaning suggests if, applying the rule with that understand-
ing to a given set of circumstances, would lead to an “absurdity” or “moral
repugnance.” Alternatively, I might reasonably be viewed as invoking, as important
or valuable (perhaps morally), what is commonly termed as the “mischief rule,”
according to which if interpreting a rule as its plain meaning suggests leads to a
result which frustrates eradication of the “mischief ” the rule was intended to
remedy (or perhaps realization of the very goal(s) the rule was intended to achieve),
then the rule is not be interpreted that way. In either case, it is important or
valuable to recognize that the rule and the norms of legal rationality do not
necessarily dictate the result that plain meaning suggests.80
With the foregoing in mind, now consider this: it is absurd or perhaps even
morally repugnant, to allow James to bring Ben into the terminal, when someone
with a domestic animal such as a dog is prevented from doing so—simply, on the
ground that bears were not expressly included in the prohibition by the proper
legislative authority. In other words, it would be absurd and perhaps morally
repugnant were the proper administrative or executive authority, i.e. the policeman
in James’s case, bound to apply the rule as it was literally drafted and consequently
78 This is a variant of a well-known example, which is presented in a slightly different form and can
be traced both to Gustav Radbruch and Leon Petrasyski. See Recaséns Siches (1959: 645–7).
79 Flores (2011: 168). See Hart (1961: 131) and (1994: 135).
80 I am grateful to Wil Waluchow for bringing this point to my attention and asking me to be more
explicit about it.

refrain from preventing his entry with Ben into the terminal. It would be no less
absurd or perhaps morally repugnant to require in any case where a decision to
disallow entry has been challenged, that the proper adjudicative authority, say
a judge, decide according to the rule interpreted strictly, i.e. in terms of plain
meaning.81
Let’s now examine this case with an eye towards the progression from R1 to R5
mentioned above. Concerning R1, the species “dog” is simply not the same as the
species “bear” and although both species are part of the genre “animals,” there is no
way to include, linguistically, the concept of “bear” in the concept of “dog,”
especially since they correspond to different natural kinds with their own genotypes
and phenotypes. Regarding R2, we can assume, for the sake of the argument, that
there is not, or at least not yet, a prior piece of legislation or judicial decision or
precedent, which has previously equated dogs, bears, and other animals to the
extent that the prohibition is already clearly equally applicable to them all. On the
contrary, if there is such precedent the justification will be provided by appealing to
R2 itself. If neither R1 nor R2 is helpful in providing the rationale to justify
extending the prohibition from dogs to bears, it is at least important or valuable
to look elsewhere for its justification.
Once R1 and R2 have run out, it is important or valuable to proceed to R3 and
search for ends, functions, interests, principles, purposes, or values that lie behind
the rule. Here we might appeal, at least in some cases, to what the author of the rule
intended to accomplish, i.e. his or her intentions. Alternatively, we might appeal to
what the rule does, in fact, serve to accomplish in order to justify extending the
prohibition from “dogs” to “bears.” By widening the scope of the rule prohibiting
dogs to cover other animals such as bears we might further accomplish these
objectives. However, once we cross the threshold of R3, a critic might claim, we
will almost certainly end up not at stage R4 but R5. Let me address this possible
objection briefly.
Suppose a critic of the view herein defended objects as follows: It is patently clear
that the explicit or express intention of the author of the prohibition was directed
only to dogs and not to bears. Had the drafter wanted to prohibit “bears”—or any
other animals for that matter—he or she would have included them in the prohib-
ition next to “dogs,” i.e. he would have said that “it is prohibited to enter with bears
and dogs and . . . ”. Alternatively, he or she would have introduced a more ample or
generic formula such as “animals,” “animals of/with certain size,” “dangerous
animals,” or “animals that may affect or threaten the users of the terminal.” In
reply, I offer the following response. The fact that the author explicitly mentioned
only dogs merely suggests that he or she had in mind, as the most probable or likely
case, one where someone might attempt to enter with a dog, which is the domestic
animal par excellence. In addition, the fact that the author expressly mentioned
dogs neither excludes nor precludes the possibility of intending to prohibit or at
81 Keep in mind both the maxim of the Digest 50, 17, 1: “Non ex regula ius sumatur, sed ex iure quod
est regula fiat” and Lon L. Fuller’s distinction between “intelligent and unintelligent fidelity to law”
(1999/1949: 1858). See also Fuller (1958: 630–1).
120 Imer B. Flores
least implicitly prohibiting by way of analogy other animals such as bears. The
justification for including bears within the scope of the prohibition derives from the
authentic or genuine wishes and true commitments intended for prohibiting dogs
in the first place.82 Clearly, the prohibition is extended to a bear such as Ben, not
because he is (for some reason) to be considered a dog, which he clearly is not, but
because in line with the legal maxim: ubi eadem ratio iuris, ibi eadem iuris dispositio,
creatures like Ben—and pet dogs like Benji—actually represent a danger or threat
to users of the airport terminal and so should be included within the scope of the
rule. In my opinion, it is very unlikely that the drafter of the prohibition, if
asked, will say that he or she indeed had only dogs in mind or that the prohib-
ition should not be extended to cover the case of bears and other animals. Rather,
he or she is more likely to point to any one of the aforementioned interpretive
maxims—or similar ones—as offering a means by which the interpreter can
decide in a way reflective of his or her actual intentions. Since the norms of
legal rationality include more than one maxim of interpretation available to
justify extending the prohibition from “dogs” to “bears,” widening the scope of
the rule prohibiting dogs from the terminal so as to cover other animals such as
bears is, if not necessary, at least important or valuable. It is, in other words,
important or valuable to recognize that legal actors are sometimes permitted to
move on from R1 and R2 to R3, R4, and even R5 in the search of the better
interpretation.
It might be thought that the need to move in these (and no doubt other) ways
beyond levels R1 and R2 might be better served, had the drafter of the rules
chose his terms more wisely. Perhaps, instead of using “dogs,” the drafter might
have used “animals” or perhaps “animals of a certain size.” But this strategy is not
necessarily as helpful as might initially appear. It will likely remain impossible to
answer all questions of application by remaining at levels R1 and R2. After all, the
expression “animals” brings its own degree of imprecision or vagueness to the
extent that it will appear to be equally applicable not only to dogs and bears but
also to a wide range of different animals from insects, such as ants, ladybugs, and
spiders, on to the largest land and sea mammals, such as elephants and whales.
To the extent that it will cover not only frogs and turtles, deer and woodpeckers
but also chipmunks, guinea pigs, rabbits, and squirrels. Employing a term like
“animals of/with certain size” or perhaps the less cumbersome expression “large
animal” might help reduce the vagueness inherent in the wider term “animals,”
but even they remain fairly imprecise and vague. They will appear clearly
applicable to larger dogs, e.g. those that are more than 75 pounds in weight or
25 inches in height, such as Great Danes and Saint Bernards, and clearly not to
very small ones, such as Chihuahuas, Dachshunds, Scotch and Yorkshire Terriers.
But there will be many borderline cases, such as Airedale Terriers, German
Shepherds, and Labradors where it will be far from clear whether, by the very
82 See Waluchow (2007: 85–91, and 225–6).

meaning of its terms, the rule applies. By way of contrast, terms like “dangerous
animals” or “animals that may affect or jeopardize the well-being of terminal
users” are more likely to help solve our problem, at least to some extent. The
term “dangerous animals,” even though it may be helpful in reducing vagueness
is of course still quite an imprecise or vague expression. It will appear clearly
applicable to certain animals, such as lions and tigers, and not applicable to some,
such as frogs and turtles. But once again there will be numerous borderline cases,
e.g. chipmunks, guinea pigs, hamsters, rabbits, and squirrels where it will be less
than clear whether the rule applies or not. Lastly, the phrase “animals that may
affect or jeopardize the well-being of terminal users,” despite being itself a
relatively imprecise or vague expression, is perhaps the most useful of our
candidate terms. This is because it likely corresponds, in a straightforward way,
to the actual purpose behind the prohibition of dogs, which seems to be equally
applicable to bears and other animals who may affect or jeopardize the users of
the airport terminal by posing a real or imminent threat to them and to their
well-being. In that sense, if we consider that the reasons that justify the prohib-
ition of dogs are analogically the same—perhaps even more so—in the case of
bears and any other animals that pose a real or imminent threat, such as cats due
to people’s allergies, it is clear that extending the prohibition from dogs to other
animals such as bears and cats by widening the scope of the rule will be justified.
This is because the prohibition is equally applicable to them to the extent that in
fact that is what the rule requires all the way.
Before responding to the objection that it is necessarily the case that we must
always end up in R5 if the line of reasoning sketched in the preceding few
paragraphs is pursued, let me respond to another possible objection. It might be
said that there is no need to consider levels R2–R5 in an account of law and its
inherent legal rationality since R1 is always sufficient to determine the application
of a legal norm. In other words, it is neither necessary, nor is it ever important or
valuable, for a decision maker to move to the so-called “upper levels,” i.e. R2–R5 if
his objective is to determine what the law requires. In response to this objection,
I have several replies.
First, even if it is necessary to start with R1, it is clear that it is not always
sufficient due, not only to the indeterminacy of language, e.g. the so-called “open
texture” of general terms and their often ambiguous and vague nature, 83 but also to
the possibility of errors and mistakes, on one side, and of oversights and even
silence, on the other, including antinomies and gaps.84
Second, even if not necessary to proceed in all cases to upper levels, it is at least
important or valuable to do so in some cases, such as in the case of extending the
prohibition from dogs to bears—and even cats—by widening the scope of the rule’s
application, to the extent that not doing so will amount to either an absurd or moral
repugnant outcome or to an irrational result.
83 See Hart (1961: 124) and (1994: 128).

84 See Fuller (1999/1949: 1859).
122 Imer B. Flores
Third, and finally, regarding the objection that it is necessarily the case that, once
we permit rule-appliers to move beyond R1, they will inevitably end up appealing
to R5, i.e. to the idea that laws and legal decisions must be fair or just and as a result
can neither admit an injustice nor the violation of basic principles and rights,
including moral ones, let me respond by simply denying that this is necessarily so.
Rather, it is at least morally important or valuable that they do so in some cases
only, i.e. in those cases in which not doing so will amount to admitting a serious
injustice or the violation of basic principles and rights, including moral ones, such
as the “golden rule.”
To see why this is so, let’s reconsider our previous example: Imagine that
someday a blind person—called him Louis—attempts to enter the airport terminal
not with a grizzly bear like Ben or a dog like Benji, but with an assistance dog—call
him Lazarus. Let’s suppose further that the proper administrative or executive
authority, i.e. the policeman, grants him entry. Or suppose, in the unlikely case
that Louis is prevented from doing so by the policeman, that the proper adjudi-
cative authority, i.e. the judge, reverses the decision and grants him the right to
enter. In my opinion, this case is the exceptio probat regulam in casibus non exceptis.
Notwithstanding this point, someone might argue in light of R1 that an assistance
dog is a dog after all, and that entering the terminal with Lazarus is expressly
prohibited. Similarly, someone might argue, by appealing to R2, that the exception
excluding assistance dogs—or perhaps service dogs in general—from the prohib-
ition has to be explicitly consigned to be applicable, following the maxim: Ubi lex
non distinguit, non distinguere debemus. However, deciding at level R1 or R2 will
amount to a violation of R5, by constituting discrimination against the blind
person’s basic rights and liberties, such as freedom of movement in conditions
similar to those in which non-blind persons enjoy this freedom. Therefore, it is not
merely important or valuable, but morally important or valuable and even (morally)
imperative or necessary that a move be made first to R3 and R4 and perhaps later on
to R5, in order to reach the following conclusion: Although the prohibition is
clearly directed at dogs and, analogously, bears, cats, and other animals that may
seriously threaten the well-being of terminal users, it excludes from its application
certain dogs, such as assistance dogs. And this is so regardless of whether or not this
exclusion has been mentioned explicitly. So the relevant question is whether an
assistance or service dog poses a real or imminent threat due to the fact of being
intrinsically perilous and representing a risk to users of the terminal and their well-
being. The answer to this question is, in my opinion, quite clearly no. All else being
equal or caeteris paribus: a blind person’s assistance dog in no way poses a real or
imminent threat to terminal users and their well-being. This claim, from my point
of view, justifies letting Louis enter with Lazarus as this is what the rule actually
requires, and is grounded not only in R3 and R4 but also in R5. This is because such
a permission avoids a serious injustice or violation of principles, including moral
ones. It serves justice and other principles, including moral ones, by legitimating
the rule as applied to the case of Louis and Lazarus and might be extended as a way
of precedent to other service dogs.
V
In closing, let me recall that my main claim, made at the very outset, was that even if
we should concede that there are no necessary and sufficient features of law, and in
particular any necessary connections with morality, the following points should be
acknowledged. First, there are at least some generally recurring features of law that
are important or valuable and even morally important and valuable; and second
that an adequate theory of law owes us an account of these features. An adequate
legal theory must provide not only a description or explanation of these important
legal phenomena, but also an evaluation and perhaps even a justification of them.
With the foregoing in mind, I suggested in section II that we bracket any
discussion of whether certain features of law are sufficient and/or necessary and
proceed to a discussion of those features we consider important or valuable. In
section III I affirmed, following Holmes, that law not only includes several features
that play an important and valuable (if not necessary) role in legal reasoning, such as
logic and experience, but also that morality is among them. In section IV, with the
idea of figuring out whether all, some, or some combination of these features is not
merely important or valuable but morally important or valuable, I analyzed the
concept of legal rationality and arrived at a three-part conclusion: first, legal
rationality comprises five different levels, spheres, or types, which are at least
important or valuable; second, one of these five levels, spheres, or types invokes
that which is morally important or valuable, i.e. R5; and, third, it is necessary (and
important and valuable) that the five different levels, spheres, or types be integrated
into a complex, but workable, legal rationality.
To conclude, let me draw attention to one possible implication of my interpret-
ation of the nature of law vis-à-vis legal rationality revisited. If I am right in turning
the tables to discuss the morally important or valuable aspects of law, we might end
up realizing at some point that something we deem morally important or valuable
in legal practice is actually necessary as well. And so we might find that law and legal
rationality neither exclude necessary references to morality nor merely include
important or valuable, but contingent, references to morality. What we might
find, instead, is that law and legal rationality actually integrate necessary references
to morality. Moreover, we might discover this additional fact: that law and legal
rationality do—and indeed must—integrate such references does not mean that
law and legal rationality exhaust them or that those references are sufficient as some
versions of natural law theory appear to claim.
References
Austin, J. (1998/1832). “The Province of Jurisprudence Determined,” in The Province of
Jurisprudence Determined and The Uses of the Study of Jurisprudence. Indianapolis:
Hackett.
Black, M. (1954). Problems of Analysis. London: Routledge & Kegan Paul.
124 Imer B. Flores
Cardozo, B. N. (1921). The Nature of the Judicial Process. New Haven: Yale University
Press.
Cardozo, B. N. (1924). The Growth of the Law. New Haven: Yale University Press.
Cohen, F. S. (1935). “Transcendental Nonsense and the Functional Approach,” Columbia
Law Review, 35: 809–49.
Coleman, J. (2001). The Practice of Principle: In Defense of a Pragmatist Approach to Legal
Theory. New York: Oxford University Press.
Dickson, J. (2001). Evaluation and Legal Theory. Oxford and Portland: Hart Publishing.
Dworkin, R. (1978). Taking Rights Seriously. Second edition, “With Appendix: A Reply to
Critics.” Cambridge, MA: Harvard University Press and London: Duckworth.
Dworkin, R. (1985). A Matter of Principle. Cambridge, MA: Harvard University Press.
Dworkin, R. (2006). Justice in Robes. Cambridge, MA: Harvard University Press.
Dworkin, R. (2011). Justice for Hedgehogs. Cambridge, MA: Harvard University Press.
Ehrenberg, K. M. (2009). “Defending the Possibility of a Neutral Functional Theory of
Law,” Oxford Journal of Legal Studies, 29: 91–113.
Flores, I. B. (1997). “La concepción del derecho en las corrientes de la filosofía jurídica,”
Boletín mexicano de derecho comparado, 90: 1001–36.
Flores, I. B. (2005). “The Quest for Legisprudence: Constitutionalism v. Legalism,” in
L. J. Wintgens (ed.), The Theory and Practice of Legislation: Essays on Legisprudence.
Aldershot: Ashgate, 26–52.
Flores, I. B. (2007). “Legisprudence: The Forms and Limits of Legislation,” Problema:
Anuario de filosofía y teoría del derecho, 1: 247–66.
Flores, I. B. (2009a). “Legisprudence: The Role and Rationality of Legislators—vis-à-vis
Judges—towards the Realization of Justice,” Mexican Law Review, 1: 91–110.
Flores, I. B. (2009b). “The Living Tree Constitutionalism: Fixity and Flexibility,” Problema:
Anuario de filosofía y teoría del derecho, 3: 37–74.
Flores, I. B. (2010). “Ronald Dworkin’s Justice for Hedgehogs and Partnership Conception of
Democracy (With a Comment to Jeremy Waldron’s ‘A Majority in the Lifeboat’),”
Problema: Anuario de filosofía y teoría del derecho, 4: 65–103.
Flores, I. B. (2011). “H. L. A. Hart’s Moderate Indeterminacy Thesis Reconsidered: In
Between Scylla and Charybdis?,” Problema: Anuario de filosofía y teoría del derecho, 5:
147–73.
Fodor, J. A. (1975). The Language of Thought. Cambridge, MA: Harvard University Press.
Fuller, L. L. (1958). “Positivism and Fidelity to Law: A Reply to Professor Hart,” Harvard
Law Review, 71: 630–72.
Fuller, L. L. (1968). The Anatomy of Law. New York: Frederick A. Praeger.
Fuller, L. L. (1969). The Morality of Law. Second edition. New Haven: Yale University
Press.
Fuller, L. L. (1999/1949). “The Case of the Speluncean Explorers,” Harvard Law Review,
112: 1851–75.
Gallie, W. B. (1965). “Essentially Contested Concepts,” Proceedings of the Aristotelian
Society, 56: 167–83.
Griswold, E. N. (1967). “Intellect and Spirit,” Harvard Law Review, 81: 292–307.
Hart, H. L. A. (1961). The Concept of Law. Oxford: Oxford University Press.
Hart, H. L. A. (1983a). “American Jurisprudence through English Eyes: The Nightmare
and the Noble Dream,” in Essays in Jurisprudence and Philosophy. Oxford: Oxford
Hart, H. L. A. (1983b). “Jhering’s Heaven of Concepts and Modern Analytical
Jurisprudence,” in Essays in Jurisprudence and Philosophy. Oxford: Oxford University
Press, 265–77.
Hart, H. L. A. (1994). The Concept of Law. Second edition, “With a Postscript.” Oxford:
Holmes, Jr., O. W. (1880). “Book Notices,” American Law Review, 14: 233–4.
Holmes, Jr., O. W. (1919). “Dissent,” in Abrams v. United States, 250 US 616.
Holmes, Jr., O. W. (1920a). “Dissent,” in Missouri v. Holland, 252 US 416.
Holmes, Jr., O. W. (1920b). “Law in Science and Science in Law,” in Collected Legal Papers.
London: Constable and Co., 210–43.
Holmes, Jr., O. W. (1991/1881). The Common Law. New York: Dover.
Holmes, Jr., O. W. (1997/1897). “The Path of the Law,” Harvard Law Review, 110:
991–1009.
Jhering, R. (1951). “In the Heaven of Legal Concepts,” in M. R. Cohen and F. S. Cohen
(eds.), Readings in Jurisprudence and Legal Philosophy. New York: Prentice-Hall, 678–89.
Kant, I. (1896/1781). Critique of Pure Reason. Translated by F. Max Müller. London:
Macmillan.
Kripke, S. A. (1980). Naming and Necessity. Cambridge, MA: Harvard University Press.
Marmor, A. (2001). Positive Law and Objective Values. Oxford: Oxford University Press.
Marmor, A. (2006). “Legal Positivism: Still Descriptive and Morally Neutral,” Oxford
Journal of Legal Studies, 26: 683–704.
Moore, M. S. (1992). “Law as a Functional Kind,” in R. P. George (ed.), Natural Law
Theory: Contemporary Essays. Oxford: Oxford University Press, 188–242.
Pound, R. (1923). Interpretations of Legal History. Cambridge, MA: Harvard University
Press.
Pound, R. (1960). Law Finding through Experience and Reason. Atlanta: University of
Georgia Press.
Putnam, H. (1975). Mind, Language and Reality. Cambridge: Cambridge University Press.
Quine, W. V. O. (1969). “Natural Kinds,” in Ontological Relativity and Other Essays. New
York: Columbia University Press, 114–38.
Raz, J. (1994). “The Problem about the Nature of Law,” in Ethics and the Public Domain:
Essays on the Morality of Law and Politics. Oxford: Oxford University Press, 195–209.
Recaséns Siches, L. (1959). Tratado general de filosofía del derecho. México: Porrúa.
Schauer, F. (2010). “Necessity, Importance, and the Nature of Law,” Virgina Public Law
and Legal Theory Research Paper, No. 2010–19, available for download at the Social
Science Research Network Electronic Paper Collection: <http://ssrn.com/
abstract=1594930>.
Smith, M. (1909). Jurisprudence. New York: Columbia University Press.
Stammler, R. (1925/1902). The Theory of Justice. Translated by Isaac Husik. New York:
Macmillan.
Stavropoulos, N. (1996). Objectivity in Law. Oxford: Oxford University Press.
126 Imer B. Flores
Waluchow, W. J. (1994). Inclusive Legal Positivism. Oxford: Oxford University Press.
Waluchow, W. J. (2007). A Common Law Theory of Judicial Review: The Living Tree.
Cambridge: Cambridge University Press.
White, M. (1949). Social Thought in America: The Revolt against Formalism. New York:
Viking.
Wittgenstein, L. (1953). Philosophical Investigations. London: Blackwell.
PART II
THE POWER OF LEGAL SYSTEMS
6
Law as Power
Two Rule of Law Requirements*
Bruno Celano
1. Introduction
“Law as power” is a somewhat neglected topic in contemporary analytical jurispru-
dence. To be sure, attention has been paid, from Hart (and Kelsen) onwards, to
normative legal powers (i.e. norm-creating and norm-applying powers, instituted
by legal power-conferring norms). “Brute” social power, however, and law’s relation
to it, are, in post-Hartian jurisprudence, largely ignored.1
This will be my topic in this chapter. I shall discuss, that is, the way the law
operates as a mode of the exercise of social power. I approach the issue from a
specific angle (section 2). The subject of my inquiry is the shape social power takes
when the rule of the law is envisaged as an ethico-political ideal. I discuss, that is,
the Rule of Law as a specific mode of the exercise of social power, and what is
valuable in it (a preliminary characterization will be offered in section 3).
I concentrate on two Rule of Law requirements, consistency (roughly, the
avoidance of conflicts) and compliability (roughly, conformity to the principle
“ought” implies “can” ) (section 4). These two desiderata contribute, I argue, to
defining a distinctive mode of social power, one that is characterized by its showing
at least a minimum of respect for the dignity of its subjects. As we shall see, power
can be effectively exercised by systematically flouting these two desiderata. Entire
power structures may flourish around vast areas of inconsistent or non-compliable
directives. Consistency and compliability are required, however, if the laws are to
treat their subjects as autonomous, responsible agents (sections 4, 5). Under both
* I wish to thank Francesca Poggi and Stefan Sciaraffa for their comments on an earlier version of this
chapter.
1 Hart (1961: ch. 4) rightly criticized the view that the law should be understood as the gunman
situation writ large—thus, he rejected Austin’s concept of the sovereign as the source of all legal rules—
as naive and simplistic. Nothing comparable, however, was put in its place. Law’s foundations are seen
in the practices of the club of officials. Brute social power—the role of law as a cog in the workings of
social powers—is, in a way, Hart’s blind spot. (I have argued for this claim in Celano (2012). Nothing
in the argument of this chapter depends on the correctness of this assessment of Hart’s jurisprudence.)
130 Bruno Celano
respects, it turns out (section 6), power exercised through the Rule of Law qualifies
as public, rational, and non-paternalistic.
2. Social power through the lens of the Rule of Law

Law’s standing vis-à-vis social power is a central issue in legal theory—witness the
works of Weber, Kelsen, Ross, and, in Italy, N. Bobbio. Not to mention, of course,
Hobbes, Bentham, Austin, or Marx.
One obvious way of approaching the subject is by investigating the relations
between law and coercion, and the different ways in which law may be coercive.
One does not have to endorse the (implausible) view that coercion is the differentia
of law as contrasted with other kinds of social order,2 nor the claim that coercion is
an essential feature of law, in order to grasp the significance of this connection—law
operating as a means of social power through its being coercive. Even if we grant
that coercion is not a necessary feature of law, it would be wrong, I think, to deny
jurisprudence the task of clarifying the issue. For this to be a legitimate task for
(general) jurisprudence it is sufficient, I think, that typically (epi to polu) law is
coercive.3
Be that as it may, my focus is not, in this chapter, on the coerciveness of the law.
What I am interested in, is law understood as a distinctive mode of the exercise of
social power. This evokes grand sociological generalizations—e.g. the law as one of
the manifestations of the state’s monopoly of legitimate force; law as an instrument
of ideological hegemony, or social oppression; the law as the organization of force in
society, involving the institutionalization of the use and the threat of physical force.
I am going to approach the issue, however, from a different angle.
The object of this chapter is the shape social power takes when the rule of the law
is envisaged as an ethico-political ideal. By “the rule of the law as an ethico-political
ideal” I mean what is usually called “the Rule of Law” (RoL, for short). I understand
the RoL in the formal, “thin,” sense,4 which has now become usual among
jurisprudents. So understood, the RoL comprises a loose cluster of (1) formal
features of the laws (prospectivity, publicity, relative generality, relative stability,
intelligibility and relative clarity, compliability, consistency), plus (2) institutional
and procedural desiderata (such as, for instance, that the making of individual
norms, applying to individual cases, be guided by general rules).5 The RoL will be
the lens through which I will look at my subject. I shall consider the RoL, in the
specified sense, as a distinctive mode of the exercise of social power.
2 Kelsen (1945: part I, ch. 1). 3 Cf. Schauer (2010). 4 Tamanaha (2004: 91).
5 For a list of these institutional and procedural requirements see e.g. Raz (1979: 215–18): “the
making of particular laws (particular legal orders) should be guided by open, stable, clear, and general
rules;” “the independence of the judiciary must be guaranteed;” “the courts should have review powers
over the implementation of the other principles;” “the courts should be easily accessible;” “the
discretion of the crime-preventing agencies should not be allowed to pervert the law.” For similar
lists of the RoL requirements, see Fuller (1969: ch. 2); Finnis (1980: 270–1); Marmor (2004: 5ff.);
Kramer (2007: ch. 2).
Law as Power 131
Before we start, two clarifications are in order. (1) My concern is, first, with what
the RoL, regarded as a distinctive mode of social power, is—what its main features
are. And, second, with what is valuable in it: what the values are that this kind of
power serves, and how. (Under this respect, mine will be a substantive ethico-
political inquiry.) I sidestep the issue of whether the concept of law entails the idea
of the RoL, or perhaps some minimal version of it. The answer to this question is,
to my present purposes, immaterial. (2) The subject of this chapter is the law as a
vehicle of “social power,” or of “brute” power (as contrasted with legal normative
powers; see section 1). What should be meant by these terms?
Power is, of course, a puzzling, highly problematic notion. The word is poly-
semic. According to some,6 the concept of power—power in society—is an
essentially contested concept. In what follows, I shall rely on an intuitive under-
standing of what social power is, what kinds of social relationships do qualify as
power relationships, and so on. There are, in fact, accounts of social power which
impose very stringent conditions on the applicability of the notion. According to
some such austere accounts, some of the “varieties of power” that I shall mention
may not qualify as forms of power at all. A fully-fledged deployment of my
argument would require a fully developed account of this notion. As I said,
I shall not attempt this here. I rely on an intuitive idea, no doubt vague, imprecise,
generic, of what power relations in society are, and of their varieties. Thus,
I stipulate, very roughly, that “social power” consists, basically, in this: someone
exercises a measure of control over someone else’s behavior, thought, or feeling, in
the pursuit of his ends or interests (which may include controlling someone else’s
behavior, thought, or feeling). My hope is that the legitimacy of this intuitive,
unrefined notion, and of the characterization of the phenomena we shall encounter
as varieties of “power”, will be apparent.
Thus, the concept of social power will remain unanalyzed. It will be used as a
primitive. A word, however, is needed about the characterization of the family of
power relations I shall discuss as forms of “brute” social power. In this context,
“bruteness” is to be understood as relative:7 the relevant forms of power may be said
to be “brute” relative to the layer of normative powers (powers to create and apply
norms) instituted by legal power-conferring rules. This does not mean that they
may not be complex or institutionalized phenomena in their own right, relative to
further, more basic kinds of facts.
3. Rule of Law power: a preliminary characterization

Laws meeting RoL requirements may have almost any content. What is peculiar, as
regards the RoL, is, however, the form that the exercise of power takes. The RoL is,
in the first instance, a specific mode of the exercise of power.
6 Lukes (2005). 7 Anscombe (1958).

132 Bruno Celano
It is not at all unusual to understand the RoL as a specific mode of the exercise of
power, in particular, political power. In order to grasp its specificity we have to
consider what other modes of the exercise of power, political and generally social, it
should be contrasted with.
As a specific mode of power the RoL is usually contrasted with “managerial
governance or rule by decree.”8 Or it is contrasted with “arbitrary” power, meaning
by this either power not exercised in conformity with the rules it should conform
to;9 or public power exercised in the pursuit of private interests.10 It is also often
contrasted with “manipulation,” or forms of “manipulative” power.11
I take my cue from this last suggestion. RoL power will be contrasted, in what
follows, with varieties of “manipulative” power.
“Manipulation” is a generic, vague, and ambiguous word, carrying evaluative and
strong emotive connotations. In order to render the contrast between RoL power and
“manipulative” varieties of power meaningful, and illuminating, we have to make its
meaning determinate. I propose to do this by stipulating two conditions, which our
account of RoL power—and, by contrast, of the “manipulative” varieties of power it
is, ex hypothesi, opposed to—will have to satisfy. The first condition concerns the
descriptive import of the word. The second concerns its evaluative implications.
(1) “Manipulation” evokes the idea of treating human beings as things, to be
modified, altered, etc. as one sees fit, irrespective of their beliefs, attitudes, or will.
Conversely, a non-manipulative form of power (e.g. the RoL) will have to be
understood as characterized by the treatment of its subjects as sources of deliber-
ation, capable of making choices in the light of their beliefs and attitudes, and of
acting on the basis of these choices.
Admittedly, this does not say much, and can easily be shown to be insufficient.
(So, for instance, when X makes Y do A by relying on Y’s belief that in circum-
stances C one ought to do A, and by deceiving Y so that he mistakenly believes
that circumstances C do in fact obtain, isn’t X “manipulating” Y’s behavior?12)
But it will be enough as a preliminary characterization. We shall make it more
determinate as we go on.
(2) “Manipulation” is, presumably, power which does not respect the dignity of
its subjects.
It is a widely held opinion that the RoL is a mode of power characterized by
(a minimum of) respect for the dignity of human beings, or at least that “observance
of the rule of law is necessary if the law is to respect human dignity.13 Our account
will have to explain in what sense this is so, and why.
8 Waldron (2008a: 78). 9 Bobbio (1999: 182–3). 10 Raz (1979: 219–20).

11 Fuller (1969: 163–7), and, on Fuller’s views thereabout, Waldron (2008b: 17); Raz (1979: 221);
Finnis (1980: 273); Kramer (2007: 184). The RoL is famously contrasted with “the rule of men.” An
understanding of this Protean contrast, however, is—barring rhetorical formulae—the end, rather than
the beginning, of theorizing.
12 Cf. Raz (1979: 221).
13 Raz (1979: 221). That the point of (many of) the RoL requirements is to make possible, inter
alia, (certain forms of) respect for human dignity, is common ground. Cf. Fuller (1969: 162–3); Finnis
(1980: 272–3); MacCormick (1985: 26); Marmor (2004: 21, 32); Waldron (2008a: 76).
Law as Power 133
The first condition gives us the beginning of the required explanation. Presum-
ably, “manipulative” power does not respect the dignity of its subjects because
it treats them as things. This, however, is only the beginning of an explanation.
We have to show in what ways subjects, under the RoL, are not treated as things,
and how—and to what extent—these contribute to an attitude of respect for the
dignity of human beings.
So, RoL power will have to be understood, in contrast with “manipulative”
forms of power, as power that, first, treats subjects as centers of deliberation and
that, second, in so doing (pro tanto) respects their dignity as human beings. I try to
show that this is so, and why, by focusing on two RoL requirements, consistency
and compliability.
4. Two Rule of Law requirements

How does law’s conformity to these two requirements contribute to defining
a distinctive mode of social power—one that is not “manipulative,” and thus
(pro tanto) respects the dignity of those over whom it is exerted?
The conception of the RoL adopted here (section 2) is mainly instrumental.
(Fuller’s tale concerning the different ways in which Rex may fail in his enterprise
aptly illustrates this kind of approach.14) The RoL is understood, basically, as
comprising what is instrumentally necessary in order to achieve a certain goal.
Which goal? A first, approximate answer is: guiding human behavior. So, for
instance, directives typically have to be prospective, and intelligible (prospectivity
and intelligibility are, as we have seen, two among the various RoL desiderata), if
they are to be capable of guiding human behavior.
This is, however, generic and imprecise. There are many ways of guiding (taking
the word “guide” at its face value) human behavior. Most of the requirements of the
RoL (in particular, those in the first group; section 2) specify, more or less directly,
what is instrumentally required in order to achieve the end of guiding human
behavior through rules.15
These two ends—“guiding human behavior,” generally, and “subjecting human
behavior to the guidance of rules”—are different. The latter is a specification of the
former: subjecting human behavior to the guidance of rules is a specific way of
guiding it. As we shall see, the difference between these two goals is crucial in
explaining how RoL power satisfies the two conditions set out in the preceding
section (not treating subjects as things and, to this extent, respecting human
dignity).
What is involved, then, in “subjecting human behavior to the guidance of rules,”
as the proper end which accounts for RoL desiderata? One obvious ingredient is
generality: the word “rule” designates, we may assume, a general directive. But this
14 Fuller (1969: 33–8).

15 In Fuller’s phrase, “the enterprise of subjecting human conduct to the governance of rules”
(1969: 106).
134 Bruno Celano
is only part of the answer. By focusing on consistency and compliability, rather than
generality, I mean to explain further facets of what is involved in the idea of
subjecting human behavior to the guidance of rules, in its difference from the
idea of guidance, as such.
In other words. “Subjecting human behavior to the guidance of rules” is the basic
purpose grounding, in its main outlines, the RoL. In order to understand what the
RoL is, and what it requires, we have to get this purpose right. What is involved in
it, in addition to the idea of guidance through general directives, and what follows
from this? Focusing on the two requirements, consistency and compliability, is a
way of answering this question.
In understanding what the end which explains RoL desiderata is, and how it
differs from that of guiding (without specification) behavior, the crucial point is
what may be termed “normal,” or standard, guidance: that is, guidance through an
understanding, by the subject, of what the law, according to its tenor, requires of
him, and that it is required of him. The RoL comprises a set of conditions which
have to be satisfied for the normal, standard guidance (so defined) of human
behavior.
The simplest case is given by the idea of “telling people what they should do”
(and, maybe, threatening them of unpleasant consequences in case of non-compli-
ance). When X tells Y to do A, typically, X tries to guide Y’s behavior in the
direction of doing A through the understanding, by Y, of what is required of him
(namely, doing A), and that it is required of him. This is the standard case. All sorts
of non-standard cases—non-standard forms of guidance—can be devised,16 as will
be borne out by our examination of the requirements of consistency and complia-
bility (below, in this section, and in section 5). In accounting for RoL desiderata,
and for their point, the general form of the explanation will have to be, in outline: it
is necessary, in order for the laws to achieve their standard purpose—guiding
human behavior, according to their tenor, through their being understood by
their subjects—that they satisfy condition C, for such-and-such reasons.
So, the task is to show why consistency and compliability are instrumentally
necessary for standard guidance, and what follows from this. We have to ask
ourselves, first, what kinds of power may be served by the systematic violation of
these two requirements, and, second, in what ways these may turn out to be
“manipulative” in character, and thus prove disrespectful of the dignity of those
who are subject to them.
In answering these two questions, I shall work with a simplified model of the
RoL. I regard the RoL as an ethico-political ideal concerning the relations between
a lawgiver (in the widest sense) who issues prescriptions and those who are
subject to these prescriptions17—an ideal, then, concerning the shape prescriptive
16 A simple case of non-standard guidance: parents may sometimes, in order to make their young
children do A, tell them not to do A, relying on their children’s standing disposition to do the opposite
of what they tell them to do.
17 “Prescriptions are given or issued by someone. They ‘flow’ from or have their ‘source’ in the will
of a norm-giver . . . They are, moreover, addressed at some agent or agents, whom we shall call
Law as Power 135
relationships (i.e. the kind of relationship which comes into being, by virtue of the
happy issuing of a prescription, between a lawgiver, on the one hand, and those
to whom her prescription is addressed, on the other hand) should take. How to
transpose this model to a more complex reality (some legal norms, we may plausibly
assume, are not prescriptions) is, admittedly, a problem.18
This is a “legislative” conception of the RoL. At least prima facie (but see in this
section, sub (2)) legislating plainly is the issuing of prescriptions. So understood,
then, the RoL is the rule of legislation. What follows concerns only this version
of the RoL.
My justification for adopting this model (apart from the gains of simplicity) is
twofold.
(1) In the jurisprudential tradition from which the conception of the RoL
adopted here (“thin” RoL; section 2) stems, the RoL is, in fact usually
understood along these lines (witness, for example, Lon Fuller’s tale of
Rex, which is the matrix of these theorizations),19 as a matter of the
conditions enacted directives have to meet if the purpose intended by
those who have enacted them is to be achieved.
(2) In contemporary legal systems, the role of legislation is paramount. “Legis-
lative” RoL captures at least a very important layer of the idea of the
RoL. True, the apparently obvious claim that legislation is the issuing of
prescriptions—that, then, statutes are prescriptions—is by no means apro-
blematic. Legislation proper, as it occurs in developed legal systems, has
many complex, articulated procedural and institutional features, which have
no obvious equivalent in the case of simple acts of prescribing.20 Most
important, there is no obvious way in which a multi-membered legislature,
composed by individuals and groups who sharply disagree with each other
on the relevant issues, and making decisions on the basis of majority rule,
may be assimilated to an individual, enacting his own will by expressing it
in the form of a prescription.21 These difficulties notwithstanding, it seems
to me, the model—a “prescriptive” understanding of the RoL—remains
norm-subject(s).” Von Wright (1963: 7). Further features which Von Wright sees as “characteristic of
norms which are prescriptions” are of no interest for us here.
18 In order to make room for power-conferring rules (and, especially, for rules conferring to private
individuals the power to achieve some ends of theirs: “If you wish to do this, this is the way to do it”
(Hart (1961: 28)), we should understand “prescribing” as including cases of telling people how to pursue
the goals they want to achieve (or telling people how to do what they want to do). Cf. Raz (1979: 215):
“power-conferring rules are designed to guide behavior.” Prescribing, so understood, covers both the
issuing of mandatory directives and the issuing of power-conferring rules. In allowing in the text for the
possibility that some legal norms are not prescriptions, I have in mind (not, power-conferring rules,
but) norms that, at least apparently, are not “issued” by any “lawgiver.” This topic will not be further
discussed here. “Prescriptive” RoL at least captures, I shall assume, one facet of the RoL.
19 Waldron (2007: 109–10) rightly observes that Fuller’s treatment of the subject in Fuller (1969:
ch. 2) “illustrates a strong . . . tendency to associate the rule of law with formal features of legislation, as
opposed to other modes of law and law-making.”
20 MacCormick (1973: 114–15).
21 Waldron (1999a: part I) and (1999b: 26–8).
136 Bruno Celano
illuminating, especially insofar as the RoL is envisaged—as it, in fact, usually
is—as an ethico-political ideal concerning the relations between those who
govern and the governed, between rulers and ruled. Pondering the formal
features of prescriptions remains, it seems to me, the first step in understand-
ing the nature of legislation (I shall not defend this view here; the reader may
take it as an axiom).22
Thus, RoL requirements (in particular, those in the first group, section 2) will be
understood as features prescriptions have to exhibit if they are to be capable of
achieving, and of achieving well, the purpose of standard guidance of the behavior
of their addressees. Prescribing is a form—a paradigmatic form—of “telling people
what they should do.”23 Thus, standard guidance is the typical, normal purpose of
the activity of prescribing—it may be termed its constitutive purpose. The practice
(or the “institution”) of prescribing would not exist, standard and non-standard
cases alike, were it not commonly and (usually) rightly assumed, when prescriptions
are issued, that they aim at standard guidance.
RoL desiderata (those in the first group), then, will be features that prescriptions
typically exhibit, and conditions prescriptions normally meet, in varying degrees.
The possibility of non-standard prescriptions is not ruled out. In particular, as we
shall see in a minute, abuses are possible: people can exploit the opportunity of
issuing prescriptions—they can make a strategic use of the institution of prescrib-
ing—in order to pursue in deviant, indirect ways, non-standard, undisclosed
purposes.
So, what kinds of power may be served by the issuing of prescriptions that
systematically violate either of our two requirements, and in what ways may such
legislation be “manipulative” of those who are subject to it, and not respect their
dignity?
Imagine a power relation of the following kind: the lawgiver does not intend
the addressee to do what he prescribes him to do. On the contrary, he often
issues inconsistent or non-compliable prescriptions (e.g. by commanding actions,
or omissions, which run counter to the most deeply ingrained inclinations of
the vast majority of human beings). What the lawgiver intends, is precisely that
the addressee should fail to comply, and, as a consequence, develop a sense of
puzzlement, fear, anxiety, or a strong feeling of guilt. Thanks to that, it will be easy,
then, to make the addressee do, by further means, what the lawgiver actually wants
him to do.
This is, I submit, a paradigm of a power structure that works through the
manipulation of subjects. I shall now try to unpack what is implied in this minimal
Gedankenexperiment. In so doing, by contrast, the main features of RoL power will
22 As is well known, arguments against a “prescriptive model” of governance through law are legion.
Cf. Hart (1961: 20–2); Hurd (1990).

23 “Telling” people what they should do, as I mean it here, refers to cases of issuing prescriptions,
not to “detached” statements of what the subject should do according to a given set of prescriptions.
Raz (1979: 153–7).
Law as Power 137
emerge—and it will be apparent in what sense, and why, the RoL meets the two
conditions laid out in section 3.24
4.1. Consistency
I begin with some assumptions concerning the notion of “inconsistent” prescrip-
tions, and how purported logical relations between prescriptions should be under-
stood. These will be the conceptual tools to be employed in the inquiries that
follow. These assumptions will be merely stated, here. Their defense is deferred to
another occasion.
By “inconsistency” (between prescriptions) I mean three kinds of cases: (1)
conflicts proper (one and the same action A is both obligatory and forbidden:
OA & FA); (2) contradictions (it is both obligatory that A and permitted that not
A; or, A is both forbidden and permitted);25 (3) cases where two conditional
directives, referring, respectively, to the condition that p and the condition that
q, reconnect to these conditions either conflicting or contradictory deontic conse-
quences (as defined in cases (1) and (2)), and, further, p and q jointly occur (it is left
open the possibility that p and q are one and the same condition; when this is so,
trivially, when the one occurs the other also necessarily occurs). In the first case, and
the corresponding case in the third group, whatever one does is wrong (these are
dilemmatic situations). In all cases, the relevant prescriptions cannot both be
obeyed or satisfied.26 This marks a difference between statements of fact (or,
propositions), on the one hand, and prescriptions, on the other hand. Statements
of fact (or propositions) may be true or false; that two statements of fact (or
propositions) are inconsistent means that there is no possible world in which
they are both true. Prescriptions are neither true nor false. That two prescriptions
are inconsistent means that there is no possible world in which they are both
followed or satisfied (with one qualification, to be spelt out soon).
The first of our two RoL desiderata requires the avoidance of inconsistencies.
Why? Inconsistent prescriptions (on any plausible account of the existence condi-
tions of prescriptions), or the issuing of such prescriptions, are not a logical
impossibility. A world in which prescriptions which cannot both be obeyed or
24 RoL requirements interact, and they work together in enabling the achievement of the purpose
and values the RoL serves. Consistency and compliability are not—neither separately, nor jointly—
sufficient conditions. So, the argument is not that when the laws are consistent, and it is possible to
comply with them, then they satisfy the two conditions laid out in section 3. Rather, an understanding
of the point, and import, of the two desiderata illustrates what the nature of RoL power is. Analogous
arguments could be developed, mutatis mutandis, with regard to other RoL desiderata.
25 I assume, following one usual approach in deontic logic, that OA entails PA (i.e. A is permitted),
and, correspondingly, that FA entails that ¬A is permitted. So understood, the conceptual relations
between the operators “obligatory,” “forbidden,” and “permitted” mirror those holding between the
alethic modalities “necessary,” “impossible,” and “possible.” Cf. von Wright (1951: 1–3) and (1968:
142–3).
26 I stipulate—following von Wright (1983: 139); but cf. also Hart (1982: 116)—that a permission
is satisfied when its subject(s) at least sometimes avail themselves of the permission—i.e. they, at least
sometimes, perform the permitted action. This seems fair to what may be taken to be the point of
permissive norms.
138 Bruno Celano
satisfied exist, or are issued, is a possible world (as opposed to a world in which
some of the propositions which are true in it are not true). Purported “logical”
relations between prescriptions, I assume,27 mirror the logical relations proper
holding between their respective obedience- or satisfaction-statements (meaning,
by this, statements of fact, or propositions, describing the actions which satisfy the
relevant prescriptions). Specifically, two prescriptions are inconsistent if and only if
their respective obedience- or satisfaction-statements are inconsistent.28
But why should we single out, as a separate class, those relations between
prescriptions that track the logical relations (for example, inconsistency) holding
between their respective obedience- or satisfaction-statements? Granted, “logical”
relations between prescriptions mirror logical relations between their respective
obedience- or satisfaction-statements. But why do they matter? Why is it important
to single out, in the form of principles of a purported “logic” of prescriptions, the
conditions under which sets of prescriptions are, by logical necessity, jointly
satisfiable? Why is it important to single out the class of those prescriptions—
inconsistent prescriptions—which cannot (logically) jointly be satisfied?
The answer is, I think, this:29 purported logical principles governing prescriptive
discourse have to be understood as criteria of rational law-giving (i.e. as principles a
rational law-giving will should conform to), under one crucial assumption: whoever
prescribes X to do A wants X to do A. It is under this assumption that the
conditions under which sets of prescriptions can (logically can) be jointly satisfied
provide criteria of rational law-giving. A rational lawgiver cannot want what is
(what she believes to be) logically impossible to be the case. This is why the issuing
of inconsistent prescriptions may normally be deemed irrational.
The relevant assumption is, then, that the lawgiver wants the addressee to do
what she tells him to do;30 and, further, she intends, by uttering the prescription, to
make him perform the action. This assumption is crucial, for this reason: it is by
virtue of this assumption that, in general, the issuing of prescriptions becomes
possible. Understanding an utterance as a prescription involves the (defeasible)
ascription, to the utterer, of the assumed intention: the lawgiver wants that the
subject perform the prescribed action; and he intends, by prescribing it, to make
her perform the action.
This is what I shall call “normal,” or “standard” legislative intention.31 Normal
legislative intention is the intention to pursue a prescription’s constitutive purpose,
27 Following von Wright (1983), (1984), (1985); but cf. also Hart (1982: 116) and (1983: 325–7).
28 This does not apply to pairs of permissions: that the satisfaction-statements of two permissions
are logically inconsistent does not entail that they conflict (PA and P¬A are not inconsistent). Hart
(1983: 327). The need for this—logically untidy—qualification is a consequence of the stipulation in n. 26.
29 Cf. von Wright (1983: esp. at 132, 141–2, 149); (1984: 452–3, 456); (1985: 271). Cf. also
Fuller (1969: 66); Hart (1982: 116); Alchourrón and Bulygin (1984: 458); Celano (1990: 268–82).
30 Von Wright (1963: 7, 119); (1983: I, 8). Cf. also MacCormick (1973: 104–5); Hart (1982: 247,
249); Celano (1990: 127).

31 The intentional structure constitutive of prescriptions is, in fact, more complex than this. The
workings of a prescription rest, typically, on a set of nested, self-reflexive intentions. Grice (1957);
Strawson (1964: 256–7); Schiffer (1988: 19); Celano (1990: 127–51, 205–13, and 2010: section 6).
Cf. also Raz (1996: 283). Typically, the lawgiver has (1) the intention to make the addressee perform a
Law as Power 139
namely, standard guidance (section 4). Non-standard legislative intentions are

possible. Prescriptions can be insincere.32 Ascription to a lawgiver of a normal
legislative intention is a defeasible presumption. It is, however, standardly true;
and an explicit denial of this condition (“I hereby order you to do A, but I don’t
care whether you do it or not”) would prevent a prescription from coming
into existence. The institution of prescribing would not exist, standard and non-
standard cases alike, were it not commonly and (usually) rightly assumed, when
prescriptions are issued, that the lawgiver wants that the subject perform the
prescribed action, and he wants, by prescribing her to perform it, to make her
perform the action.
Prescriptive inconsistency, then, is no logical impossibility. Its avoidance is,
rather, a matter of rational law-giving. “Rational” in what sense? Instrumentally
rational, relating to a certain goal: standard guidance, i.e. guiding the behavior of
subjects in the prescribed direction—making the subject perform the prescribed
action—through their understanding of this.
So, the existence of jointly non-satisfiable prescriptions is a logical possibility; the
issuing of inconsistent prescriptions is possible. And—what is crucial to our present
purposes—it may well be instrumentally rational, in the pursuit of non-standard
legislative goals. Possible reasons—i.e. non-standard legislative intentions—are the
following. (I take as paradigmatic those cases where whatever one does is wrong; the
items in this list may be weakened or qualified in order to account for the other
possibilities.)
(1) The lawgiver may wish to induce in the subject a state of puzzlement (“What
on earth is required of me?”), of frustration, anxiety, a sense of being “entrapped,”33
a nervous breakdown,34 or cognitive dissonance.35 (2) He may want to have a
(purported) reason for punishing her anyway. (3) He may want to induce fear, or
terrorize her. (4) He may have the purpose of humiliating her, or (5) of inducing
feelings of guilt.
These are all non-standard legislative intentions (the list has no pretension of
being exhaustive) which may render instrumentally rational the issuing of inconsist-
ent prescriptions. They all identify cases of abuse of the institution of prescribing:36
specifically, the lawgiver exploits—i.e. makes a strategic use of—the form of
certain action; (2) he intends to make the addressee perform the action as a consequence of his uttering
a sentence; and (3) he intends to make the addressee perform the action (as a consequence of his
uttering a sentence) by virtue of the recognition, by the addressee, of these very same intentions, (1) to
(3). There is, however, no need for us to follow these complications here.
32 Austin (1962: 16); Searle (1969: 60, 64ff.); MacCormick (1973: 105); Hart (1982: 249).
33 Cf. Raz (1979: 222).
34 Fuller (1969: 66).
35 A lawgiver issuing inconsistent prescriptions may want to put the subject in a conflict situation;
and, through the latter, to engender cognitive dissonance. Festinger (1962: 39–40) distinguishes
between conflict (in a choice situation) and dissonance (in a post-choice situation, following the choice
of one of the conflicting options). In conflict, one is pushed (by the conflicting reasons for and against
conflicting options) in opposite directions. When dissonance occurs, after the choice, one is pushed in
one single direction, toward the reduction of dissonance.
36 Cf. Austin (1962: 16); Hart (1982: 247).
140 Bruno Celano
prescriptive discourse in order to bring about a result different from, and irreducible
to, standard guidance.37
Now, it is important to realize that robust power structures, having a firm hold
on their subjects, may in fact come into being and prosper—in families, work-
places, churches, associations of various sorts—around the issuing of inconsistent
prescriptions, and through it.38 Flouting the consistency requirement may be a
perfectly rational way of engendering puzzlement, frustration, fear, guilt, etc., thus
rendering the subjects malleable to the lawgiver’s will.
Power structures that operate in this way. I submit, may well be characterized, in
a straightforwardly intelligible sense, as of a “manipulative” kind (our first condi-
tion, section 3). The non-standard modes of control on our list are ways of forcing
the subject in an artificially created quandary, of inducing in her unpleasant feelings
or emotions, or a contrived experience of uneasiness, and to capitalize on this.
These are ways of, metaphorically, treating subjects as things to be modified as one
sees fit: not treating them as centers of deliberation—i.e. as at least potentially
autonomous agents, who can be (part) authors of their life.39
Considering these various non-standard legislative intentions is, then, a good
way of making determinate, or at least less generic, the idea of (not) treating human
beings as things, and, thus, of seeing under what respects RoL power is opposed to
the manipulation of human beings. Correspondingly, power structures that rely on
the non-standard modes of control on our list prove, in various ways, disrespectful
of the dignity of their subjects (our second condition, section 3).
In acting in these ways, first, the lawgiver does not respect the subjects’ status as
potentially autonomous agents (we have just seen this). By acting in these ways,
second, he may positively intend to dismantle, and succeed in dismantling, the self-
esteem of subjects—provided that the latter are disposed, for whatever reason
(section 4.2), to take seriously the lawgiver’s dictates. This is a further way of not
respecting them. And, third, in all these ways, when apparent—to both lawgiver
and subjects or to a third party (cf. section 4.2)—the lawgiver does not respect the
dignity of subjects because, and insofar as, he shows disrespect for them. (One way
of not respecting someone is to show disrespect for him.) The attitude evinced
through the issuing of inconsistent prescriptions may be rendered as “It doesn’t
37 A further hypothesis is that of a “pluralist” lawgiver, who issues conflicting prescriptions because
he wants different, conflicting values to be pursued by his subjects. (Thanks to M. C. Redondo for
pressing this point on me.) This hypothesis may perhaps be of some interest in accounting for the
conflicts engendered by the various rights, principles, and values which make up a substantial part of
the content of contemporary (especially European) constitutions, and of many supranational instru-
ments. I shall not discuss this issue here, and concentrate on the kinds of “strategic” uses of inconsist-
encies, and the power structures building upon them, discussed in the text.
38 A special case, by no means a marginal one, are “double binds,” generated, within deep or
standing personal relationships, by the repeated occurrence of conflicting directives of different levels of
abstraction. J. Elster (in (1993: 81); see also (1983: 60ff.)) discusses “contradictory injunctions,” such
as “Don’t be so obedient!” or “Be spontaneous!:” their “overt content” “contradicts their pragmatic
presuppositions;” they thus place subjects “in an impossible situation,” requiring of them “an impos-
sible undertaking.”
39 Cf. for this characterization of autonomy Raz (1986: 369).
Law as Power 141
matter what you may think you should do. Calculations are useless.40 I can do
whatever I want of you;”41 and “I don’t care about what you may think I think,
believe, or will. I am free from any discoursive obligation toward you” (“You may
think you are somebody worth talking to, but you are not”). Such messages show
disrespect for those to whom they are addressed,42 and may work as tools in
demolishing their self-esteem.
These are different, cumulable, and mutually reinforcing ways of not respecting
the status of human beings as at least potentially autonomous centers of deliber-
ation and agency. It is in this way that the non-standard modes of guidance on our
list prove disrespectful of the dignity of subjects.
4.2. Compliability (the technology of guilt)

The requirement of compliability may be understood in various ways, according to
how we understand the idea that it should be “possible” to perform the required act
(or, to forbear from acting in the relevant way). So, for instance, the principle may
be understood as imposing the condition that the required act should be logically,
or conceptually possible (thereby ruling out, for example, the injunction to draw a
square circle, and the like). Here, as usual in the literature, I shall understand the
principle as requiring the prescribed acts (or omissions) to be “humanly possible”—
i.e. as requiring them to be, in addition to being logically or conceptually possible,
and physically possible, acts whose performance is generally within the scope of the
abilities and capacities of normal (whatever this may mean) human beings.43
The principle “ought” implies “can,” I assume, is not a logical principle: it is not
logically inconsistent to claim that Xs ought to do A, although they cannot do
A (the former claim does not entail the negation of the latter). Relating to the
field of prescriptive discourse, it expresses, however, a conceptual constraint:
prescriptions flouting this constraint are abuses of the institution, parasitical on
the normal or standard case. This, once again, should be understood in the sense
that conforming to the principle is a matter of rational law-giving.44 “Rational”—
instrumentally rational—with respect to a certain purpose: the purpose of standard
guidance. Typically, understanding an utterance as a prescription involves the
ascription, to the utterer, of a normal legislative intention (a defeasible presump-
tion; section 4.1): whoever prescribes to X that he should do A wants X to do A,
and he wants to make X do A. This is why prescribing to X the performance of A,
and at once openly admitting that A is, for human beings, impossible, defeats the
40 On this, see section 6, sub (2).

41 Not caring about consistency is a sign, and an intimation, of unlimited power. Cf. for a parallel
point about compliability Fuller (1969: 70–1).
42 Cf. Marmor (2004: 32) making a parallel point about compliability.
43 Thus, what is claimed is not that laws should be tailored to suit individual capacities and abilities.
The inference from “cannot” to “therefore, he is under no obligation of doing it” does not hold, when
the “cannot” is due to individual peculiarities. Cf. Hart (1961: 174); Kramer (2007: 130, 166). Rules
are addressed to “normal” (whatever this may mean) subjects.
44 Cf. the discussion in von Wright (1963: ch. 7, especially 119, 122), to which I am strongly
indebted here. Cf. also von Wright (1985: 269).

142 Bruno Celano
prescription. It would be irrational to want somebody do something—and to try to
make her do it—if one believed it to be impossible.
So, let us ask ourselves, as we did in the preceding section, what kind of power
may be served by flouting the compliability condition, and why such forms of
power may turn out to be of a “manipulative” kind, and not respect the dignity of
their subjects.
The answers to these questions are, in fact, the same as in the case of the
consistency condition. This is no surprise, since inconsistency, as defined here
(section 4.1), is a special case of non-compliability: inconsistent directives cannot
be jointly followed (first case) or anyway satisfied (second case). The features
indicated in the preceding section as making power manipulative, and disrespectful,
in the case of inconsistent prescriptions, are, in fact, features that apply to non-
compliable prescriptions generally. The list of the reasons making instrumentally
rational the violation of the requirement—non-standard legislative intentions—
may, however, be enriched. To those listed in section 4.1 we may add the following
two items.
(1) The lawgiver may want to put on trial the subject’s disposition to obedience,
or deference to his authority.
(2) He may wish to make her try to perform the (apparently) required act. This
may happen either because he wants the subject to succeed in the assigned
task to the (by hypothesis partial) extent that she can,45 or because he wants
her to fail. And this, once again, may engender frustration, anxiety, etc.46
One especially important possibility is that of pursuing the strategy of prescribing
the impossible in order to make the addressees feel guilty. Systematically flouting
the compliability requirement may be a way of inducing in the subject the sense of
his inadequacy, and weakness, and of making him feel guilty about that (we may
call this the “technology of guilt”). Logical impossibility (due to inconsistency, or to
other forms of conceptual impossibility) may be a way of achieving this, but, if
we assume the relevant logical relations to be evident to the subject’s eye, and to
be common knowledge among lawgiver and subject, this way of pursuing the
strategy may require too much by way of irrationality on the part of subjects.
Generic non-compliability—prescribing what is (though logically possible) not
humanly possible—is a much more efficacious method (not least because, as
remarked by Fuller,47 the line between what is and what is not humanly possible
is often uncertain). One way of acquiring and exercising power over human beings
is by inducing in them guilt for their (purported) constitutive inadequacy, or
45 Fuller (1969: 71).

46 Non-compliable directives may also be rational in the light of further ends, other than the
guidance, be it standard or non-standard, of human conduct. So, for instance, if one takes the function
of law to be that of providing the resolution of disputes, unfulfillable demands may be perfectly in
order. Kramer (1999: 46–7); (2007: 131). Further, the imposition of strict liability involves a violation
of the principle “ought” implies “can.” Thus, whatever reasons may, in some contexts, make it rational
to impose strict liability, make it rational to issue unfulfillable directives.
47 Fuller (1969: 79).
Law as Power 143
weakness, and setting ourselves as their healers, because we are uniquely authorized
to guarantee them forgiveness for their faults, or because we know how, and are able
to, supplement them in their inadequacy or weakness (we have to manage to make
them believe this, of course; believing it ourselves may help). And one way of doing
this is by issuing prescriptions we know they will not be able to comply with—
setting a standard we know they will not be capable of living up to. That is, we can
do it by flouting the requirement that whoever prescribes wants the addressee do
what he prescribes her to do, and tries, by issuing the prescription, to make her
perform the desired action. In such cases, we do not actually want the addressees to
do what we (seem to) require of them; rather, we want them to fail. It is thanks
to their expected failure that we (mean to) acquire and exercise power (by prescrip-
tions, or any other means) over them. In this way, guidance—non-standard
guidance—may be enabled by non-compliable directives.48
This is the place to introduce, and discuss briefly, a complication implicit in
what has been said already (and which may have perplexed the reader in the
preceding paragraph). It concerns the import of both requirements, compliability
and consistency. The issue is what may be termed the “epistemic dimension” of the
requirements—and of the RoL generally. When we say that the RoL requires
compliability, do we mean that the relevant directives should simply be compliable
(no matter what the parties believe thereabout), or that they should be (perhaps
mistakenly) believed (by both lawgiver and addressee?), or even known to be
fulfillable?49
There are various possibilities; and, correspondingly, different versions of the
compliability condition. I consider only some of them, those that seem to me
significant re our main issue, “law as power.”
The action prescribed, A, may, unknown to the parties, in fact be impossible, or
it may be believed, either truly or falsely, to be impossible. The former hypothesis,
it seems to me, is, for our purposes, uninteresting.
The hypothesis that A is falsely believed to be impossible does not, it seems to
me, bear peculiar implications, apart from the fact that it may happen that A is in
fact performed, and thus discovered, ex post (contrary to what was previously
believed), to be possible. Let us consider, then, the hypothesis that A is truly
believed to be impossible. We can further distinguish three sub-hypotheses,
according to who believes this, either the lawgiver, or the addressee, or both. In
the first of these, issuing a prescription may be a way of making the addressee try to
perform A, fail, and (provided that, for whatever reason, the addressee is disposed to
take the lawgiver’s dictates seriously) of engendering frustration, or guilt.
Suppose now that both the lawgiver and the addressee believe A to be impossible.
We can further distinguish two possible scenarios. The fact that A is impossible
48 J. Elster (1993: 81) mentions, as an example of orders that require from subjects “an impossible
undertaking,” orders that require “willing what cannot be willed” (“Forget it!,” “Don’t even think
about that!”). Elster remarks that such orders, “if taken seriously,” may engender “pervasive guilt.” The
technology of guilt works at its best when the relevant directives concern such internal states as
believing, desiring, and feeling (“Love Big Brother!”). Cf. also Elster (1983: 60ff.).
49 I shall only discuss compliability, but the discussion also applies, mutatis mutandis, to consistency.
144 Bruno Celano
may be common knowledge among lawgiver and addressee (it is out in the open
between them that A is impossible), or not (each one of them believes A to be
impossible, but does not believe the other to believe this).50 In the former case,
prescribing the impossible may be a way of overtly showing disrespect, and (once
again, provided that, for whatever reason, the addressee is disposed to take the
lawgiver’s dictates seriously) of humiliating the addressee, or of demolishing his self-
esteem (“You are not even worthy of a compliable prescription”).51
Often, as remarked, the distinction between what is and what is not humanly
possible will be uncertain (to the lawgiver, to the addressee, to both . . . ). This may
give rise to a vast array of epistemic nuances in the technology of guilt. Generally
speaking, it seems to me, the relevant epistemic and, generally, intentional frame-
work may be largely indeterminate, inchoate, untidy, muddled, even incoherent—
consisting of more or less rational epistemic attitudes and dispositions of the parties
involved (specifically, dispositions of deference toward the lawgiver, on the part of
the subjects).52 Within a pathological parental relation, for example, a child may
feel guilt, or lose his self-esteem, for not complying with prescriptions commonly
known, by her and her parent, to be unfulfillable (and, maybe, issued with that
purpose). On the other hand, irrational lawgivers may well want subjects do what
they believe to be impossible—and, thus, sincerely set for them standards they
know the subjects will not be capable of living up to.
5. Comments
In principle, then, power can be effectively exercised by systematically flouting our

two requirements. Power structures, in various social settings, may flourish around
areas of inconsistent or non-compliable directives, by relying on their psychological
effects (ranging, as we have seen, from puzzlement to guilt).53
A Foucaultian remark may further clarify this. According to Foucault, writes
John Scott, “any form of power, other than mere force or physical repression,
depends on the formation of individuals into subjects with appropriate motives and
desires.”54 Power both requires, and brings about, “subjectivation.” Specifically,
both inconsistency and uncompliability may be means of constituting subjects apt
for subjection to (non-RoL) power. Systematically flouting our two requirements
50 Common knowledge involves an indefinite series of epistemic iterations. It is common know-
ledge that p among X and Y when X knows that p, Y knows that p, X knows that Y knows that p,
Y knows that X knows that p, X knows that Y knows that X knows that p, and so on. For refinements
and details cf. Celano (2011: section 5). But, it seems to me, if we stop short of common knowledge
there is no reason to go beyond the first iteration in the present context.
51 Cf. Fuller (1969: 71, 162).
52 On this cf. also section 6, sub (3).
53 A form of power that may usefully resort to these strategies is what M. Foucault (1981) calls
“pastoral power:” power exercised over human beings on the model of the shepherd’s leading of his
flock.
54 Scott (2001: 92).
Law as Power 145
does not bring about specific motives, beliefs, or attitudes. It may, however, be
useful in inducing or reinforcing a generic habit of submission and docility.
It is not necessary that the strategy be intentional. As we have seen (section 4.2)
the epistemic and, generally, intentional dimension of non-standard guidance may
be indeterminate, inchoate, untidy, incoherent, both on the part of the subjects,
and on the part of lawgivers. It may be enough that the relevant directives are, and
are more or less confusedly perceived as being, unfulfillable or inconsistent, that
the effects are those described (ranging from puzzlement to guilt or fear), and that
these effects feed back on the guidance structure, providing for its maintenance.
The whole process may go on without anybody intentionally practicing a strategic
use of non-compliable or inconsistent legislation—or, specifically, of the technol-
ogy of guilt.
So, the issuing of non-compliable or inconsistent directives can both be strategies
for gaining a secure hold on subjects: ways of laying down the foundations of
robust power structures. Not, however, respectful ways. Rather, they can be ways of
forcing subjects in a no-way-out situation (and, thus, of suspending their agency),
of dismantling their self-esteem, of showing disrespect for them, in order to make it
easier to guide them (by whatever means may be conducive to the end) in the
desired direction. When effective in making the subjects docile to guidance, power
structures of this sort may well be characterized as manipulative. Subjects are not
treated as agents, but as things to be pushed or modified according to the power-
holder’s will. And, in so manipulating subjects, non-standard guidance, as I have
characterized it here (through nonconformity with our two requirements), falls
below the threshold of a minimum of respect for human dignity.
6. Rule of Law power: some central features
The two RoL requirements we have considered illustrate a threefold conclusion:

RoL power is public, rational, non-paternalistic.55
(1) We have considered the possibility of power structures built around the
systematic flouting of the requirements of consistency or compliability. In both
cases, typically, power is non-public, in this sense: the power-holder keeps hidden
something which is essential for the intended workings of the power structure—
namely, the reason why he is issuing inconsistent or non-compliable directives
(in order to punish the subject no matter how she will act; or, in order to induce
feelings of guilt, etc.). If that were to become public, the power structure would
cease to work. Power exercised by flouting the two requirements is liable to be
“unmasked.”
55 These features are grounded, in different ways, in all the RoL desiderata, and how they work
together. Here, we consider them only to the extent that their instantiation is illustrated by the two
desiderata (satisfying the two desiderata is not, however, a sufficient condition; see n. 24).
146 Bruno Celano
The RoL is, in this sense, publicly exercised power: the power structure works, if
and to the extent that the subject understands that, and how, it is working. This
understanding is part of the power structure itself.
In other words, RoL power may be characterized as “public” in the following
sense: the strategies and forms of power which systematically resort to inconsistent
or non-compliable directives are, typically, secret: their explicit avowal would refute
them, both in the sense of making them inefficacious, and in the sense of showing
them to be disreputable. In such “secret” power structures, inconsistent, or non-
compliable, directives are useful, only so long as it is not publicly known why they
are issued.56
Thus (or at least this is the pro tanto conclusion the examination of our two
desiderata leads us to), RoL power works through the understanding, by the
subject, of its workings: the subject understands that something is required, and
what is required, of her, and this understanding is a necessary condition for the kind
of guidance intended—for her doing what she is intended to do, in the way she is
intended to do it.
I conclude that, to the extent that it satisfies our two desiderata, the law is, pro
tanto, a method of social control that works openly and publicly in the production
of the desired behavior. Its mode of operation is such that it has to stand out in the
open for it to work in the intended way.
(2) As we have seen (section 4), power exercised by flouting the two require-
ments can be instrumentally rational. It all depends on what the (non-standard)
purpose of the lawgiver is. There is, however, a specific respect under which RoL
power is rational power: it is power for rational subjects. Specifically, I shall now
argue, power exercised by complying with our two requirements is power for
rational people.
Think of the attitude which may be expressed by the issuing of inconsistent
prescriptions: “It doesn’t matter what you may think you should do, calculations—
or the weighing of reasons for and against alternative courses of action—are useless.
What will happen to you does not depend on your deliberation.” Guidance
through the issuing of prescriptions, when it is standard guidance, is a rational
mode of guidance, in this sense: it is a method of control which relies on—and, in
the first instance, leaves room for—the subject’s calculations and deliberation. The
issuing of inconsistent prescriptions, for the purposes of some form of non-standard
guidance, suspends this. The addressee is not being treated as a rational (calculat-
ing, reasons-weighing, deliberating) agent anymore.
The same holds in the case of non-compliability, where this is an instrument for
non-standard guidance. When non-compliability is common knowledge among
56 Publicity, as I have said (section 2), is usually listed as a further RoL requirement. Are we
conflating different phenomena, blurring conceptual distinctions? No, for two reasons. (1) What the
RoL requires is publicity of the laws. Here, we are considering publicity as a property of the way in
which laws conforming to the RoL influence the subjects’ behavior. (2) Different RoL requirements are
interconnected, in various ways (so, for instance, prospectivity, intelligibility, and publicity are, in the
RoL, mutually related features). So, it is no surprise that consistency and compliability should point to
a facet of the ideal of publicity. Further aspects of this ideal are sorted out in Celano (2011).
Law as Power 147
the parties, the attitude expressed by the lawgiver is very similar: “You are not
worthy of being told what you are supposed to do. Your fate does not depend on
your understanding or deliberation.” When the issuing of non-compliable direct-
ives is an—undisclosed—way of making the subject try to perform the (apparently)
required action, or of inducing feelings of inadequacy and guilt, the subject is, once
again, not being treated as a rational agent.
This leads us to an important point, which has surfaced already at times (sections
4.2, 5), but is worth emphasizing. I have summarily described some non-standard
forms of guidance through the issuing of prescriptions, built upon the violation of
our two RoL requirements. It is important to note that power structures of this
kind may rely, for their efficacy and operation, on irrational attitudes and on
fallacious inferences on the part of the subjects. To take two extreme cases, it is
(barring, perhaps, far-fetched scenarios) a wholly irrational reaction to lose one’s
self-esteem in the face of inconsistent directives, or to feel guilt for not complying
with patently non-compliable prescriptions. Both attitudes may be the upshot
of unconscious drives, misplaced affections, unwarranted fears, or any sort of
irrational attitude. Power structures relying on these sorts of psychic mechanisms
are no less real for that.
Thus—and this is an important methodological point—when we consider the
RoL as a specific mode of power, and when we try to grasp what its features are by
contrasting it with other varieties of power, it is not only power as effective in
controlling rational subjects that we have to contrast it with. There is no reason to
exclude from the comparison forms of (non-standard) guidance that rely on the
irrationality of subjects.
When seen in the light of these possibilities, RoL power—specifically, power
conforming to our two requirements—looks, in a sense, quite naive: X tells Y to do
A, and announces to Y what the consequences of non-compliance will be, in the
expectation that Y will ponder whether to comply or not. It is in this, minimal,
sense that RoL guidance—as illustrated by our two requirements—is guidance for
rational subjects.57
(3) From what has just been said about RoL power as power for rational subjects
it follows that RoL guidance, as illustrated by a power structure that meets our two
requirements, is non-paternalistic. Not in the sense (the focal sense of “paternalis-
tic”) that it rules out interference—or, specifically, coercive interference—in the
activities of subjects in order to promote their own good. Laws conforming to
the RoL desiderata may well be paternalistic in this sense. It is, rather, a matter of
the kind of attitude expressed in the exercise of RoL power.
When the government treats its subjects in accordance with the RoL, it treats
them as adults, capable of making their own decisions on the basis of their
own preferences and their own understanding of the relevant facts. It tells them
57 Cf. Raz (1979: 222): “a legal system which does in general observe the rule of law . . . attempts to
guide [people’s] behaviour through affecting the circumstances of their action. It thus presupposes that
they are rational . . . creatures and attempts to affect their actions and habits by affecting their deliber-
ations.”
148 Bruno Celano
explicitly “I want you to behave in such and such a way; these will be the
consequences—I shall inflict you such and such a harm—in case you don’t; now
it’s up to you.”
This mode of exercising power over a human being should be contrasted with
the way in which children are often treated. In order to make children do what we
want them to do we sometimes tell them lies (“Candy shops are closed now”); we
fake non-existing unpleasant consequences (“The wolf will come and get you”); in
various ways, we distort reality. Or we try, unknown to them, directly to manipu-
late the environment, or their preferences, by working behind their back, so to
speak. Or, again, we rely, in trying to make them do what we want them to do, on
an aura of parental authority, or on symbols. In acting in these ways, we do not
recognize to children the dignity of responsible agents, capable of autonomous
choice; we do not treat them as autonomous agents capable of—and entitled to—
making their own choices on the basis of preferences and beliefs which are in fact
their own. It is characteristic of standard guidance through prescriptions—thus, of
RoL power—that it gives the subject a choice—albeit, often, a forced one.58
Furthermore, in a prescriptive relationship the subject to whom a prescription is
addressed is kept at a distance, so to speak. She is not regarded by the lawgiver as an
appendix to, or an extension of, his own body, as merely a tool, or as one
commodity among others at his disposal, or again as something in the environment
to be manipulated (be it in view of its own good). Causal efficacy on her conduct is
mediated by her own understanding of its being exerted, and how (cf. in this
section, sub (1))—and this is common knowledge between the two. In this way,
too, RoL power is non-manipulative.
These two features—giving the subject a choice (albeit a forced one) and keeping
her at a distance (in the specified sense)—render RoL power very different from the
kind of power parents often exercise over their young children, shepherds over the
flock, propaganda experts over their target, etc. It may be said that, if the supreme
achievement of power consists in securing the willing compliance of its subjects,59
it is characteristic of RoL power that it aims at securing unwilling compliance.
Under both respects, one distinctive feature of RoL power is that rulers regard
their subjects, literally, as addressees—i.e. as subjects capable, and worthy, of being
addressed. To borrow a phrase from Strawson, their dealings with them, as
addressees, are not premised on “objectivity of attitude:” a “purely objective view
of the agent as one posing problems simply of intellectual understanding, manage-
ment, treatment and control.”60
58 This should not be taken as holding across the board. Think, for instance, of threats having a
“Your money or your life” structure. These do not in fact offer the subject the choice they apparently
offer her. In case the subject complies, the gunman will get her money. In case she doesn’t, the gunmen
will get both her life and her money. This is, in fact, no (well-formed) alternative. The latter hypothesis
includes the former—they are not logically independent. The real alternative (and the choice) is
between losing or not losing one’s life: one’s money is lost anyway.
59 Lukes (2005: 10–11).
60 Strawson (1962: 87). “To adopt the objective attitude to another human being is to see him,
perhaps, as an object of social policy; as a subject for what, in a wide range of sense, might be called
treatment; as something certainly to be taken account, perhaps precautionary account, of; to be
Law as Power 149
This is what happens in cases of standard guidance, where (inter alia) our two
requirements are satisfied. Things stand differently within power structures
where flouting consistency and compliability may turn out to be instrumentally
rational. Where, for example, pastoral power is being exerted, inconsistent or non-
compliable directives may be good means in the enterprise of leading the subject
in the desired direction, be it his own good. Or, again, in issuing patently non-
compliable directives in order to induce feelings of guilt power-holders are not
treating subjects as addressees (in the specified sense).
Thus, under all three respects—publicity, rationality, non-paternalism—RoL
power, as illustrated by laws meeting our two requirements, qualifies as a mode of
power which (minimally) fits the dignity of adult, rational, autonomous agents, to
be contrasted with manipulative forms of power.61 When the government treats its
subjects in this way, it recognizes to them the dignity of beings worthy of being
publicly, openly addressed, and of being guided through their understanding of the
way in which power is being exerted over them, and their deliberations on the
merits of the case. In short, it treats them with, and shows them, respect.
This concerns only the form of the relationship, not the laws’ content. It is
compatible with all sorts of disrespect and unjust discrimination. But it positively
is, it seems to me, a valuable feature of laws meeting RoL desiderata.
References
Alchourrón, C. E. and Bulygin, E. (1984). “Pragmatic Foundations for a Logic of Norms,”
Rechtstheorie, 15: 453–64.
Anscombe, G. E. M. (1958). “On Brute Facts,” in G. E. M. Anscombe, Ethics, Religion and
Politics. Oxford: Blackwell, 1981, 22–5.
Austin, J. L. (1962). How to Do Things with Words: The William James Lectures. Delivered at
Harvard University in 1955. Second edition edited by J. O. Urmson and M. Sbisà.
Oxford: Oxford University Press, 1975.
Bobbio, N. (1999). Teoria generale della politica. Edited by M. Bovero. Turin: Einaudi.
Celano, B. (1990). Dover essere e intenzionalità: Una critica all’ultimo Kelsen. Turin:
Giappichelli.
Celano, B. (2010). “Normative Legal Positivism, Neutrality, and the Rule of Law,” paper presented at
the conference “Neutrality and Legal Theory” (Girona, Spain, May 21–2, 2010), <http://ssrn.com/
author=1605026> and <http://www.filosofiayderecho.es/pt/tienda/?page_id=3&lang=eshttp://
www.filosofiayderecho.es/pt/tienda/?page_id=3&lang=es> (accessed March 20, 2011).
managed or handled or cured or trained” Strawson (1962; 79). Strawson writes that “if your attitude
towards someone is wholly objective, then though you may fight him, you cannot quarrel with him,
and though you may talk to him, even negotiate with him, you cannot reason with him.” But this, it
seems to me, downplays what is involved, by way of reasoning with someone, in talking to him, or in
negotiating with him.
61 I say “minimally” because conformity to the RoL, as remarked in section 3 (following Raz),
should be understood as being only a necessary condition for the law to respect human dignity.
150 Bruno Celano
Forthcoming in J. Ferrer Beltrán, J. J. Moreso, and D. M. Papayannis (eds.), Neutrality and Theory
of Law. Dordrecht: Springer.
Celano, B. (2011). “Publicity and the Rule of Law,” forthcoming in L. Green and B. Leiter
(eds.), Oxford Studies in Philosophy of Law, vol. 2. Oxford: Oxford University Press.
Celano, B. (2012). “Hart’s Blind Spot: Il tassello mancante in The Concept of Law,” Rivista di
filosofia del diritto, 1: 405–26.
Elster, J. (1983). Sour Grapes: Studies in the Subversion of Rationality. Cambridge:
Elster, J. (1993). Political Psychology. Cambridge: Cambridge University Press.
Festinger, L. (1962). A Theory of Cognitive Dissonance. Second edition (first edition 1957).
Stanford, CA: Stanford University Press.
Finnis, J. M. (1980). Natural Law and Natural Rights. Oxford: Clarendon Press.
Foucault, M. (1981). “Omnes et Singulatim: Towards a Criticism of ‘Political Reason’,” in
S. McMurrin (ed.), The Tanner Lctures on Human Values. Salt Lake City: University of
Utah Press, 225–54.
Fuller, L. L. (1969). The Morality of Law. Revised edition. New Haven: Yale University
Press.
Grice, H. P. (1957). “Meaning,” in H. P. Grice, Studies in the Way of Words. Cambridge,
MA: Harvard University Press, 1989, 213–23.
Hart, H. L. A. (1961). The Concept of Law. Oxford: Clarendon Press.
Hart, H. L. A. (1983). Essays in Jurisprudence and Philosophy. Oxford: Clarendon Press.
Hart, H. L. A. (1982). Essays on Bentham: Studies in Jurisprudence and Political Theory.
Oxford: Clarendon Press.
Hurd, H. M. (1990). “Sovereignty in Silence,” Yale Law Journal, 99: 945–1028.
Kelsen, H. (1945). General Theory of Law and State. Cambridge, MA: Harvard University
Press.
Kramer, M. H. (1999). In Defense of Legal Positivism: Law without Trimmings. Oxford:
Kramer, M. H. (2007). Objectivity and the Rule of Law. Cambridge: Cambridge University
Press.
Lukes, S. (2005). Power: A Radical View. Second edition (first edition 1974). Basingstoke:
Palgrave.
MacCormick, N. (1973). “Legal Obligation and the Imperative Fallacy,” in
A. W. B. Simpson (ed.), Oxford Essays in Jurisprudence. Second series. Oxford:
Clarendon Press, 171–201.
MacCormick, N. (1985). “A Moralistic Case for Amoralistic Law?,” Valparaiso Law Review,
20: 1–41.
Marmor, A. (2004). “The Rule of Law and its Limits,” Law and Philosophy, 23: 1–43.
Raz, J. (1979). The Authority of Law: Essays on Law and Morality. Oxford: Clarendon Press.
Raz, J. (1986). The Morality of Freedom. Oxford: Clarendon Press.
Raz, J. (1996). “Intention in Interpretation,” in J. Raz, Between Authority and Interpretation:
On the Theory of Law and Practical Reason. Oxford: Oxford University Press, 2009,
265–98.
Schauer, F. (2010). “Necessity, Importance, and the Nature of Law,” paper presented at the
conference “Neutrality and Legal Theory” (Girona, Spain, May 21–2, 2010), <http://
www.filosofiayderecho.es/pt/tienda/?page_id=3&lang=eshttp://www.filosofiayderecho.
Law as Power 151
es/pt/tienda/?page_id=3&lang=es> (accessed March 20, 2011). Forthcoming in J. Ferrer
Beltrán, J. J. Moreso, and D. M. Papayannis (eds.), Neutrality and Theory of Law.
Dordrecht: Springer.
Schiffer, S. (1988). Meaning. Second edition (first edition 1972). Oxford: Clarendon Press.
Scott, J. (2001). Power. Cambridge: Polity Press.
Searle, J. R. (1969). Speech Acts: An Essay in the Philosophy of Language. Cambridge:
Strawson, P. F. (1962). “Freedom and Resentment,” in P. F. Strawson (ed.), Studies in the
Philosophy of Thought and Action. Oxford: Oxford University Press, 1968.
Strawson, P. F. (1964). “Intention and Convention in Speech Acts,” in P. F. Strawson,
Logico-linguistic Papers. London: Methuen, 1971.
Tamanaha, B. Z. (2004). On the Rule of Law: History. Politics, Theory. Cambridge:
Von Wright, G. H. (1951). An Essay in Modal Logic. Amsterdam: North Holland.
Von Wright, G. H. (1963). Norm and Action: A Logical Enquiry. London: Routledge &
Kegan Paul.
Von Wright, G. H. (1968). “The Logic of Practical Discourse,” in R. Klibansky (ed.),
Contemporary Philosophy: A Survey, vol. 1: Logic and the Foundations of Mathematics.
Florence: La Nuova Italia, 141–67.
Von Wright, G. H. (1983). “Norms, Truth and Logic”, in G. H. von Wright, Practical
Reason. Oxford: Blackwell, 130–209.
Von Wright, G. H. (1984). “Bedingungsnormen: Ein Prüfstein für die Normenlogik,” in
W. Krawietz et al. (eds.), Theorie der Normen: Festgabe für O. Weinberger zum 65.
Geburtstag. Berlin: Duncker & Humblot, 447–56.
Von Wright, G. H. (1985). “Is and Ought,” in E. Bulygin et al. (eds.), Man, Law and
Modern Forms of Life. Dordrecht: Reidel, 263–81.
Waldron, J. (1999a). Law and Disagreement. Oxford: Oxford University Press.
Waldron, J. (1999b). The Dignity of Legislation. Cambridge: Cambridge University Press.
Waldron, J. (2007). “Legislation and the Rule of Law,” Legisprudence, 1: 91–123.
Waldron, J. (2008a). “Hart and the Principles of Legality,” in M. H. Kramer et al. (eds.),
The Legacy of H. L. A. Hart: Legal, Political, and Moral Philosophy. Oxford: Oxford
Waldron, J. (2008b). “The Concept and the Rule of Law,” Georgia Law Review, 43: 1–61.
7
A Comprehensive Hartian Theory of Legal
Obligation: Social Pressure, Coercive
Enforcement, and the Legal
Obligations of Citizens*
Kenneth Einar Himma
Law characteristically regulates behavior by creating obligations. Both officials and

citizens are subjects of legal obligations. Citizens are obligated to honor their
contracts and to refrain from violence under most circumstances; these are first-
order obligations defined by primary norms. Judges are obligated to decide cases
under the relevant norms; these are second-order obligations created by recognition
norms.
Hart appears to have at least the beginnings of a comprehensive theory of legal
obligation in the form of his “practice theory of rules.” Hart believes that legal
obligation is a form of social obligation. A social obligation, on Hart’s account, is
defined by a “social norm” accepted and practiced by members of the group and
thought sufficiently important by group members to back with social pressure to
conform. The existence of second-order legal obligations governing the behavior of
officials in their capacity as officials, then, is explained by officials taking the
internal point of view towards the rule of recognition, conforming their behavior
towards that rule, and backing the rule with the threat of social pressure.
Although Hart rejects Austin’s sanction theory of obligation as not accurately
expressing either the sense in which civil law binds or the sense in which officials
are bound, he certainly intimates that first-order legal obligations of citizens are
* I presented earlier versions of this chapter at the University of Bristol School of Law, the Ono
School of Law Conference on Positivism, Democracy and Constitutional Interpretation, the National
Autonomous University of Mexico (UNAM), and the Conference on the Nature of Law: Contempor-
ary Perspectives at McMaster University. I am indebted to those who offered comments and sugges-
tions at these talks. I am also grateful to the following for their very helpful comments on earlier drafts
of this chapter: Larry Alexander, Brian Bix, Patrick Capps, Jules Coleman, Jorge Fabra, Imer Flores,
Mark Greenberg, Nina Guzman, Matthew Kramer, Marc McGee, Julian Rivers, Scott Shapiro, Brian
Tamanaha, and Wil Waluchow. I am especially indebted to Stefan Sciaraffa and Alex Langlinais for
comments that prevented me from making some very obvious mistakes.
A Hartian Theory of Legal Obligation 153
explained by the availability of institutional coercive mechanisms for enforcing

first-order legal norms against citizens. As Hart puts the point, “the typical form of
legal pressure may very well be said to consist in such threats [of physical punish-
ment or unpleasant consequences]” (Hart 1994: 179, 180).
The problem with the view, as I have explained it, is that the practice theory of
rules lacks the resources to conceptually explain the first-order obligations of
citizens. For Hart, social obligations arise from social norms. As Hart understands
the latter notion, social norms are established when and only members of the
relevant group take the internal point of view towards the norm (i.e. accept it as a
standard governing the members of the group) and generally conform their behav-
ior to the norm. The problem here as it pertains to giving a conceptual account of
first-order obligation is that first-order legal norms are not generally social norms as
Hart articulates that notion since they are not conventions practiced by citizens
who take the internal point of view towards them. Rather these obligations are
imposed on them by legal officials through first-order legal norms.
In this chapter, I attempt to remedy this gap in the coverage of the Hartian
theory of social obligation by broadening its scope to include two kinds of social
norms that can create social obligations: (1) norms that govern a group because the
members of the group accept and practice those rules; and (2) norms that are imposed
on a set of subjects who acquiesce and conform to the requirements of the norm.
I retain the use of “social” to pick out both sets of rules to call attention to the
positivist thesis that all laws are created through some kind of social process; a
conventional norm is the outcome of something resembling a deliberative conver-
gence of behavior and attitude on the norm, while other social norms are manufac-
tured through social processes like those set forth by a rule of recognition and
imposed on non-members of the group.
Strictly speaking, this departs from Hart’s account by rejecting his assumption
that it is a necessary condition for a social norm to create an obligation that binds
members of the social group that they take the internal point of view towards the
norm. If, as seems reasonable, valid first-order legal norms requiring a behavior
create legal obligations even in legal systems in which citizens merely acquiesce to
the norms of the system and the de facto authority of the system, the internal point
of view cannot be a conceptually necessary condition for a norm to obligate a
person. There must be some other attitude on the part of the subject that is
sufficient to bind the subject to the rule—“acquiescence,” as I will call it—and
this attitude is consistent with much that Hart states about the existence conditions
for law and legal validity.
The theory of social obligation sketched here also retains what I take to be the
Hartian idea that it is the authorization of social pressure for violations of the norm
that partly constitutes the norm as obligatory. In particular, the theory presented
here explains and identifies the particular binding mechanism for social obligations:
a distinctive reason for action that explains the sense in which a social norm,
including a valid law, has a special binding force and is necessarily, if not conclu-
sively, normative. The authorization of coercive enforcement mechanisms, as Hart
observes, for non-compliance is a form of social pressure.
154 Kenneth Einar Himma
The analysis here presupposes that the concepts central to law are social con-
structs arising from the common core understanding of legal practice by legal
practitioners; I assume, with Raz, that legal practitioners cannot be systemically
mistaken about the nature of the core practices of law. Thus, an adequacy
constraint on any theory of legal obligation is that it conforms to the ordinary
core understanding of judges, lawyers, and legislators. Insofar as a theory of
obligation departs from this understanding, the burden of proof is on the theorist
to explain the unusual move (in conceptual analysis, traditionally understood) of
utterly rejecting the ordinary core understandings of legal practitioners that are
usually thought to help fix the content of our concepts—a burden that, to my
knowledge, dissenters have not even attempted to meet.
This assumption has an important methodological consequence. The starting
point is not some preconception of what a “real” or “genuine” obligation looks like
that so constrains the conceptual inquiry that one might conclude that first-order
mandatory legal norms do not necessarily give rise to legal obligations because they
lack the characteristics of such obligations. The assumption that these norms create
legal obligations is grounded—in and hence treats the notion as constructed by—
our ordinary legal and linguistic practices, and thus requires an examination of legal
practices regarding legal obligation to identify certain characteristics of this practice
that have the properties that all things accurately characterized as obligations have.
So the methodology here is not to look, so to speak, into an objective realm of
concepts to identify the content of mind-independent concepts of obligation and
legal obligation and then determine whether the essential furniture of legal systems
as legal practitioners conceive them have features that conform to this mind-
independent concept of legal obligation. To do that would presuppose some kind
of epistemic access to the space of mind-independent concepts; about the only
important epistemological view that would seem to allow that is some kind of
intuitionism that is even less plausible for an epistemology of conceptual analysis
than it is for a moral epistemology.
Rather, the methodology is to adopt the common views among legal practition-
ers about legal obligation and then to find appropriate mechanisms that would
adequately line up with those common views. The reason for this, again, is that law is
a social artifact or construct, so the concepts central in legal practice would also be
social constructs. If, as Raz believes, legal practitioners cannot be systemically
mistaken about the content of the concepts fixed by their self-conscious adoption
of certain norms of legal discourse and practice, then this seems to be the only
appropriate methodology. The argument given in this chapter cannot be understood
without understanding the methodology and assumptions on which it is based.
1. The problem of law’s normativity
It is frequently thought to be a conceptual truth about law that it is normative in

some sense that requires philosophical explication (i.e. the Normativity Thesis).
Indeed, the problem of explaining law’s normativity is commonly thought one of
the central problems in conceptual jurisprudence. As Andrei Marmor puts it,

“elucidating the conditions of legal validity and explaining the normativity of law
form the two main subjects of any general theory about the nature of law” (Marmor
2011).
It is essential to note at the outset that there are two different interpretations of
the Normativity Thesis:
(A) It is a conceptual truth that laws are normative in the sense that they actually
provide reasons for action; and
(B) It is a conceptual truth that laws are normative only in the weaker sense that
they purport to provide reasons for action.
(A) and (B) differ, then, in an important respect: whereas (A) takes the position that
law cannot fail in providing a reason for doing what the law requires (although that
reason might be outweighed by other reasons for not doing what the law requires),
(B) entails that a law might purport to provide a reason for doing what law requires
but does not succeed in providing any reason at all.
What kind of explication is required of law’s normativity depends on which
view one holds. Insofar as one holds (A), one must give an explication of what
reasons for action law characteristically provides. Insofar as one holds (B), one
must give an explication that shows how it is possible for law to provide reasons
for action. For example, Jules Coleman conceives the problem of law’s norma-
tivity as showing how “to make intelligible law’s claim to alter the set of
permissible actions” and claims “what is missing [from Austin’s theory of obliga-
tion] . . . are the resources that would illuminate law’s claim to obligate” (Coleman
2001: 159). So solving the problem of legal normativity requires showing how it
is possible for law to provide reasons for action—i.e. make intelligible law’s claim
to provide reasons for action—and hence views (B) as expressing the problem of
law’s normativity.
I should point out here that the problem of legal normativity should not be
confused with that of state legitimacy. The problem of state legitimacy is usually
conceived as concerned with identifying the conditions under which it is morally
permissible for a legal authority to coercively restrict behavior. Solving the problem of
state legitimacy is usually thought to require showing how legal norms give rise to
content-independent moral obligations to obey owed to the state. If this can be
shown, the state is thought justified, as a matter of political morality, in coercively
enforcing compliance with those obligations. The problem of legal normativity is
concerned not with how law can or does provide moral reasons for action but how
law provides reasons for action that are distinctively legal and are characteristic of law.
One can have legal obligations to perform acts that violate one’s moral obligations.
In any event, this article assumes that (A) is the correct way to conceive of law’s
conceptual normativity, which is a truism about legal practice among officials and
legal practitioners. It also appears to be a majority view among legal philosophers.
As one leading conceptual theorist observes in an entry on legal obligation in
the most important encyclopedia of philosophy, “[e]very legal system contains
obligation-imposing laws” (Green 2010).1 Indeed, it seems a grave mistake for a
positivist to deny (A) if (1) law is an institution, as positivism presupposes, that is
wholly constructed by certain social practices; (2) our concept of law is fixed by our
ordinary legal talk and practices that define the institution; and (3) it is our concepts
that the philosopher of law wishes to explicate. To deny (A) is to imply that this
common understanding is systemically mistaken—something that cannot be true
of concepts that are ours in the relevant sense.
Further, the denial suggests that our legal practices would have to be radically
changed or recharacterized in ways that seem counterintuitive. If some laws do not
give rise to legal obligations, then the violation of those laws cannot, as a conceptual
matter, justifiably be met with the application of coercive enforcement mechan-
isms; since the legal justification for the application of such mechanisms is the
violation of a legal—as opposed to moral—obligation, legal norms that purport but
fail to give rise to legal obligations cannot, from the standpoint of our ordinary legal
practices, justifiably be enforced. Again, it bears reiterating that the problem here
has nothing to do with the moral legitimacy and justification of enforcing laws or
the moral authority of the law; it is a matter of legal practice, justification, and legal
authority.2
One potentially tempting argument for denying (A) is grounded in the claim
that legal obligations must be “real” in some sense and certain legal norms cannot
establish some acts as “really” obligatory. But it is hard to see what explains why and
how a valid legal norm requiring X to do p would purport, but fail to give rise to a
legal obligation. If “real” refers to obligations that have moral force (the only sense
that I can make of that claim), then that claim, from the standpoint of positivism, is
clearly false. I doubt that this, however, is what is intended by the use of “real;” that
much I can say. But I know of no other remotely plausible alternative analysis of the
claim that legal obligations must be real obligations. That claim is commonly
accepted even though there seems to be no philosophically rigorous explanation
of what it means. Either way, this idea cannot do any meaningful work in a theory
of law and obligation and should be rejected once this is understood. It seems to
amount to empty sloganeering. Thus, the idea that legal obligations must be real
obligations, which would support (B) by denying (A), lacks adequate support.
What I will do here by way of justifying (A) is to point to its ubiquitous role in
legal practice, which strikes me as plausible if not conclusive, and point to a number
of theorists who support (A). Since this falls short of being compellingly persuasive,
the thesis should be construed as conditional: if (A) is correct, then the theory of
1 That this analysis of the conceptual relationship between certain laws and obligations appears in a
respected encyclopedia speaks to its being a majority view among theorists. For example, David Lyons
observes that “ ‘legal obligation’ [is] something that is automatically created by legal requirements and
prohibitions, regardless of their moral quality and history and independent of the specific relation of
the individual to the political system” (Lyons 1993: 98).
2 When I use the term “legal authority” here, I am principally concerned with one conceptual
feature of authority, namely the capacity to issue directives that create obligations that bind subjects.
One utterly uncontroversial claim about authority is that it functions as a mechanism for producing
obligatory directives.
legal obligation advanced here would provide at least part of the answer to the
question of law’s normativity. The trick here is to build a positivistic account of
law’s normativity that explains how law gives rise to legal obligation in a way that is
not only internally consistent but coheres with the analysis given by positivists of
other concepts central to legal practice. This means that the goal is to provide a
theory that conforms to a comprehensive positivist account of core concepts to legal
practice that coheres with the assumption that (A) is true.
2. The theoretical significance of the concept

of obligation in law and legal practice
2.1. Centrality of obligation talk to legal practice
The concept of obligation is ubiquitous in legal practice. For example, a plaintiff in a
contract dispute typically claims the defendant is obligated to perform some act,
while the defendant argues that her non-performance is excused by the plaintiff ’s
own breach of obligation. Likewise, a prosecutor will argue that the defendant
breached some obligation defined by the criminal law, while the defense will argue
that the defendant did not breach such an obligation. Finally, judges frequently
couch their decisions in terms of what some party is obligated to do. In, for example,
Henningsen v. Bloomfield Motors, Inc., the court held that “[i]n a society . . . where the
automobile is a common and necessary adjunct of daily life, and where its use is so
fraught with danger to the driver, passengers and the public, the manufacturer is
under a special obligation in connection with the construction, promotion, and sale
of his cars.”3
The claims that obligations that arise under law are legal in character and hence
conceptually distinct from moral obligations are truisms for legal practitioners.
According to Black’s Law Dictionary:
[A] moral obligation [is a] duty that is based only on one’s conscience and that is not legally
enforceable; an obligation with a purely moral basis, as opposed to a legal one.
It is, of course, arguable that moral obligations are “based only on one’s con-
science;” however, the crucial point here is that the idea that moral and legal
obligations are analytically distinct is so entrenched in law that it is recorded in a
leading mainstream legal dictionary.
This, of course, is not to suggest that moral obligation is irrelevant to ordinary
talk about legal obligation; it is simply to assert ordinary legal talk and practice
presupposes the existence of legal obligations analytically distinct from moral
obligations. Although the content of law and the content of morality frequently
converge, they frequently diverge as well; in such cases, however, the law defines a
legal obligation if not a moral one.
3 161 A.2d 69 (1960), at 85 (emphasis added).

The law regulates behavior by a variety of means, including power-conferring
norms like those governing the creation of binding contracts and wills, but
characteristically constrains the behavior of citizens by creating such obligations.
This is the view that Hart takes. Hart observes, for example, that Austin correctly
assumes that systems of law necessarily create legal obligations:
[T]he theory of law as coercive orders, notwithstanding its errors, started from the perfectly correct
appreciation of the fact that where there is law, there human conduct is made in some sense non-
optional or obligatory. In choosing this starting point the theory was well inspired, and in
building up a new account of law in terms of the interplay of primary and secondary rules we
too shall start from the same idea. (Hart 1994: 82; emphasis added)
Further, Hart asserts that it is a conceptual truth that primary legal norms generally
define legal obligations (some confer legal liberties): “Rules of the first type impose
duties [i.e. primary rules]; rules of the second type [i.e. secondary rules] confer
powers, public or private” (Hart 1994: 80–1). If Hart is correct, then it is a
conceptual truth that law regulates the behavior of citizens, at least in part, by
primary norms that create obligations that are legal in source and character.4
The most plausible interpretation of Hart’s remarks on the features of norms and
obligation is that he intends them as conceptual claims. To begin, it would be
uncharitable in the extreme to suppose Hart rejected Austin’s conceptual theory of
law on the ground that his conceptual theory of legal obligation lacks the resources to explain
obligation, which he regards as a foundational feature of law—without even attempting
an analysis of the concept of legal obligation! The absurdity of attributing such an
oversight to a thinker of Hart’s gifts militates, decisively on my view, against interpreting
Hart this way. At the very least, one would need a very good argument to justify this
insulting conclusion—and I have yet to see one made that is even remotely plausible.
Further, Hart makes many remarks indicating his concern with the conceptual
problem of legal obligation: (1) “[I]t is crucial for the understanding of the idea of
obligation to see that in individual cases the statement that a person has an
obligation under some rule and the prediction that he is likely to suffer for
disobedience may diverge;” (2) “It is clear that obligation is not to be found in
the gunman situation, though the simpler notion of being obliged to something may
well be defined in the elements present there;” (3) “To understand the general idea of
obligation as a necessary preliminary to understanding it in its legal form, we must
4 A legal system without primary legal norms is a conceptual impossibility on Hart’s view. Since it is
a matter of natural necessity that a legal system includes norms reflecting the minimum content of the
natural law, it follows that every legal system includes primary legal norms; see Hart (1994: 199). Hart
also views law’s adoption of coercive enforcement mechanisms as a natural necessity for efficacious
regulation of human behavior: “There are no settled principles forbidding the use of the word ‘law’ of
systems where there are no centrally organized sanctions, and there is good reason (though no
compulsion) for using the expression ‘international law’ of a system, which has none. On the other
hand, we do need to distinguish the place that sanctions must have within a municipal system, if it is to
serve the minimum purposes of beings constituted as men are. We can say, given the setting of natural
facts and aims, which make sanctions both possible and necessary in a municipal system, that this is a
natural necessity; and some such phrase is needed also to convey the status of the minimum forms of
protection for persons, property, and promises which are similarly indispensable features of municipal
law” (Hart 1994: 199).
turn to a different social situation which, unlike the gunman situation, includes the
existence of social rule; for this situation contributes to the meaning of the statement
that a person has an obligation in two ways;” (4) “The statement that someone has
or is under an obligation does indeed imply the existence of a rule;” and (5) “Rules
are conceived and spoken as imposing obligations when the general demand for
conformity is insistent and the social pressure brought to bear upon those who
deviate or threaten to deviate is great” (Hart 1994: 85–6; emphasis added).
There are three points worth noting here. First, as an interpretive matter, Hart
explicitly commits himself to adopting as a conceptual “starting point” Austin’s
view that “where there is law, there human conduct is made in some sense non-
optional or obligatory:” “In choosing this starting point the theory was well
inspired, and in building up a new account of law in terms of the interplay of
primary and secondary rules we too shall start from the same idea” (Hart 1994: 82).
Hart clearly regards an analysis of the concept of legal obligation as foundational
(a “starting point”) to providing a conceptual theory of law and hence understands
the concept of obligation as being in need of philosophical explication.
Second, Hart seems to be conceiving the problem of law’s normativity as the
problem of explaining how law necessarily provides reasons for action, (A) in
section 1, and not as the problem of showing merely how it is possible for law to
provide reasons for action. Hart seems to believe that it is a conceptual truth that
obligations provide reasons for action: he characterizes the notion of obligation as
characteristically “involving sacrifice or renunciation” (Hart 1994: 87), something
that can only be explained by obligations being or entailing reasons for action.
Similarly, Hart argues that “where rules exist, deviations from them are not merely
grounds for a prediction that hostile reactions will follow . . . but are also a reason
or justification for such reaction” (Hart 1994: 84). If deviations from an obligation
N by A provide other persons for reasons for action in the form of justifying the
expression of censure or hostility for the deviation, then the relevant N must
provide A with a reason for conformity; it makes no sense to condemn someone
for something they had no reason to do. “There is no reason to do p” is a relevant
answer to the question “Why shouldn’t you do p?”
This is not an uncommon claim among philosophers of law. For example,
Joseph Raz views the relationship between ought-claims and reasons as follows:
[S]tatements of the form “x ought to ” are logically equivalent to statements of the form
“There is a reason for x to .” . . . Every conclusion entailed by one statement is also entailed
by the other. . . . General “ought” sentences are often used to assert that there is a case, which
is not necessarily a conclusive one, for acting in a certain way. (Raz 1990: 29).
In any event, it is important to realize that Hart’s view that it is a conceptual truth
that law creates legal obligations, and Raz’s view that obligations neccessarily
express reasons, entail that it is conceptual truth that law provides reasons for
action. This latter claim entails nothing about the strength of the reason law
provides relative to other reasons (e.g. moral) and is consistent with law’s norma-
tivity being outweighed, by other salient normative considerations.
The reason here is a normative reason, and not necessarily a motivational reason.
The difference between the two is this. The distinguishing feature of the idea that
obligations give rise to normative reasons is that the mere claim that the law
requires -ing necessarily counts as a reason for -ing. The distinguishing feature
of the idea that obligations give rise to motivational reasons is the implausible claim
that the law requires -ing entails that some people are inevitably motivated
by the law.5
Third, the denial that, as noted above, law necessarily creates legal obligations
seems to be inconsistent with legal practices utterly central to modern municipal
legal systems resembling that of the US and England. To deny this, then, is to say
that legal practitioners are systematically mistaken about the nature of what they
do. This is problematic, as noted above, because conceptual analysis of our legal
concepts is grounded in what we take to be legal practices that are central to law.
Moreover, the idea that we are systemically mistaken about our legal concepts and
basic legal practices seems to entail something resembling an error theory of law
because it would entail denying the shared core understanding of legal practitioners
on all the corresponding legal practices. Justifying an error theory of law, such as is
entailed by (B), requires satisfying a very difficult burden of proof on the propon-
ent. As I indicated above, I have never seen even a cursory attempt to justify the
denial of (A)—much less one that would satisfy the evidentiary burden for an
adequate defense of rejecting (A) in favor of (B).
2.2. A conceptual relation between legal obligation and mandatory

legal norms: the Equivalence Thesis
The centrality of legal obligation to ordinary practice is neither mysterious nor
surprising. Legal norms characteristically require or prohibit certain behaviors of
subjects, and the notion of a legal requirement is conceptually related by ordinary
linguistic practices to the notion of a legal obligation in the following way: the
phrase “is legally obligated to do a” is extensionally equivalent to “is required to do a
by a legal norm” such that these phrases can be substituted for the other in any
sentence without change of truth value. Though such substitutions can result in
subtle changes of meaning or intension, they are not significant enough to change
the truth value.
This fact about our practices and conceptual usages, however, has an important
consequence regarding the logical relationship of certain obligation-statements and
norm-statements. It implies that the formula “P is legally obligated to do a” is
logically equivalent to “P is required to do a by a legal norm” and hence that the
following is presupposed or implied by ordinary practices regarding the terms “law”
and “obligation:”
5 I am indebted to Alex Langlinais for making me realize that I needed to explain and use this
distinction.
The Equivalence Thesis: It is a conceptual truth that P is legally obligated to do a
in circumstance C if and only if there is some legal norm that, in circumstance C, requires
P to do a.
The Equivalence Thesis asserts that, as a conceptual matter, mandatory legal norms
(i.e. norms requiring some act or omission) create legal obligations. While a
mandatory legal norm might not give rise to a moral (or “real”) obligation, it creates
a legal obligation if, as ordinary usage and legal practice suggest, the Equivalence
Thesis is true.
At the risk of belaboring the point, this seems to be the view that Hart takes.
Hart rejects John Austin’s command theory of law on the ground that the materials
Austin provides cannot explain the fact that the requirements of law create obliga-
tions. On Hart’s view, since a command backed by force, by itself, can oblige but
not obligate compliance, Austin’s theory fails. But the inability of Austin’s account
to explain obligation is a defect only insofar as the concept of law is part of a
network of concepts that include logical interconnections with the concept of
obligation. It is not a criticism of a conceptual theory of law that it cannot explain
some non-conceptual, purely contingent feature of law; conceptual theories must
explain only those features of law that are conceptually required for something to
count as an instance of “law.” That Hart rejected Austin’s conceptual account of
law because it could not make sense of legal obligation adds to the textual evidence
adduced above that shows Hart believes it is a conceptual truth that at least certain
kinds of law define obligations.
Conversely, where there is no legal norm of any kind requiring behavior B, there
is no legal obligation to B because not-B is legally permissible. This is true not only
of a situation where the law is agnostic with respect to B in the sense that no valid
legal norm says anything about the performance or non-performance of B, but also
of a situation where the law encourages the performance (or non-performance) of
B but does not, in any theoretically significant sense, make the performance
(or non-performance) of B mandatory. For example, the enactment of a statement
that the government encourages citizens to be charitable might fairly be character-
ized as a “law” if enacted by the proper procedures and recorded in a book of
statutes but it does not give rise to anything fairly characterized as a legal obligation.
3. The concept of obligation
If ordinary talk or legal practice is any indication, there are different types of
obligation. We distinguish, for example, moral, social, and legal obligations and
speak as if these types of obligation are conceptually distinct. Even so, many
theorists believe they are instances of the same general type. As Joseph Raz puts
it: “normative terms like ‘a right’, ‘a duty’, ‘ought’ are used in the same sense in
legal, moral, and other normative statements” (Raz 1979: 158). Likewise, Hart
observes, “[m]oral and legal rules of obligation have . . . striking similarities enough
to show that there common vocabulary is no accident” (Hart 1994: 172). While
moral, social, and legal obligations differ in important ways, there are certain
elements essential to the notion of obligation and these elements are present in
moral, social, and legal obligations.
Legal and moral obligations presumably have different properties, but both satisfy
the application-conditions for the concept-term “obligation” in the following
sense: satisfaction of the application-conditions for “obligation” will be necessary
(though not sufficient) for something to count as either a “legal obligation” or a
“moral obligation.” If so, then the set of application-conditions for “obligation” will
be a subset of the set of application conditions for “moral obligation” and “legal
obligation.”
If this is correct, then we cannot understand the concept of legal obligation
without understanding the general concept of obligation. In sections 3.1 to 3.4,
I will briefly sketch what I take to be the central elements of the general concept of
obligation.
3.1. Obligations and mandatory prescriptions

Obligations are associated with prescriptions, which correlate with claims about
what some specific person (particular prescriptions) or some class of persons (general
prescriptions or, more commonly, norms including rules and principles) ought to do
in some state of affairs, but not every prescription expresses or implies an obligation.
Although all prescriptions are concerned to commend some behavior (or abstin-
ence), not all prescriptions require them; there are things I ought to do I am not
obligated to do. On one view of charity, a person should give to charity but she is
not obligated to do so.
The only prescriptions that create or express obligations are mandatory
prescriptions—i.e. prescriptions that require some act rather than merely urge its
performance. That is, it is a necessary condition for someone’s being obligated to
perform some act that there is a prescription that expresses a requirement and hence
an obligation to perform that act. If there is no mandatory prescription requiring a,
then there is no obligation to perform a; the claim that a is obligatory but not
required by a mandatory prescription seems self-contradictory. It therefore appears
to be a necessary condition for P to be obligated to do a that there is a mandatory
prescription that requires that P do a.
3.2. Obligations as reasons

Obligations seem to correlate with normative reasons. On this view described
above, the claim that X has an obligation to do a implies that X has a normative
reason to do a. If Y asks X for a justification for X’s doing a, “X was obligated to do
a,” if true, is always relevant in assessing whether doing a was justified from the
standpoint of practical rationality—i.e. whether X should have done a.
The reason can be moral, but need not be. Some obligations are associated with
moral reasons but not all obligations are. If, as many theorists believe, it is not true
that the status of a norm as law does not afford a prima facie moral reason to obey it
even in reasonably just states, then it is reasonable to think that it is not a

conceptual truth about law or legality (as opposed to legitimacy) that one has
even a prima facie moral reason to obey law; clearly, there can be no prima facie
moral reason to obey laws or legal systems that are wicked. There are, of course,
clearly other kinds of reasons—although the number of different types of “basic”
reason (i.e. reasons that are irreducible to other reasons) are limited. Although a
person might lack a moral reason to obey wicked laws and legal system, she might
have a normative prudential reason (her safety) but one that would be defeated by
normative moral reasons, given the wickedness of the law.
It is very difficult to think of any other basic reasons than prudential and moral
reasons. Perhaps there are aesthetic reasons as well. But if there are no other basic
reasons, then every other kind of reason, including legal reasons, will ultimately be
“compound” in character, ultimately constituted by some combination of members
of the set of basic or atomic reasons.
The reason might be conclusive, but it need not be. It seems that, as an objective
matter of practical rationality, we have a conclusive reason for doing what we are
morally obligated to do all things considered.6 I have a reason not to torture another
innocent person no matter what else might be true and hence a conclusive reason
for not doing so. But whatever prudential reasons Nazis may have had to do
morally wicked things, they were clearly not conclusive; taking into account the
relevant moral reasons, they had a conclusive reason not to do these things. The
idea that a reason is normative, again, tells us nothing about its weight relative to
other reasons.
If ordinary talk is any indication, obligations are reasons. Again, it is always a
relevant consideration in justifying the performance some act a or, relatedly, in
deliberating whether to do a that one has an obligation of some kind to do a.
“Because I had an obligation to do a” might not be an adequate answer to the
question “why did you do a?” It might be false that I had an obligation to do a, or it
might be true that I had such an obligation but it was outweighed by a more
important obligation. But if, as seems reasonable, only reasons can practically
justify an act, then obligations are reasons. Genuine obligations are necessarily
normative and hence are normative reasons for action.
3.3. Obligations as exclusionary

Obligations are defined by valid mandatory prescriptions, and mandatory prescrip-
tions, as a conceptual matter, are “exclusionary” in the following stipulated sense:
they define limits on what kinds of consideration justify non-performance. There
are a couple of mechanisms by which a mandatory prescription might limit the
scope of justified non-performance. It could be that mandatory prescriptions
simply reject, as irrelevant, certain kinds of consideration as constituting valid
6 Indeed, some theorists hold that it is a conceptual truth that moral obligations trump other kinds
of obligation in the event of a conflict; that is, morality defines a set of norms that are “supreme” and
“categorical” (Frankena 1966).
reasons for non-performance. Alternatively, a mandatory prescription might simply
defeat these considerations by outweighing them. Either way, the exclusionary
character of mandatory prescriptions, as I have described it, helps to explain why
the relevant acts are properly thought of as mandatory or required: an act that
people are generally free not to perform because non-performance is trivially
justified is not accurately described as an act that is required by the relevant
prescription.
Indeed, it appears that the claim p is required by a mandatory prescription N is
inconsistent with the claim that non-performance of p can be justified under N, as a
general matter, by purely prudential considerations—in much the same way that
the claim that p is a bachelor is inconsistent with the claim that p is married. The
idea that obligations are defeasible by anyone’s prudential interests, no matter how
trivial, is as incoherent as the idea that some bachelors are married. It might be true
that in some cases a person’s prudential reasons might rise to a level that justifies not
doing p. But this does not imply that prudential reasons sometimes defeat the
reasons provided by a mandatory prescription; such cases might be characterized as
involving exceptions to the prescription N.
No claim is made here about the structure of the reasons provided by law—that,
for example, they are “preemptive” in the Razian sense. The idea that mandatory
prescriptions are exclusionary claims or presupposes nothing about the structure of
the relevant reasons. The claim is merely that a mandatory prescription is exclusion-
ary in the limited sense of excluding or defeating certain stories as justifying or
excusing non-performance.
3.4. The special normative force of obligations: obligations as binding

The concepts of obligation and wrongness are related to the concept of being
(normatively) bound. Obligation-talk is frequently couched in terms of a relation-
ship in which the subject of the obligation is bound to the norm. As Hart puts
the point, “The figure of a bond binding the person obligated . . . is buried in the
word ‘obligation’” (Hart 1994: 87). Obligations, according to ordinary intuitions,
bind us.
In what sense? The term “must” (and, less frequently, the term “shall”) is
frequently used to express that we have an obligation—and are hence bound—to
perform some act. We may do what is permissible and should do what is good, but
we must do what is obligatory.
It might be tempting to explain the concept of bound in terms of some sort of
psychological or physical compulsion. The idea here is that persons are bound by a
rule creating an obligation in the sense that they are psychologically or physically
“unfree” to do other than what the rule requires. But not every obligation, as a
conceptual matter, is supported by compulsion of this kind. Many persons do not
feel psychologically compelled (i.e. psychologically unfree) to satisfy moral obliga-
tions. Further, there are many obligations not supported by physical compulsion;
we are not physically compelled (i.e. unfree in some physical sense) not to lie. Here
coercion and compulsion, it should be remembered, are two different things: a
gunman coerces me with the threat of death but, other things being equal, cannot
compel me to obey. Coercion leaves me with a free choice in a wrongfully imposed
forced choice situation, whereas compulsion renders me unfree to refrain from
performing the compelled act.
As that concept has been explained in this chapter, the exclusionary character of
obligations, by itself, lacks the resources to explain the binding quality of obliga-
tions. The claim that a mandatory norm is exclusionary says no more than that the
content of the norm is such that it disqualifies certain stories as justifying non-
performance. But the claim that a mandatory norm binds us is a claim about its
normative force; this is the point of the metaphor of a bond that ties us to rule.
Simply knowing that the content of a norm excludes certain considerations as
justifying non-performance does not tell us much, if anything, about the nature of
this bond or the special normative force that it has. Indeed, it does not even tell us
whether a norm that functions this way has any normative force because it tells us
nothing about whether the norm is valid or applicable. Invalid mandatory norms
are exclusionary in this limited sense, but they have no normative force and hence
do not bind.
4. Hart’s theory of social obligation
Legal obligation, as conceived by Hart and most positivists, belongs to a special

class of obligations. Since positivism explains law as a set of social practices, the
concept of obligation applicable in legal practice must itself be fully explicable in
terms of social practices. Legal obligation, then, is a species of social obligation.
A full explanation of the concept of legal obligation, then, requires an explanation
of the concept of social obligation, which must harmonize with the explication of
the general concept of obligation.
4.1. Social norms

The first element is straightforward. Although not every social norm gives rise to a
social obligation (e.g. some create social powers), social obligations arise under
general social prescriptions—or social norms, which are created, as Coleman by a
convergence of attitude and behavior. Persons in the group converge on taking the
internal point of view towards the norm, accepting it as a standard that governs the
behavior of people in the group, and generally conform to its requirements. Thus, if
people in the group (1) self-consciously converge in accepting the norm (this need
not be for moral reasons); (2) converge in general conformity to the norm; and (3)
converge in taking a critical reflective attitude using the norm to evaluate the
behavior of other members of the group, then it is, on Hart’s view, a social norm
governing behavior in the group.
4.2. Social norm and reasons

It is a conceptual truth that a social norm, as Hart explains it, gives rise to reasons
for action. Taking the internal point of view towards a mandatory norm, on Hart’s
view, involves a durable commitment to subject one’s own behavior to governance
of the rule and to evaluate the behaviors of other people according to the rule.
Someone who genuinely commits to subjecting her behavior to the rule will accept
and participate in a host of normative practices regarding the rule—including
practices that treat members of the group, including herself, as obligated. As we
will see, this is not the only reason that obligatory social norms provide for those
who are members of the relevant group but it is one such reason.
This suggests that persons who take the internal point of view towards a manda-
tory social norm and thereby commit themselves to conforming to the norm’s
requirements have a reason for action. The idea that I have committed myself to
doing something provides a reason for doing what I have committed myself to doing.
Again, it bears emphasizing that this entails nothing about the strength of the reason
the commitment provides. If a person has committed herself to doing something
morally wrong, call it p, the reason for doing p is outweighed or otherwise defeated by
moral reasons; indeed, the weight of the reason for doing p might be negligible. But
the fact that a person commits herself to doing p is always relevant in answering the
question “Why did you do p?” It therefore counts as a reason for action—although
the reason might be quite weak in comparison to other relevant reasons.
4.3. Social norms as exclusionary

This feature requires little discussion. Hart seems pretty clearly to take the position,
as he must, that mandatory social norms are exclusionary in the sense described
above. To see this, consider again the following remarks from Hart:
[I]t is generally recognized that the conduct required by these rules may, while benefiting
others, conflict with what the person who owes the duty may wish to do. Hence obligations and
duties are thought of as characteristically involving sacrifice or renunciation. (Hart 1994: 87)
This seems to express the view that mandatory social norms are exclusionary. The
claim in the quoted text above is that in cases where obligations conflict with what a
subject wishes to do, the obligation wins. Obligations could not “characteristically
involv[e] sacrifice” unless non-performance cannot be justified by “what the person
who owes the duty . . . wish[es] to do.” It follows that mandatory social norms
exclude or defeat a subject’s desires and are hence exclusionary at least with respect
to a subject’s desires.
4.4. How social obligation binds

While unilateral acceptance alone can explain a person’s adoption of a social norm
as functioning as exclusionary, unilateral acceptance, by itself, cannot explain the
normative force of the obligations to which social norms give rise. After all,
unilateral acceptance can always be given and withdrawn at will, and if that is all
there is to the story, it is hard to see how a durable social obligation could arise. If
the commitment is durable, it provides a reason for abiding by the norm but does
not by itself explain how social obligation and its distinctive normative force arise.
What explains the binding (and hence durable) quality of a social obligation owed
by a member of the social group is, in part, the attitudes of other members of the
social group towards non-compliance.
Hart seems to explain the binding character of social obligations in terms of
considerations ordinary persons are likely to regard as having normative signifi-
cance. According to Hart, “[r]ules are conceived and spoken of as imposing
obligations when the general demand for conformity is insistent and the social
pressure brought to bear upon those who deviate or threaten to deviate is great”
(Hart 1994: 85–6).7 Social pressure in the form of a hostile reaction is something
people with ordinary psychological characteristics tend to regard as having norma-
tive force. Not everyone responds in the same way to social disapproval, but it is an
empirical fact that ordinary persons tend to dislike criticism and hostility and are
willing to take minimal steps to avoid it.
A couple of points deserve attention. First, deviating behavior under the norm is
generally regarded as a reason or justification for the application of social pressure.
The claim is not just that, as a general matter, deviating behavior correlates with
social pressure. Rather, it is that members who accept the rule regard the rule as a
reason for applying social pressure: “For [those who take the internal point of view
towards a rule], the violation of a rule is not merely a basis for the prediction that a
hostile reaction will follow but a reason for the hostility” (Hart 1994: 90). This will
be true, as a conceptual matter, for any form of social obligation, on Hart’s view,
including legal obligation: violation of a legal obligation functions as a justification
for the application of institutional coercive enforcement mechanisms.
Second, the claim is not that social pressure is sufficient for social obligation;
after all, the gunman exerts social pressure on his victim. Rather, it is that a
convergence of attitude and behavior on a rule, together with the appropriate kind
of social pressure, constitutes the norm as obligatory. Such pressure is likely sup-
ported by a belief that it is warranted (though not necessarily morally warranted),
which is related to two factors: (1) the acceptance of the social norm; and (2) the
belief that the norm is important because “necessary to the maintenance of social
life or some highly prized feature of it” (Hart 1994: 87).
Hart’s explanation of social obligation can be summed up as follows:
Hartian Theory of Social Obligation (HTSO): X has a social obligation to do p if and
only if
(1) members of the relevant group converge in attitude and behavior on a mandatory norm
N governing X that requires X to do p ; and
7 Such social pressure “may take only the form of a general diffused hostile or critical reaction”
(Hart 1994: 86), but may also rise to the level for “physical sanctions” (Hart 1994: 86); in this latter
case, the rules are properly regarded as a “rudimentary” or “primitive” form of law (Hart 1994: 86).
(2) N is supported by significant social pressure because
(3) N is thought important because necessary to the maintenance of social life or some
highly prized feature of it.
According to HTSO, it is the presence of the appropriate social pressure in a context
that includes the existence of a practice along with certain beliefs about the
importance of the norm that explains the sense in which the obligation is, as a
conceptual matter, binding: “social pressure appears as a chain binding those who have
obligations so that they are not free to do what they want” (Hart 1994: 87).8 No matter
how important a social norm N might be thought by relevant members of the
group, it is incorrect to characterize it as defining an obligatory and hence binding
requirement if not supported, in some way, by the appropriate level of social
pressure. As Hart puts the view, such pressure is the “primary” characteristic of
obligation (Hart 1994: 87).
Indeed, it is important to note that Hart seems to view pressure of some kind as a
necessary feature of any kind of obligation—and, most likely, the one that explains
the distinctive normative force of the relevant obligation. To begin, Hart believes
that social pressure in the form of coercion is at the very least a naturally, if not
conceptual, necessary feature of legal obligation:
There are no settled principles forbidding the use of the word “law” of systems where there
are no centrally organized sanctions . . . . On the other hand, we do need to distinguish the
place that sanctions must have within a municipal system, if it is to serve the minimum
purposes of beings constituted as men are. We can say, given the setting of natural facts and
aims, which make sanctions both possible and necessary in a municipal system, that this is a
natural necessity; and some such phrase is needed also to convey the status of the minimum
forms of protection for persons, property, and promises which are similarly indispensable
features of municipal law. (Hart 1994: 199; italicized emphasis added)9
Similarly, Hart views pressure of some kind a necessary feature of moral obligation,
which he seems to regard as a form of social obligation: “A further distinguishing
feature of morality is the characteristic form of moral pressure which is exerted in its
support” (Hart 1994: 179; emphasis added). On his view, this is a conceptually
necessary feature of moral obligation: “The fourfold criteria . . . we used to distin-
guish [moral obligation] from other forms of social standard or rule (importance,
immunity from deliberate change, the voluntary character of moral offences, and
the special form of moral pressure) are in a sense formal criteria” (Hart 1994: 180–1;
italicized emphasis added).
Social pressure of some kind seems to be partly constitutive of any kind of
obligation and would explain the normativity of the various kinds of obligation;
as Hart seems to put this point, “[m]oral and legal rules of obligation . . . are alike in
8 It is again worth noting that this quote supports the view that Hart believes that social pressure
constitutes the special binding force of social obligation. It is the social pressure that “appears as a chain
binding those who have obligations so that they are not free to do what they want” (Hart 1994: 87).
9 Indeed, he states that “the typical form of legal pressure may well be said to consist in such threats
[of ‘physical punishment or unpleasant consequences’]” (Hart 1994: 180, and 179 for the language in
the brackets).
that they are conceived as binding independently of the consent of the individual
bound and are supported by serious social pressure for conformity” (Hart 1994:
172). And here it is worth noting that the relevant legal rules of obligation are first-
order mandatory legal norms because they bind without the consent of the subject;
the social rule of recognition, in contrast, binds, in part, because subjects take the
internal point of view towards the rule. The internal point of view involves a
stronger attitude than just the consent of the subject; it requires a convergence on
accepting the relevant norm as governing their behavior, together with other beliefs.
But this normatively stronger attitude clearly involves consent as a relevant con-
stituent.
This implies neither that every person feels the force of the social pressure that
makes a social norm binding nor that any person should feel this force. The claims
here are quite limited. They are purely descriptive because they make no claims
about what people should regard as reasons. Further, they make no claim about
what any particular person in a social group might feel in response to social
pressure; as Hart points out, “there is no contradiction in saying of a hardened
swindler . . . that he had an obligation to pay the rent but felt no pressure to pay”
(Hart 1994: 88). The claim is significantly weaker: as an empirical matter, people
tend to care about social pressure enough to modify their behavior in many
circumstances.
5. Why HTSO cannot ground a comprehensive

theory of legal obligation
5.1. Second-order legal obligation as defined by

a social rule of recognition
Ultimately, there are two conditions, on Hart’s view, necessary and sufficient for
the existence of law and legal obligation. First, officials converge in taking the
internal point of view towards and conforming to a conventional rule of recogni-
tion. Second, citizens generally comply with the rules validated by the conventional
rule of recognition. First- and second-order mandatory norms in such a system
define legal obligations. In particular, mandatory recognition norms define the legal
obligations of officials.
Recall that Hart does not argue it is unilateral acceptance that binds an official to
the rule of recognition; that would be problematic because unilateral acceptance
does not provide anything that necessarily has independent normative force given
what we know about the psychology of ordinary persons. Hart argues instead that it
is the joint acceptance by officials together with social pressure on each to conform to
the rule of recognition that together warrant characterizing the rule of recognition
as being “obligatory.”
Such pressure is likely to have normative force for officials because they can be
presumed to care about what other officials think. Voluntary membership in a social
group governed by norms signals that the member regards at least some of the
beliefs and actions of the other group members as having significant motivational
force. It is, thus, reasonable to think that someone who seeks out membership in a
social group, at least if their motivations are sincere and non-subversive, will regard
such pressure to conform as having significant motivational force.
This does not imply that the motivation for conforming to a social norm must be
explained in terms of a desire to avoid the social pressure. I assume that most people
want to avoid the condemnation accompanying a murder conviction and hence
regard the prospect as having motivational force—and this includes people who
commit murders. But the motives that explain why most people do not commit
murder make no reference at all to these prospects. What explains why most people
do not commit murder is, in part, a subjective moral reaction to murder (any
decent person would be horrified at the thought of committing such an act) and a
lack of extreme anger and hostility.
The point of these sorts of social mechanisms in Hart’s analysis, then, is not to
explain why officials accept the rule of recognition. Officials who take the internal
point of view towards the rule of recognition are presumably motivated to conform
to the rule by whatever desires brought them to officialdom to begin with. While
officials would also presumably want to avoid the disapprobation of other persons
in the relevant groups, Hart is not committed to explaining their behavior in terms
of some necessary motivation to avoid such social pressure. Social pressure contrib-
utes to the explanation of how the rule of recognition obligates. It is no part of
Hart’s theory to suggest that such social pressure explains why officials accept the
rule. Indeed, the social pressure is exerted only on members of the group; and it is
acceptance of the rule that constitutes one as a member of the relevant group.
In any event, if it is reasonable to think that there are social recognition norms
that legally obligate officials, the Hartian explanation of how social norms obligate
is a plausible one. Since the notion of legal obligation is largely constructed by our
legal practices, it is appropriate to look for features of legal practice that exhibit
characteristics that would meet the adequacy constraints on the general concept of
obligation and hence could serve as the foundation for an explanation of legal
obligation. Joint acceptance of, and general conformity to, a social rule of recogni-
tion that is backed by social pressure because the rule is thought to protect
important aspects of social life seems to do just that job insofar as one rejects, as
one should, the rarely explained assumption that social and legal obligations must
be “real” or “genuine” obligations.
5.2. A potentially fatal gap in the Hartian account: explaining

first-order legal obligation
Hart’s theory of second-order obligation (HTSO) will not explain first-order legal
obligation. Merely showing officials can obligate themselves through some mech-
anism does not show that their acts qua officials can obligate citizens. The claim that
you and I have obligated ourselves to behave in a particular way does not entail any
claim about the obligations of other people.
The problem here arises for Hart because he holds that (1) it is a necessary
condition for a person to be obligated by a social norm that she take the internal
point of view towards the norm and (2) it is not a conceptual truth that citizens take
the internal point of view towards the rule of recognition—or, for that matter,
towards the authority of the legal system or towards the norms valid under the rule
of recognition. If it is a conceptual truth that mandatory valid first-order norms
create legal obligations owed by citizens, then either (1) or (2) is false. Since (2) is
clearly true, (1) must be false. Citizens are not members of the relevant group—i.e.
the officials taking the internal point of view towards the rule of recognition. Thus,
if mandatory primary legal norms obligate citizens, those obligations cannot be
explained by HTSO.
Of course, non-members might be obligated to follow rules of groups to which
they do not belong. Non-Muslims are required to abide by certain conventions that
Muslims have accepted regarding behavior inside mosques, but this is explained by
other standards to which non-Muslims are subject; non-Muslims have a duty to
respect those conventions when in mosques. Since admission to mosques is condi-
tioned on consent to abide by certain standards, one should not enter a mosque
unless prepared to abide by the relevant standards.
But nothing like this is true of the legal obligations of citizens. There is no
standard external to law that would create a legal obligation on the part of citizens
to obey first-order mandatory legal norms. In a legitimate legal system, there is an
external standard that might entail that citizens have an obligation to obey manda-
tory legal norms. But such a standard would be moral and hence would create, at
most, a moral obligation on the part of citizens to obey mandatory legal norms.
Accordingly, HTSO lacks the basic resources to explain the legal obligations of
citizens. The significance of this problem should not be understated. Assuming the
Equivalence Thesis is true, the conceptual intimacy of the relationship of legal
obligation to legality ensures that the success of positivism’s theory of legal validity
depends on the success of its theory of legal obligation. If it is a conceptual truth
that legal norms define legal obligations and positivism proves incapable of produ-
cing a viable story about legal obligations, then this ultimately casts doubt on
positivism’s story about legal validity. If legal obligation cannot be explained in
terms of the social facts that the positivist believes explain legal validity, then there
must be more to the explanation of legal validity, given the conceptual relationship
between the two notions, than those social facts. And this, of course, threatens the
view that the law can be explained entirely in terms of social facts, which is the very
core of positivism’s conceptual foundation. A comprehensive account of legal
obligation is absolutely crucial to the continuing viability of legal positivism.
5.3. Legally obligated or legally obliged?

One might be tempted to deny it is a conceptual truth that mandatory first-order
legal norms create something accurately characterized as “legal obligation.” On this
line of analysis, a social norm of any kind cannot obligate someone who has not
taken the internal point of view toward the norm, the system of norms, or the
recognition norms creating that system. It is true that the coercive enforcement of
these norms obliges citizens to conform their behavior to these norms, but that is all
the work that can be done by coercive enforcement mechanisms. Accordingly, the
failure of HTSO to explain the legal obligations of citizens is not a problem for Hart
because it is not a conceptual truth that mandatory first-order legal norms legally
obligate citizens. At the very most, the presence of these elements, together with
coercive enforcement mechanisms in a legal system might “legally oblige” the
subject to obey, but it would not “legally obligate” the subject to obey.10
This reasoning is problematic. To begin, as we have seen, the idea that it is not a
conceptual truth that primary legal norms requiring citizens to behave in certain
ways create legal obligations simply does not line up with ordinary use of the
concept-term “legal obligation,” which figures centrally in our ordinary talk about
core legal practices and in those core practices themselves. In the absence of some
reason to think that such talk is mistaken or incoherent, there is little reason to
reject the associated conceptual presuppositions.
Moreover, the concept of an obliging, unlike the concept of an obligation, picks
out an “ought” that is prudential in character and hence grounded in probabilistic
assessments of self-interest. One is obliged by self-interest to comply with the
gunman’s order because one can calculate that the expected cost of not complying
(i.e. the cost of not complying multiplied by the probability of incurring the cost)
dramatically exceeds the expected benefit of not complying (i.e. the benefit of not
complying multiplied by the probability of achieving the benefit); in consequence,
the expected value (i.e. expected benefit of complying minus the expected cost) is
quite high. But if the expected value of not complying is greater than the expected
value of complying, one is obliged not to comply. Whether or not a person P is
obliged to do a, as a conceptual matter, depends on exactly such a probabilistic
assessment of the effects of doing a and of not doing a on P’s self-interest. This, of
course, is precisely the implication that correctly kills predictive theories that
explain legal obligation in terms of the probability of incurring a sanction.
6. Modifying HTSO to provide a viable comprehensive

theory of legal obligation
HTSO is a plausible account of certain kinds of social obligation. Some obligations—

like moral obligations if morality is objective—might be categorical and apply to
everyone regardless of their attitudes. But it is clear from social practices, as well as
our folk theory of obligation, that people can voluntarily perform acts that place
themselves under an obligation to do something. In the moral realm, promising is
one way of doing so. In the legal realm, the exchange of mutual promises supported
by consideration is another way of doing so.
10 I am indebted to Brian Tamanaha for this concern.

Indeed, it is not unreasonable to think that the only form of obligation that does
not require some deed or attitude by the subject to bind her to a norm defining an
obligation is moral obligation—assuming, again, morality is objective. On this
assumption, a being X is subject to the requirements of morality if and only if
(1) X is sufficiently rational to grasp basic moral norms (“knows the difference
between right and wrong” is one common, but sloppy, way to describe this
requirement); and (2) X has the capacity to freely choose acts that conform to
these norms.
In contrast, Hart seems correct that a person cannot be subject to the mandatory
norms of a social group without being a member of that social group (or, in
anticipation of modifying HTSO, without bearing some kind of theoretically
significant relationship to members of that group). Further, this would seem to
require a voluntary act that is intended, or is known by the actor, to make her a
member of the group or somehow subject to the directives of the social group—and
this act succeeds in doing so. I am not bound by a comprehensive and nuanced
system of etiquette unless I voluntary place myself in a group that is governed by
those norms.
This is, in part, why Hart’s practice theory of rules remains influential in
conceptual legal theory. A group of people must voluntarily join, or subject
themselves to the authority of, the social group made up of people who will serve
as officials of the legal system in order to be bound by any legal norms that regulate
the behavior of officials by creating legal obligations in some cases. The class of
persons functioning as officials in a legal system is just one kind, albeit a special
subtype, of social group as Hart conceives it. Regardless of whether one thinks that
Hart intended to claim that obligations are partly constituted by the authorization
of social pressure for nonconformity, he clearly took the position that social
obligations apply only to people who have taken an attitude that makes them a
member of the group and hence subject to the social norms governing the group.
It is difficult to see how a person could be subject to the mandatory norms of a
legal system without some act that explicitly or implicitly places one under the
authority of that system. Indeed, it is one of the conceptually necessary features of a
legal system, according to Hart, that citizens generally “obey” laws valid under the
social rule of recognition—and obedience involves a certain attitude towards the
authority that expresses itself in voluntary conformity to the law because it is law;
the notion of obedience entails a certain kind of acceptance of the authority—even
if it does not rise to the level of the internal point of view.
For example, there can be no legal system and hence no legal obligations owed by
citizens in a society in which a sufficiently large number of people openly revolt
against the system by fighting authorities and disobeying many laws either as a
matter of protest or chronic widespread lawlessness. Such a situation lacks the
requisite stability of an efficacious institutional system of norms to warrant charac-
terizing the system as a legal system. Such resistance might not rise to the level of a
civil war or a revolution; as long as it reaches a level that precludes a system from
being sufficiently efficacious to count as a legal system, there can be no legal
obligations. Such persons are not under a legal obligation because their behavior is
not such as to place them under the governance of a legal authority.
The key to a successful comprehensive account of legal obligation, then, is to
identify some voluntary act that places anyone under the appropriate authoritative
group or mechanism responsible with manufacturing the law. In the case of
officials, it seems quite plausible to think, with Hart, that, as a conceptual matter,
it is enough for officials to place themselves under the legal obligations defined by a
rule of recognition that they converge on taking the internal point of view toward
the rule of recognition and on conforming their acts as officials to that rule.
But Hart rightly maintains that citizens need not take the internal point of view
towards the rule of recognition, the law, or the authority of the system. For this
reason, it could not be a conceptual truth that citizens voluntarily place themselves
under the authority of the law by taking the internal point of view towards the rule
of recognition, the law, or the authority of the legal system. We must look to some
other attitude or act on the part of citizens if we are to successfully explain how
citizens are legally obligated by mandatory first-order legal norms—and this will
necessarily require a departure from HTSO in letter, if not in spirit. Since HTSO
remains one plausible piece of a comprehensive theory of legal obligation, the
departure will amount to a friendly modification of HTSO that augments it by
adding a second set of conditions for placing oneself under the authority of some
social institution, like law.
Intriguingly, Hart frequently suggests that other attitudes entail acceptance of a
system of putatively authoritative norms might be sufficient to create obligations.
For example, he argues that:
In what sense, then, are we to think of the continuity of the legislative authority of the
Queen in Parliament, preserved throughout the changes of successive legislators, as resting
on some fundamental rule or rules generally accepted? Plainly, general acceptance is here a
complex phenomenon, in a sense divided between official and ordinary citizens, who
contribute to it and so to the existence of a legal system in different ways. The officials of
the system may be said to acknowledge explicitly such fundamental rules conferring
legislative authority: the legislators do this when they make laws in accordance with the
rules which empower them to do so. . . . The ordinary citizen manifests his acceptance
by acquiescence in the results of these official operations (Hart 1994: 61; emphasis added).11
Of particular importance here is that this discussion takes place in the context of
evaluating Austin’s view that a mere habit of obedience figures into the existence of
a legal system and legal obligation. Hart rejects Austin’s view on the ground that
“habits are not ‘normative’” (Hart 1994: 60), which suggests that acceptance in the
11 Similarly, Hart speaks of two different forms of acceptance that have an internal aspect: the
internal point of view taken by officials and the acquiescence of citizens who generally conform (“obey”
is the word Hart frequently uses) to directives of the legislature or courts: “if the truth of this
presupposition [i.e. internal statements of law presuppose some form of acceptance] were doubted, it
could be established by reference to actual practice: to the way in which courts identify what is to count
as law, and to the general acceptance of or acquiescence in these identifications” (Hart 1994: 108;
emphasis added).
form of “acquiescence” is sufficiently normative to bind acquiescing subjects to the

mandatory first-order legal norms of the legal system.
Hart repeatedly emphasizes the role of even passive forms of acceptance like
acquiescence among citizens as an essential feature of a legal system: “The existence
of a legal system . . . involves the attitudes and behaviour involved in the voluntary
acceptance of rules and also the simpler attitudes and behaviour involved in mere
obedience or acquiescence” (Hart 1994: 201; emphasis added). Hart is commonly
thought to hold that the internal point of view that officials take towards the social
rule of recognition is the only form of acceptance conceptually required to explain
the existence of a legal system and the content of conceptually paradigmatic legal
practices. But the discussion here makes clear that some weaker form of acceptance
on the part of citizens is needed as well, and he characterizes that form of acceptance
as “acquiescence.”
Acquiescence to a system of norms, unlike taking the internal point of view
towards the norms, does not involve accepting the norms as legitimate in any sense
and does not involve taking a “critically reflective view” of deviations from the
norms. It merely involves passive acceptance (conceived as something no stronger
than tolerance) by the citizen of the imposition of norms as regulating her behavior,
together with a willingness to conform to the norm because it is a member of
the system. This is a weaker normative attitude to the law than taking the
internal point of view towards it but it is also weaker than the Hartian requirement
for the existence of a legal system that valid rules “must be generally obeyed”
(Hart 1994: 116).
One plausible logical consequence of the analysis in the last paragraph is that
“willingness” is not synonymous with “obedience.” “Obedience” connotes that one
consciously conforms to the norm in deference to the will of the authority that
manufactures the norm (or, if the norm is not a social artifact, deference to the
authoritative quality of the norm) in circumstances where one is disposed to do
something else. My behavior conforms to, or complies with, the law prohibiting
murder but it is not true that I obey the law prohibiting murder because my
motivation for not having committed murder makes no reference to the fact
that murder is illegal; instead, my not having ever committed murder is explained
by moral considerations and certain empathetic emotions.
Obedience is different from mere conformity or compliance. Obedience requires
a deferential attitude towards the norm or authority that treats the norm’s status as
authoritative as a reason to ensure one’s behavior does not violate, or conforms to,
the norm. As Robert Wolff puts the matter, “Obedience is not a matter of doing
what someone tells you to do. It is a matter of doing what he tells you to do because
he tells you to do it” (Wolff 1998: 9). Similarly, Green 2010 observes that “to obey
is not merely to comply with the law; it is to be guided by it.” The relevant form of
deference here is doing what the authority requires because the authority requires it
and this necessarily involves being guided by it.
Mere conformity to or compliance with a norm entails no more than that one’s
behavior is consistent with the requirements of the norm; in particular, to charac-
terize someone’s behavior as conforming to, complying with, the law is to say
nothing at all about her motivations for the relevant behaviors. We know that the
behavior of most people generally conforms to the law in a legal system but, as that
says nothing about what motivates conforming behavior, we cannot say with
confidence that most people generally obey the law. In fact, I would be surprised
if the vast majority of people who refrain from violent felonies do so out of a
deferential attitude towards the law.
For this reason, Hart overstates one of his minimum conditions for the existence
of a legal system when he asserts that one necessary condition for the existence of a
legal system is that “those rules of behaviour which are valid according to the
system’s ultimate criteria of validity must be generally obeyed [by citizens]” (Hart
1994: 116). Insofar as obedience, as a conceptual matter, requires a specific form of
deference to the authority that takes the form of doing what the authority directs
because she directs it, it presupposes that the obeying subject take something very
close to the internal point of view towards the authority or authoritative directives.
This sort of deference makes no sense if one does not believe that others should
obey the norms of the authority and involves general conformity to those norms.
These are, or closely resemble, the constituents of the internal point of view as Hart
defines them.
Acquiescence to a set of putatively authoritative norms involves not defying
the authority when conformity to one of the norms is inconsistent with one’s
desires. It is constituted by general conformity to the norms without regard to the
motivation and by an attitude of willingness to do what the authority says in most
cases in which the subject is not inclined to conform. This willingness is less
normatively robust as the form of deference to the authority is endorsed by the
internal point of view but it is nonetheless a form of deference that is plausibly
characterized as a form of acceptance. Accordingly, HTSO must be augmented as
follows to include this form of acceptance:
HTSO +: X has a social obligation to do p if and only if
(1a) X is a member of the relevant group in virtue of taking the internal point of view
towards a mandatory norm N governing X that requires X to do p; or
(1b) the members of the group subject to the system of mandatory norms including N,
generally acquiesce to the application of this system of norms to X’s behavior as well as
their own;12 and either
(2a) the members of the group generally conform to N; or
(2b) N is a norm imposed by the group on non-members who generally conform to the
norm and are willing to conform to norms conflicting with their desires, at the very
least, to avoid conflict with members of the group; and
(3) N is supported by significant social pressure because
(4) N is thought important because necessary to the maintenance of social life or some
highly prized feature of it.
12 Thus, universal disregard of just one particular law is not enough to defeat condition (1b) and
hence does not defeat the claim that X has a social obligation to do p. I am indebted to Alex Langlinais
for the formulation of (1b).
It is important to note here that every condition here expresses a social requirement
defined by social acts and attitudes.
7. Comprehensiveness achieved: a positivist account

of first-order legal obligation
As Hart’s theory of the social rule of recognition seems a plausible positivist account
of the second-order legal obligations that bind officials in their capacity as officials,
Hart’s account serves as one piece of a comprehensive positivist theory of legal
obligation. It bears reiterating that the methodology here is to find salient elements
of legal practice that function in a way that coheres with the core commitments of
positivism; with the general account of obligation; and with HTSO+. Hart’s theory
of the social rule of recognition does all these things and successfully meets the
standard for a plausible explanation for second-order legal obligation.
It remains to develop an account of first-order legal obligation that conforms to
these core positivist commitments, the general theory of obligation, and HTSO+.
As we will see, coercive enforcement together with acquiescence on the part of
citizens will satisfy all three requirements for a viable account of first-order legal
obligations and hence fills in the gap in Hart’s theory of legal obligation.
Once law is explained in terms of a social rule of recognition accepted by officials
in an efficacious legal system, citizen obligation in modern municipal legal systems
seems best explained in terms of the authorization of formal institutional mechan-
isms of coercive enforcement. The idea here is not that coercive enforcement of a
norm, by itself, constitutes the norm as obligatory; rather, it is that coercive
enforcement of a social norm in a system that satisfies certain properties—including
the institutionalization of the relevant set of norms—constitutes it as legally
obligatory. Coercive enforcement of a legal norm constitutes it as legally obligatory
upon citizens, in part, because (1) the norm belongs to an institutionalized system
of norms (2) grounded in recognition norms accepted and practiced by officials and
is (3) minimally efficacious in regulating citizen behavior.
Here it is important to emphasize the normative dimension of this practice.
While officials of the legal system need not regard a first-order law as a moral
justification for enforcing the law against non-compliance, they regard it as a legal
reason or justification (i.e. a reason that is internal in the sense that it is within the
system of law) for such enforcement. Obligation is explained by a normative web of
practices that includes the legal authorization of formal enforcement mechanisms as
a legal justification for applying them to citizens for non-compliance.
Formal institutional enforcement should be distinguished from sanctions.
Enforcement sometimes involves punitive intent, as it does in the case of a
defendant who is being prosecuted for murder under the criminal law. But it
need not involve such intent, as in the case of a judge ordering damages for breach
of contract. Such enforcement mechanisms include sanctions but also include the
court’s power of contempt, which backs every court order. Moreover, the court’s
authority over these mechanisms includes the authority to refuse to enforce or
recognize a defective instrument of some kind, which might include a contract,
will, or even a statute. Refusal to enforce a defective contract is part of how courts
coercively enforce the laws governing formation of a contract.
What constitutes a mandatory norm as legally obligatory in modern municipal
legal systems is that coercive enforcement is legally authorized. If the application of
coercive force for violations of a valid legal norm N is authorized by some valid legal
norm as a normative response to non-feasance, then N is legally obligatory and its
binding force is constituted by the authorization of the relevant coercive mechan-
isms. Of course, it is probably true that it is also a necessary condition for the
existence of a legal obligation that the relevant coercive mechanisms are reliably
applied in cases where they are authorized. But this is not part of what constitutes a
norm as legally obligatory.
8. Supporting considerations
8.1. The centrality of coercive enforcement in modern judicial practice
The availability of formal, institutional coercive enforcement mechanisms is a
central feature of law in modern municipal legal systems. Most obviously, the
criminal law is characteristically backed with punishment. But such mechanisms
also play a central role in civil law: the point of bringing a civil lawsuit is to get a
court order requiring the defendant to do something. Sometimes the plaintiff seeks
damages; sometimes the plaintiff seeks specific performance. However, any plaintiff
who brings a civil suit in any legal system remotely resembling this one is asking the
court not only for a judgment, but also a court order.
The court has authority to enforce its lawful orders by a formal, institutional
coercive mechanism known as the contempt sanction. It is this power that enables
the judge to enforce her orders in civil cases where they cannot plausibly be
characterized as imposing direct or indirect sanctions. In systems like ours, every
court order is backed by the legal authorization of the contempt sanction for non-
compliance.
This suggests that coercion is central to legal systems resembling that of the
US. Since the contempt sanction is both coercive and universally available to courts
to enforce its orders in civil and criminal matters, it follows that every criminal and
civil law is ultimately backed with a coercive mechanism (since the court’s con-
tempt sanction is coercive). The authority of the court to issue coercively enforced
orders is foundational to its ability to decide disputes in systems like this one.
At the very least, this much seems reasonable: in cases where (1) formal coercive
mechanisms are generally authorized for non-compliance and (2) officials lack
authority to apply these mechanism in enforcing a particular judgment, norm, or
order with coercive mechanisms, it is implausible to characterize the judgment,
norm, or order as “obligatory.” Such norms are more fairly characterized as
“advisory” because there is no sense in which the relevant behavior is made
mandatory by mechanisms reasonably presumed to have normative relevance given

certain general psychological facts about human beings.
This is not to suggest that legal obligation cannot exist in a legal system without
formal, institutional coercive mechanisms, which would entail that such mechan-
isms are a conceptually necessary feature of law—that is to say, that law is
necessarily coercive. Although both HTSO and HTSO+, as I have construed
them, entail that social pressure is a necessary condition for social obligation, the
account of legal obligation here assumes only that some form of social pressure is a
necessary condition for legal obligation. The specific view that the authorization of
formal, institutional coercive mechanism constitutes the binding force of first-order
legal obligation applies only to modern municipal legal systems like that of the
US. The theory defended here neither assumes nor implies that coercive enforce-
ment is a necessary feature of law—a virtue because the idea that law is necessarily
coercive is contentious.
In any event, the authorization of such measures is a more reliable indicator of a
legal obligation than the language in which the relevant law is expressed.
A statement asserting that the defendant “must” or “shall” perform some act is,
despite its language, best characterized as “advisory” if no coercive legal conse-
quences are authorized for failure to comply.13 Further, a statute asserting
that people “should” perform some act is, despite its language, best characterized
as “obligatory” if courts are authorized to incarcerate people who do not perform
the act.
As a general matter, officials are quite careful to ensure that the words of an
authoritative statement of law adequately signal whether coercive enforcement
mechanisms are available, but this is explained by non-conceptual considerations.
Conscientious officials want to ensure that authoritative statements of law convey
appropriate notice of what is required. The terms “must” and “shall,” in contrast to
“should” and “ought,” signal that some behavior is required and provide construct-
ive notice to citizens that courts have recourse to some coercive mechanisms—
though such terms do not say anything about the nature or severity of such
mechanisms.
Still, it is the availability or non-availability of coercive mechanisms, and not the
language in which a rule of law is expressed, that ultimately determines whether
that rule defines a legal obligation. When the language in which a legal norm N is
expressed and the availability of coercive enforcement mechanisms do not agree, it
is the latter that determines whether N is fairly characterized as “legally obligatory”
upon citizens.
It is crucial to understand here that there is a difference between the authoriza-
tion of coercive enforcement mechanisms and the application of such mechanisms
in a case of non-compliance. These are two distinct notions. The idea that such
13 There are, of course, some laws that authorize sanctions but are chronically unenforced. Whether
or not these count as legally obligatory will be determined by whether they count as valid under that
system’s recognition practices. In some legal systems, the chronic failure to enforce or apply a norm
suffices to invalidate the law, a situation sometimes described as “repeal by desuetude.”
mechanisms are authorized for non-compliance simply means that officials have
authority to use these mechanisms as legally justified responses to non-compliance.
The idea that such mechanisms are applied simply means that those coercive
mechanisms have been used against someone on the ground that he failed to
comply. But it is important to note that this does not entail even that the use of
such mechanisms are legally justified—as one would expect if legal mistakes are
possible. The authorization of coercive enforcement of a legal norm provides a legal
justification for the appropriate application of the relevant mechanism for non-
compliance with the norm.
One might object that the violation of a legal obligation justifies the application
of coercive mechanisms and thus that a legal obligation cannot be constituted by
coercive enforcement applications.14 This misunderstands the thesis of the chapter.
The claim being defended here is that the authorization of such mechanisms for
non-compliance is, in part, what constitutes a legal norm as binding and hence
legally obligatory and hence provides the justification for application in genuine
cases of non-compliance. The obligation is constituted, in part, by the authorization
of such mechanisms and is not identical with the existence or application of such
mechanisms.
8.2. The binding force of obligation

This theory explains the bindingness of mandatory legal norms in terms of consid-
erations likely to be regarded by subjects as normatively relevant. First, being
subject to coercive enforcement is a clear sense in which that norm can plausibly
be characterized as being non-optional. Second, the authorization of coercive
enforcement mechanisms (including the contempt power) is something that is
normatively relevant to any rational citizen. This, again, is not to claim that citizens
are necessarily motivated to obey the law by a fear of sanctions; rather, the point is
merely that rational self-interested citizens are, as descriptive matter, likely to care
about avoiding the coercive enforcement power of the state.
One might worry, however, that the sort of reason for action that a legal
obligation provides, according to this theory, is, as a conceptual matter, the
wrong kind of reason. In particular, one might object that this account explains
the normative legal obligation in terms of prudential considerations and hence
reduces legal reason to first-order prudential reasons. This is problematic insofar as
one thinks (1) prudential reasons are not the only basic reasons constituting a legal
reason and (2) legal reasons are pre-emptive reasons.
As to (1), it seems clear that legal reasons, on a positivist view, being the product
of a human artifact manufactured by social processes (i.e. a legal system) would
have to be a compound reason reducible to basic reasons. And it is clear that a
positivist cannot hold that it is a conceptual truth that a legal reason is partly
reducible to a moral reason without violating the Separability Thesis that there are
14 I am indebted to Scott Shapiro for this line of argument.

no necessary moral constraints on the content of law. As, we saw above, there is a
limited palate of basic reasons to choose from: there seem to be no other kinds of
basic reason other than prudential, moral, and possibly aesthetic reasons. And it
should be clear that legal reasons are not constituted, even in part, by basic aesthetic
reasons if such there be. If legal reasons are compound, the only kind of reason they
could be reduced to are prudential reasons.
But it is important to understand that the claim that first-order legal obligation
creates a reason for action is prudential in character is not inconsistent with the idea
that such obligations are exclusionary.15 It is the authorization of a coercive
response to non-compliance that constitutes the binding force of legal obligation.
While this is a prudential consideration, it is nonetheless constitutive of the binding
force of the obligation and functions differently than the kind of prudential
considerations excluded as justifications or excuses for non-performance. The latter
considerations normally figure, from the standpoint of purely prudential rational-
ity, in estimating expected values of doing or not doing something. The binding
force of a first-order legal obligation might be constituted by prudentially salient
considerations but the exclusionary nature of the obligation, as defined by the terms
and application conditions of the relevant legal norm, conclusively defeats the
balance of these other prudential reasons regardless of their comparative weight
by making the weighing process irrelevant by one of two mechanisms: (1) applying
the norm regardless of whether the prudential reasons constituting the binding
force of the obligation outweigh the balance of other prudential considerations; or
(2) excluding these other prudential considerations as excuses or justifications for
non-compliance.
As to (2), the idea that legal reasons are preemptive reasons is contentious. While
it is clear that mandatory legal norms are, by the very nature, exclusionary in the
sense that they exclude certain justifications for non-performance, this does not, by
itself, imply that the reasons created by such norms are preemptive in the Razian
sense. Given the fact that the Razian account of authoritative reasons is conten-
tious, the objection simply begs the question against the account offered here.
Given the consistency of this account of first-order legal obligation with
the exclusionary character of first-order law, the idea that the authorization of
coercive enforcement constitutes a mandatory norm as legally obligatory harmon-
izes nicely with another important idea concerning legal obligation—namely that
there is no prima facie moral reason to obey the law. Most theorists have come to
reject not only the idea that the law necessarily gives rise to moral obligations but
even the weaker idea that it is necessarily the case that we have even a prima facie
moral reason to obey legal requirements. If this plausible view is correct, then the
fact that a mandatory legal norm creates a legal obligation does not imply that it
creates a moral obligation to obey it—or even that there is a prima facie moral
reason to obey it.
15 I am indebted to Stefan Sciaraffa for impressing on me the need to address this concern.
This harmonizes nicely with the theory of first-order legal obligation defended
here. The only reasons for action that are necessarily provided by a legally obliga-
tory norm, if the theory here is correct, are prudential in character—even if law’s
exclusionary character rules out the possibility of these prudential reasons being
defeated by the balance of all other prudential characteristics. Clearly, first-order
legal obligation would be prudentially normative on the story offered here: it is not
in the interests of a person, other things being equal, to be subject to the sorts of
coercive mechanisms that are used to enforce mandatory legal norms. Equally
clearly, first-order legal obligation is not necessarily morally normative on this story.
This is exactly what we would expect if the prevailing view that law does not
necessarily give rise to prima facie moral reasons to obey the law is correct. An
analysis of legal obligation that implies we have even a prima facie moral reason to
satisfy our legal obligations would be inconsistent with this view. The fact that, on
the analysis offered here, legal obligation is not necessarily morally normative is a
strong point in its favor.
9. Conclusions
In this chapter, I have attempted to ground a comprehensive theory of legal

obligation that explains the nature of both first- and second-order legal obligations
in terms that are consistent with a minor modification of Hart’s theory of social
obligation. In particular, I have argued that first-order legal obligations are consti-
tuted by the authorization of coercive enforcement mechanisms for non-compli-
ance, together with acquiescence to the norms of the system on the part of the
subject, in an institutional system of norms that includes a social rule of recognition
that is supported by social pressure because the rule is thought to advance some
important feature of social life.
References
Coleman, J. (2001). “Conventionality and Normativity,” in E. Sosa, E. Villanueva, and

J. Tomberlin (eds.), Social, Political, and Legal Philosophy. Oxford: Blackwell Publishers.
Frankena, W. (1966). “The Concept of Morality,” Journal of Philosophy, 63: 688–96.
Green, L. (2010). “Legal Obligation and Authority,” in E. Zalta (ed.), Stanford Encyclopedia
of Philosophy; available at <http://plato.stanford.edu/entries/legal-obligation/>.
Hart, H. L. A. (1994). The Concept of Law. Revised edition. Oxford: Oxford University Press.
Lyons, D. (1993). Moral Aspects of Legal Theory: Essays on Law, Justice and Political
Responsibility. Cambridge: Cambridge University Press.
Marmor, A. (2011). “The Nature of Law,” in E. Zalta (ed.), Stanford Encyclopedia of Philosophy;
available at <http://plato.stanford.edu/archives/win2011/entries/lawphil-nature/>.
Raz, J. (1990). Practical Reasons and Norms. Oxford: Oxford University Press.
Raz, J. (1979). The Authority of Law. Oxford: Clarendon Press.
Wolff, Robert (1998). In Defense of Anarchism. Berkeley, CA: University of California Press.
8
Law and the Entitlement to Coerce*
Robert C. Hughes
1. Introduction
A long tradition in political and legal philosophy regards coercion as central to the
very idea of law. Some historical figures, such as Hobbes, Locke, and Austin, took
the position that there can be no law without a coercive sanction. Many philoso-
phers of law, most famously H. L. A. Hart, have called this view into question.1
Nonetheless, many political and legal philosophers continue to believe that law
is necessarily connected with coercion in a subtler way. Whenever government is
entitled to make a law that imposes a direct requirement on conduct, it is entitled to
use coercion to enforce this requirement. Some endorse this position explicitly.2
Others commit themselves to it when they argue against certain kinds of laws or
legal arrangements by claiming that coercive enforcement of those laws would be
objectionable. Such arguments presuppose that a government is justified in making
law only if it has some sort of justification for enforcing law coercively. The view
that the entitlement to make law necessarily comes with an entitlement to coerce is
challenged rarely, if ever. Nonetheless, this view is mistaken.
I shall argue that the justification of the power to make law does not entail the
existence of an entitlement to coerce. It is possible for there to be a legal system in
* I am grateful to Jorah Dannenberg, Joshua Dienstag, Sonny Elizondo, Barbara Herman, Louis-
Philippe Hodgson, Yannig Luthra, Eliot Michaelson, Letitia Meynell, Calvin Normore, Arthur
Ripstein, Stefan Sciaraffa, Seana Shiffrin, Luca Struble, Wil Waluchow, Alan Wertheimer, and
audiences at Northwestern and UCLA for their comments on this chapter and on earlier papers
from which this chapter is derived.
1 Hart notes, for example, that there is no coercive sanction directly attached to a law requiring
someone who wants to make a will to get two witnesses’ signatures (1994: 28). Lawrence Sager
provides another counterexample. The United States Supreme Court is bound by law in deciding cases,
but its members are not subject to sanction if their opinions fail to follow the law. Congress can remove
Supreme Court Justices from office by impeachment and trial, but presumably it is not in virtue of this
rarely applied sanction that Justices are bound by law in making decisions (1978: 1222). Joseph Raz
and Scott Shapiro have argued that is conceptually possible for there to be an entire legal system that
lacks coercive enforcement. Raz (1999: 157–61); Shapiro (2011: 169–70).
2 Arthur Ripstein does so in his defense of Kant’s political theory (2009; 2004). Kant expresses this
view in the Metaphysics of Morals at Ak. 6:232 (1996: 388). Grant Lamond defends the related position
that law necessarily claims the right to exercise coercion. He criticizes the view that law is necessarily
coercive (2001).
184 Robert C. Hughes
which making law is justified but neither government nor any private party has any
entitlement to enforce law coercively. In societies like ours, governments may
sometimes have an entitlement to coerce, but this entitlement does not follow
from the bare fact that some citizens are inclined to violate legal restrictions their
governments are entitled to impose. Whether a government is entitled to enforce
any given legal restriction coercively depends on concrete empirical facts, not only
on abstract truisms about human nature, and on moral features of the restriction
in question. Empirically, it matters to what extent compliance with a particular
restriction could be secured without coercion and what effects coercive enforce-
ment would have. Morally, it matters whether the value of coercively addressing
violations of a given law is great enough to justify establishing a coercive enforce-
ment mechanism that inevitably risks injuring innocent people. Because the
answers to these questions can vary, a government’s entitlement to coerce may
not extend to all justified legal requirements on conduct, and it may expand or
contract as social conditions change.3
The fact that an entitlement to coerce does not simply follow from an entitlement
to make law has consequences for the way political discourse should be conducted.
Both civil and economic libertarians often criticize legislation by arguing that the
legislation in question cannot permissibly be coercively enforced.4 Such arguments
must be either abandoned or refined. One cannot straightforwardly argue that
a legal requirement should not be imposed by arguing that it would be wrong to
enforce this requirement coercively. Recognizing that government can be entitled
to make laws that it is not entitled to enforce coercively can also help us to make
more sophisticated diagnoses of injustice. For example, arguments for economic
egalitarianism often appeal to the injustice of using coercion to enforce an inegali-
tarian system of property.5 If there is a moral problem with an inegalitarian system
of property, it is important to ask whether the problem is only with the coercive
enforcement of the system or also with the government’s claim that people ought
to comply.
To show that the entitlement to make law does not always come with an
entitlement to coerce, I will first argue in section 2 that it is conceptually possible
for there to be a legal system that neither uses nor threatens coercion. If it were
impossible to make law without using or threatening coercion, an entitlement to
make law would, of course, require an entitlement to coerce. Sections 3 and 4
address two ways of arguing that the entitlement to make law generally comes with
an entitlement to coerce even if non-coercive law is possible. Section 3 considers
and rejects arguments that the entitlement to make law necessarily comes with an
entitlement to enforce law coercively. On this view, in any possible society, even a
society of morally very good people who do not break the law without justification,
3 Requirements on conduct contrast with requirements that must be met to acquire, to transfer, or
to exercise a legal power, such as the requirement that one have two signatures to make a valid will.
4 Robert Nozick’s libertarian arguments take this form (1974: ix).
5 Ronald Dworkin and Michael Blake both make arguments of this type. Dworkin (2002: 2); Blake
(2001).
Law and the Entitlement to Coerce 185
government has an entitlement to coercively enforce every law it is entitled to make.

Section 4 addresses arguments that the entitlement to make law always comes
with an entitlement to coerce in a society in which some people are inclined to
break the law unjustifiably.
2. The conceptual possibility of law without coercion

Before arguing that no entitlement to coerce is required to have the power to make
law or to be justified in exercising that power, I will explain why it is possible to
make laws—not mere public guidelines or recommendations, but laws—that lack
coercive enforcement. Indeed, it is at least conceptually possible for there to be an
entire legal system that lacks coercive enforcement. Showing that there can be law
without coercion is a necessary first step to showing that there can be an entitlement
to make law without an associated entitlement to coerce. If making law necessarily
involved using or threatening coercion, it would be senseless to try to argue that
lawmaking bodies are entitled to make laws they are not entitled to enforce
coercively.
I am not the first to defend the conceptual possibility of an entirely non-coercive
legal system. Raz and Shapiro have both argued for this possibility.6 My argument
for the conceptual possibility of an entirely non-coercive legal system will differ
from Raz’s and Shapiro’s arguments in that it does not presuppose a particular
account of the nature of law. In particular, it will not presuppose a position on legal
positivism, the view that non-moral social facts are the only facts that determine the
content of the law.7 Raz’s and Shapiro’s arguments for the possibility of non-
coercive legal systems both presuppose particular, fully-developed accounts of the
nature of law, which in both cases are positivist accounts. Perhaps because these
arguments presuppose positivist accounts of law, the possibility of a non-coercive
legal system has not been widely recognized among anti-positivists. Notably,
according to Ronald Dworkin’s account of the concept of law, “The most abstract
and fundamental point of legal practice is to guide and constrain the use of
governmental power,” where “power” means coercive power.8 On this view, law
would be either impossible or pointless in the absence of a coercive government
whose power needs to be constrained.
I will show that both positivists and anti-positivists can accept the conceptual
possibility of an entirely non-coercive legal system. To do this, I will offer a set
of jointly sufficient conditions for the existence of a legal system. Since these
6 Raz (1999: 157–61); Shapiro (169–70).

7 Greenberg (2004: 157–8).
8 Dworkin (1986: 93). Recently, Dworkin has taken a subtler view. He distinguishes the “socio-
logical concept of law” from the “doctrinal concept of law.” The sociological concept of law, which is
the concept relevant to the question whether a social practice counts as a legal system, “is not
sufficiently precise to yield philosophically interesting ‘essential features’ ” (2006: 228). So there is
no point in asking whether there could be a legal system, properly so-called, that lacks enforcement
institutions (2006: 3).
conditions are jointly sufficient, not necessary or constitutive, both positivists and
anti-positivists should be able to accept them. I will then show that in a society of
morally very good people, there could be a system of legislative and judicial bodies
that satisfies these conditions and that does not use or threaten coercion.9
2.1. Sufficient conditions for the existence of a legal system

One might be tempted to think that any characterization of law must mention law’s
coercive character. Intuitively, laws seem to be binding in a way that other rules are
not. The presence of public coercive enforcement is the obvious way to explain this
difference between laws and other rules. So one might think that purported legal
restrictions on conduct that lack coercive enforcement are not really laws; they are
merely public guidelines, recommendations, or advisories. A system of cooperation
that lacked coercive enforcement for its rules would not constitute a legal system
because its rules would not be binding in the way that is characteristic of law.
I suggest there is another way of seeing the distinction between laws and other, non-
mandatory public rules. Even if laws are not binding in the sense that they are
coercively enforced, laws can be morally binding in a way that guidelines, recom-
mendations, or advisories are not.
When complying with guidelines is morally mandatory, it is mandatory for
reasons that are independent of the guidelines’ existence. For instance, if complying
with safety guidelines is morally mandatory, compliance is mandatory because of
independent facts about what safety requires, not because a public body decided to
establish the guidelines. By contrast, laws do not in general aim merely to report
independently existing moral obligations. To be sure, some laws do not change
people’s moral obligations and do not aim to do so. For instance, laws criminalizing
wrongful violence aim to codify independent moral requirements.10 Most laws,
however, aim to change what obligations people have, either by directly imposing
obligations or by changing the ways in which obligations can be established,
transferred, and eliminated.11 Laws regarding the ways in which property may be
acquired and transferred, for instance, aim to affect what actions constitute morally
wrongful theft.
Of course, not all rule-making or rule-applying bodies that can change people’s
moral obligations are law-making or law-applying bodies. For instance, private
voluntary associations may be able to make rules or decisions that change their
members’ moral obligations. There are two ways in which legal systems’ normative
powers are distinctive. First, legal systems typically claim, and often in fact have, a
9 My argument resembles Raz’s in that it uses a thought experiment involving a society of morally
very good people. Unlike Raz, I do not assume that such people would be disposed to comply with law;
this is to be shown.
10 The only people whose obligations these laws affect are the people who operate the criminal legal
system. I do not claim that a society of morally very good people would have any use for laws
criminalizing wrongful violence.
11 The distinguishing feature of philosophical anarchism is its rejection of the view that citizens
have a moral duty or obligation to treat law as binding. Simmons (1987); (2009).
moral power to regulate everyone in a geographically defined society. They do not

limit their normative claims to a proper subset of the population, such as a list
of volunteers. Second, legal systems characteristically regulate certain areas of life.
Though there is probably no area of life that some legal system has not tried
to regulate, there are certain areas of life all effective legal systems do regulate.
For instance, every effective legal system has something to say about the allocation
of physical resources within its territory.
I suggest the following as a set of jointly sufficient conditions for the existence of
a legal system:
(1) There is a system of legislative and judicial bodies that makes rules and
decisions concerning a subject matter that is characteristically governed
by law—that is, a subject matter that is regulated by the legislative and
judicial branches of government in every actual society that has an effective
government.
(2) These rules and decisions are morally binding on everyone in a geographic
territory.
(3) The rules and decisions are morally binding partly because these legislative
and judicial bodies have made them. The bodies in question do not merely
identify independent moral truths and report them; their pronouncements
actually affect what moral obligations people have.
I emphasize that these conditions are jointly sufficient conditions for the existence of
a legal system, not necessary conditions. I leave open whether there could be legal
systems that do not meet one or more of these conditions (e.g. whether there could
be a legal system with a legislature and no judiciary). I also emphasize that these
conditions are conditions for the existence of a legal system. They do not claim that
an individual rule must be morally binding to count as a law. Legal systems
typically have some laws that do not alter people’s moral obligations, in some
cases because these laws do not aim to change people’s moral obligations, in other
cases because they try and fail. Such laws may count as laws not because they are
morally binding or because they are binding in some other way, but because they
are products of the same legislative process that in other cases produces morally
binding rules with characteristically legal content and applicability.
It is clear that anti-positivists can accept these conditions for the existence of a
legal system. Despite the conditions’ reference to morality, positivists should also be
able to accept them, since these conditions are supposed to be sufficient conditions
for the existence of a legal system, not necessary or constitutive conditions.
Positivists can hold that any system of legislative and judicial bodies that satisfies
(1)–(3) would have all of the non-moral, social features that it must have to count as
a legal system. Most notably, the system should satisfy positivist requirements
concerning the efficacy of law. For a system of legislative and judicial bodies to
alter people’s moral obligations, the system and its rules and decisions (or some core
subset of its rules and decisions) must be widely treated as authoritative. If no part
of a would-be legislature’s output were widely respected, the would-be legislature
could not alter people’s moral obligations; it could only describe them. So positiv-
ists and anti-positivists alike can accept (1)–(3) as jointly sufficient conditions for
the existence of a legal system. They will disagree about what facts make a legal
system exist and about what features of a legal system distinguish it from other, less
binding systems of rules. The anti-positivist will hold that the distinctive way in
which legal systems can alter people’s moral obligations distinguishes them from
other normative systems. The positivist will hold that legal systems are distin-
guished by the way in which their rules are socially regarded as binding.
2.2. Meeting the conditions

Having outlined a set of jointly sufficient conditions for the existence of a legal
system, I shall argue that in a society of morally very good people, an entirely non-
coercive system can satisfy these conditions. An entirely non-coercive legal system is
thus a conceptual possibility. Though there are various forms of law that a non-
coercive system could make, I shall concentrate on property law. A society’s system
of property provides rules for the allocation of its physical resources.12 To have a
property interest in an object is to have an entitlement to use it in some way or set
of ways, free from the interference of others. Property is a form of cooperation in
which governments are characteristically involved. All actual effective governments
make laws regulating the acquisition and transfer of property (e.g. about the
transfer of property by inheritance) and specifying what rights come with property
interests of different sorts (e.g. what rights landlords and tenants have). All actual
effective governments also settle disputes about property, such as boundary disputes
between neighbors. Moreover, governments are justified in making property law,
since it serves morally important purposes. Whether every society needs a system in
which private property is the dominant form of property ownership is debatable,
but it is clear that every society needs some system for the allocation of physical
resources.13 Having widely respected property rules is important for the function-
ing of a sophisticated economy. Having a system of property is also important
for individual autonomy. If I have no property interests, including property
interests in common or collective property, others are entitled to interfere with
anything I may choose to do with any physical resource. All human activities
require the use of physical resources. (Some activities, like thinking and dancing,
do not directly involve the use of movable objects, but even these activities
involve the use of space.) Individuals need property in order to have spheres
12 Waldron characterizes property in a similar way (1988: 31–7). He notes that some systems of
property allow people to have property interests in non-material objects, such as inventions and
reputations. He argues that interests in material objects are more central to the concept of property
because all human societies have felt the need to allocate scarce material resources. By contrast, not all
societies have felt a need for a system of intellectual property or for property in other incorporeal objects
(1988: 33–4).
13 An example of a property interest that is not a private property interest is an entitlement to use
the public roads.

of possible action in which they can make choices without needing the permission
of others.14
Under ideal social circumstances, it is possible for legislative bodies to make
morally binding rules of property without using or threatening coercion and
without authorizing private parties to coerce. To see this, imagine a society made
up of morally very good people, people who are strongly committed to treating each
other justly, competent at moral reasoning, and free of weakness of will. Suppose
that this society has a deliberative body that has published a code of property rules
and that periodically revises this code in response to objections and concerns.
A large majority of people follow the code because they recognize the importance
of having generally accepted property rules and they think the rules set out in this
code are good. For the rule-making body to have the moral authority that legal
systems typically claim to have, its code must be morally binding on everyone,
whether or not they happen to approve of the code.15 The mere fact that a large
majority of the population has embraced the code gives everyone a strong reason to
follow it. It is important for there to be a single system of property rules that
everyone follows, at least with respect to uses of important resources that exclude or
interfere with others’ use of those resources. Once a system of property rules has
obtained widespread acceptance, refusing to follow it is justified only if there is a
compelling reason to reject it in favor of some other system.
People could have two sorts of reasons for preferring another system of rules to
the one that is widely accepted. First, people could have purely self-interested
reasons for preferring another code. A commitment to treating others fairly does
not exclude reasonable self-interest. But the mere fact that the generally accepted
system of property is less to one’s advantage than some other possible system does
not justify refusing to follow the accepted system. The alternative system one
prefers would disadvantage others if it were adopted. If people were entitled to
disobey existing rules merely because some other system would be better for them,
there could be no system of property rules that everyone is morally required to
obey. That the current system is not to one’s advantage could justify disobedience
only if the disadvantage one experiences is unfair.
Moral reasons, such as considerations of fairness, are the second sort of reasons
people could have for preferring another code to the currently accepted one. If
either the content of the code or the process by which it is created and amended is
grossly unjust, people may be justified in refusing to follow it even if there is no
significant chance that doing so may help bring about consensus behind a better
code. Suppose that the code’s failings are not this great. It is morally better for
everyone to follow the existing code than for some to follow it and for others to
14 This view of the importance of property is inspired by Ripstein’s view. Ripstein claims that
interfering with someone’s use of her property interferes with her ability to set and pursue her own ends
(2006: 240–3). I agree with this, but I also take the stronger position that individuals need to have
property in order to be able to set and pursue their own ends.
15 The code may be only pro tanto morally binding; special circumstances may justify disobedience.
follow another code or no code at all. Suppose further that most people accept the
established legislative process as the only legitimate way of changing the rules of
property. Disobedience will thus have no significant chance of helping to bring
about consensus behind better rules. Under these circumstances, disobeying the
generally accepted rules of property on moral grounds will be unjustified. So the
legislature in this society makes rules of property that are morally binding on
everyone, including people who disagree with them.
A society of morally very good people could also have a judiciary that makes
morally binding decisions about property disputes. Morally very good people can
have good faith disputes about who is entitled to use what resources. They can
disagree about relevant matters of fact, and they can disagree about the interpret-
ation of the generally accepted rules of property. Since they are competent moral
reasoners, their views will be plausible even if they are false.16 If two people who
are competent moral reasoners have a good faith dispute about whether one of
them may use a particular resource in a way that excludes the other’s use of it,
it is unethical for the first party simply to start using the resource or for the
second party simply to prevent this use. Morally, the parties must reach agreement,
either by discussing the issue together or by submitting the issue to the decision of a
third party. Suppose they cannot agree either on the substance of the issue or on a
private party they both trust to settle the dispute fairly. If the generally accepted
code of property identifies a court system with the power authoritatively to resolve
disputes about the code’s application, the parties will be morally required to submit
their dispute to this court system. Unless the courts issue a grossly unjust ruling—
which they will not, assuming that the rules of property and the rules of civil
procedure are not grossly unjust and that the courts perform their duties in good
faith—both parties will be morally required to comply with the court’s decision
even if they disagree.
The society just described has a judicial system that makes morally binding
decisions in disputes about property. It also has a legislature that makes morally
binding property rules. These rules and decisions concern a subject matter that is
regulated by the legislative and judicial branches of government in every society
that has an effective government. So condition (1) is satisfied. The rules and
decisions are morally binding on everyone in the society, so condition (2) is
satisfied. The rules and decisions are binding in part because the legislature and
the courts made them; the legislature and the courts have the power to change
what moral obligations people have. So condition (3) is satisfied. This society
therefore has a legal system. Yet neither the legislature nor the courts ever uses
or threatens coercion, nor do they authorize any other party to coerce on their
behalf. Thus, it is conceptually possible to have a legal system that lacks coercive
enforcement of law.
16 I assume here that competence at moral reasoning includes not only an ability to draw plausible
moral conclusions from non-moral facts, but also some degree of competence at assessing the non-
moral facts relevant to moral questions.
3. The entitlement to coerce in ideal societies
Having shown that there can be law without coercion, I now turn to my central
claim, that a body can be justified in making law without having an entitlement to
enforce law coercively. This section will argue that there is no strictly necessary link
between the entitlement to make law and an entitlement to coerce. It is possible for
there to be a society in which a legal system is entitled to make law, but no one has
any entitlement to enforce law coercively. I will spend some time defending this
claim because there are plausible reasons for thinking that there is an entitlement
to enforce law coercively in any society, including societies of very good people.
In particular, one might think that a law-making body is entitled to establish a
coercive enforcement mechanism even if this mechanism would not address a need.
One might also think that every society needs coercive enforcement to deal with
lawbreaking that stems from good-faith disagreement. Finally, one might think
that people need a coercive assurance that others will respect their rights whether or
not violations are likely. Since these arguments are supposed to demonstrate the
existence of an entitlement to coerce even in the best social circumstances, examin-
ing their application to a society of very good people will show why they are
unsound.
3.1. An entitlement to coerce needs justification

One might think that a legal system has an entitlement to authorize coercive
enforcement of law even if coercive enforcement is not practically necessary.
Sometimes a political body is entitled to do something it has all-things-considered
reason not to do. In a society of people who are strongly disinclined to break the
law, there may be policy reasons not to establish a coercive enforcement mechan-
ism. Building prisons and hiring police officers may be a poor use of public
resources, and it may not be a good use of the legislature’s time to set penalties
for violations of law. There could still be an important sense in which law-making
bodies are entitled to enforce law coercively or to authorize other parties to enforce
law on their behalf.
A legislature can also have an entitlement to coerce in the face of certain sorts of
moral reasons against coercive enforcement. Law-making bodies could have a
general entitlement to enforce law coercively even if there were weighty moral
reasons not to enforce particular laws coercively. For instance, a government might
have reason not to enforce a particular law coercively because coercive enforcement
would deter innocent and desirable conduct that is in the neighborhood of
prohibited conduct. A law-making body might have an abstract entitlement to
coerce though none of the available forms of coercive enforcement are morally
permissible. Imagine, for instance, that a society’s severely limited resources made
corporal punishment, public humiliation, and confinement under inhumane con-
ditions the only available sanctions. But a law-making body lacks an entitlement
to coerce if there are unmet moral objections to that body’s use or threat of coercion
as such.17 It is incoherent to say that a body has an entitlement to coerce though
there are unmet moral objections to the body’s use of coercion in any form and for
any purpose.
There is in all societies a pro tanto objection to coercive enforcement of law—an
objection that rarely gets its due. Most explanations of governments’ entitlement to
coerce focus exclusively on showing either that prospective offenders have no
legitimate objection to state coercion, or that the interests of the victims of legal
wrongdoing outweigh or override the objections of prospective wrongdoers. An
adequate explanation of government’s alleged entitlement to coerce must also
address the effects of governmental coercion on those who do not violate the law
and have no desire to do so. Any coercive enforcement mechanism, however well
designed and however virtuous its agents, sometimes interferes with the entitle-
ments of innocent persons. Interference with entitlements of the innocent is not
merely a feature of particular forms of governmental coercion. It is a necessary
feature of coercive enforcement. This does not show, of course, that coercive
enforcement is impermissible. But it shows that coercion stands in need of
justification. It is necessary to show that the problem coercion solves is more serious
than the problem it introduces.
There are broadly two ways in which laws can be enforced coercively, and they
interfere with the entitlements of the innocent in different ways. One way of
enforcing a law coercively is to threaten a sanction for non-compliance. The threat
of sanctions for non-compliance sometimes interferes with the freedom of the
innocent because whoever decides when sanctions should be applied will inevitably
sometimes apply a sanction to an innocent person. Such errors need not reflect
malice or negligence on the part of law enforcers; they can result from good faith
mistakes. If a coercive enforcement mechanism is meant to address negligence and
intentional wrongdoing, as well as legal wrongs that result from good-faith, non-
negligent error, we must also be concerned about negligence and malice on the part
of those who make accusations of legally cognizable wrongdoing, those who make
decisions about guilt or innocence, and those who decide what sanctions to impose.
Legal systems can be designed to deter false accusations and to prevent people from
being wrongly convicted or found liable. However carefully a legal system is
designed, inevitably the evidence will sometimes support imposition of a legal
sanction on a person who is in fact innocent. Assuming that legal sanctions imposed
on innocent people deprive those people of things to which they are entitled, the
imposition of coercive sanctions necessarily interferes with the entitlements of the
innocent.
The other form of coercive enforcement is the direct application of force to
compel compliance. Direct application of force includes application of force against
someone’s person—for instance, dragging a trespasser off private property. It also
17 Likewise, private parties have no entitlement to enforce law coercively if there is an unmet moral
objection to coercive law enforcement by private parties. The pro tanto argument against governmental
coercion applies also to private coercive enforcement of law, except perhaps in the narrow case of self-
defense against an imminent threat.
includes at least some other actions that would, outside of the context of law
enforcement, violate a right of the actions’ target. For instance, police use force
when they break into a private home to retrieve stolen property or to obtain
evidence of crime. When direct application of force is used to enforce a law
incorrectly, or when it is used to enforce an incorrect governmental order, the
innocent target of enforcement suffers a double injury. If the police drag someone
away from a place he was legally entitled to occupy, they do not only deprive him of
his entitlement to be where he was; they also violate his entitlement not to be
manhandled. If a court incorrectly finds that an object has been misappropriated,
and it orders police to retrieve it from a private home, the residents do not only lose
an object they were entitled to possess; they also suffer a violation of privacy. Now,
in a society that lacks coercive enforcement of law, people can be subject to
incorrect orders from courts, police, or other public officials. If these orders are
made with morally legitimate authority, people may be morally required to give up
things they are entitled to, and if they are morally upright, they will comply. But
they will not suffer the additional injuries they would have suffered had the
incorrect orders been enforced by direct compulsion. Because some governmental
orders will inevitably be incorrect even if government officials always perform their
duties in good faith, the direct application of force to compel compliance with
orders will inevitably sometimes violate the entitlements of the innocent. More-
over, it will sometimes violate entitlements that would not be violated in a legal
system that lacks coercive enforcement.
Because both coercive threats and direct application of force necessarily risk
violating the entitlements of the innocent, no government can have an entitle-
ment to coerce unless establishing a coercive enforcement mechanism would
solve a problem that matters more than the problems it would create. Now,
the most obvious problem a coercive enforcement mechanism could address is
the problem of lawbreaking that arises from human moral failings. There are two
other problems coercive enforcement could address, however. These are problems
that could arise even in a society of morally very good people. One is the
problem of moral disagreement. Competent moral reasoners can make errors of
moral judgment. In particular, they can mistakenly believe that a particular law
or governmental order is not morally binding, or worse, that they have a moral
obligation to disobey. One might think that coercive enforcement of law is
needed to address violations of law that result from such errors of moral
judgment. Second, one might think that public coercion provides an important
assurance even if citizens would not in fact violate the law in the absence of
public coercion. Coercive enforcement provides each member of society an
assurance that others will not violate their legal rights without providing compen-
sation. Individuals need this assurance even if there would be no rights violations
without it because they should not have to depend on others’ good will. The next
two subsections will show that in a society of morally very good people, neither
of these arguments succeeds in showing that there is an entitlement to enforce
law coercively.
3.2. Moral error and compliance

Morally very good people, as I have characterized them, are not immune from
moral error. They reliably reason competently about moral matters, but because
some moral questions are hard, competent moral reasoning can lead to plausible
but incorrect judgments. Competent moral reasoners can reasonably disagree about
various moral questions, including the justice of an existing property system. Even
if the currently accepted system of property rules is not unjust in any way that
justifies disobedience, a competent moral reasoner might reasonably think other-
wise. Likewise, a competent moral reasoner might erroneously but reasonably think
that a court’s decision is so unjust that disobeying the court is morally permissible.
One might think that coercive enforcement of law is needed, even in a society of
morally very good people, to deal with well-intentioned but wrongful violations of
the law.
The first point to make about this line of argument is that there is only a limited
range of cases in which fully competent moral reasoners will disagree with each
other not only about what the law should be, but also about whether one should
obey the law even if the law is not what it should be. Though competent moral
reasoners can make errors of moral judgment, they do not make gross errors.
It would be a gross error to think that it is unimportant for a society to have a
generally respected system of rules for the allocation of physical resources. It would
also be a gross error not to recognize that it is often morally better to follow the
generally accepted rules of property than to follow the rules one thinks people
should be following. There are many possible systems for the allocation of physical
resources. Some of them are morally intolerable. All competent moral reasoners
recognize, however, that a variety of systems of property rules are morally tolerable
in the following sense: it is morally better for one of these systems to have every-
one’s compliance than for some to follow that system while others follow the rules
of another system they regard as morally better.
There are some cases, however, in which competent moral reasoners can disagree
not only about what rule they should follow, but also about whether following the
same rule as everyone else is more important than following the best rule or the
right rule. Imagine, for instance, that there is disagreement about whether it should
be possible to own animals of a certain species. The animals of this species have a
high cognitive capacity, though they do not have the same mental powers as human
beings. Most people believe that it is acceptable to keep these animals in captivity
under good conditions, and that it would be wrong to release a captive animal
without the putative owner’s consent. Some animal rights activists believe that
animals of this species are close enough to humans in cognitive capacity that
keeping them in confinement is tantamount to kidnapping. Releasing captive
animals of this species without the alleged owner’s consent is not only morally
permissible; it is required. The activists disagree with the majority not only about
what the rules of property should be, but also about whether in this case it is
important to follow the same rules as everyone else if the rules are bad. Suppose that
the animal rights activists are mistaken in these moral judgments. Though the
animal rights activists are in error on this matter, their error does not reflect moral
incompetence or lack of moral seriousness. It is the sort of error that competent
moral reasoners can make.18 Many of the animal rights activists will disobey
their society’s property law, though they are not morally justified in doing so.
Governmental coercion would help to protect animal owners from the well-
meaning but misguided animal rights activists.
Though moral disagreement of this sort can lead to morally wrongful violations
of law, even in a society of very good people, it does not justify establishing a
coercive enforcement mechanism. When private individuals disagree with the
public both about what the law should be and about whether individuals should
obey the law even if it is the wrong rule, sometimes the dissenters will be wrong.
Sometimes, however, the dissenters will be right. Perhaps well-designed public
bodies are more likely than morally competent private individuals to be right about
whether disobeying a given law would be morally justified. Even well-designed
public bodies will be far from infallible about this matter, however. When dissent-
ers correctly believe both that a law is bad and that it should be disobeyed, even
though it is the law, it would be wrong to attempt to coerce the dissenters to
comply. An attempt to coerce compliance would be an attempt to coerce people to
do something they are entitled not to do. It may even be an attempt to coerce
people to do something wrong.
Thus, if controversial laws are enforced coercively, sometimes the enforcing body
will prevent wrongs or address wrongs that have occurred, but sometimes the
enforcing body will commit wrongs itself. If controversial laws are not enforced
coercively, there will be wrongs by private parties that go unaddressed, but no
public body and no private party authorized to act on behalf of the public will be
guilty of wrongful coercion. Perhaps there is reason to believe that the private
wrongs coercive enforcement would address would be somewhat greater in number
than the public wrongs coercive enforcement would involve. There is no reason to
suppose, however, that in a society of competent moral reasoners, government
would necessarily be right a great deal more frequently than dissenters. There is also
no reason to suppose that the private wrongs coercive enforcement would address
are more serious wrongs than the public wrongs coercive enforcement would
involve. There are two reasons to think the public wrongs should matter more
from the perspective of the legislature. First, there is at least arguably a general
moral distinction between doing and allowing: it is worse to commit wrongs than
to allow wrongs of comparable severity. Second, when public officials commit
wrongs while acting in their official capacity, they purport to act on behalf of all
18 This assumes that the animal rights activists are committed to limiting their use of violence. It is
not plausible to think that one is justified in killing or injuring human beings in order to free captive
animals. In general, when a great many competent moral reasoners think that doing X is impermissible,
it is a gross error to believe that one is justified in killing or injuring others in order to facilitate doing
X. That many competent moral reasoners believe that X should not be done ought to give pause to
anyone who is tempted to use serious forms of violence to promote an opposing view. (I owe this point
to Thi Nguyen.)
citizens. Because their wrongful acts implicate people other than themselves, they
have more gravity than wrongs committed by private citizens not acting under color
of law. The somewhat greater frequency of private wrongdoing does not clearly
make up for the greater importance of wrongs committed by public agents. So
although moral disagreement would lead to a limited amount of non-compliance
with law in a society of morally very good people, that fact alone does not entail that
there would be an entitlement to enforce law coercively.
3.3. Rights and assurance

There is another reason one might have for thinking that a society of morally very
good people would need coercive enforcement of law. One might argue that such a
society needs a coercive enforcement mechanism to provide individuals assurance
that others will not violate their rights without providing compensation. Arthur
Ripstein offers an argument, on Kant’s behalf, that it would be important to have
such a coercive assurance even under social conditions in which violations of law do
not occur. Ripstein argues that any society needs public coercive enforcement of
legal entitlements to property, as well as certain other legal rights of individuals,
among them the right to bodily integrity.19 In order for us to enjoy external
freedom, we each need to be able to set and to pursue ends of our own choosing.
Our choice of what ends to pursue must not be “subject to the choice of another
person.”20 I can set and pursue ends of my own choosing only if my use of my
means, including my person and my property, is not subject to the choice of others.
For instance, if you can decide to lock me in a room, my use of my person is subject
to your choice. If you can decide to drive my car without my permission, my use of
my car is subject to your choice. My use of my means is subject to your choice even
if it is empirically unlikely that you will interfere. A slave does not enjoy external
freedom if he has a benevolent master who allows him to make his own decisions
and gives him access to resources with which to act. The slave is subject to his
master’s will even if as an empirical matter he can expect his master never to forbid
him from doing anything he wants to do or to require him to do something he does
not want to do.21
For us to enjoy external freedom, then, our freedom needs to be subject to
reciprocal limits. You must not be free to interfere with my use of my means, and
I must not be free to interfere with your use of your means. Interferences with
individuals’ use of their means must therefore be stopped. Violations of rights may
be coercively stopped before they can be completed; for instance, a theft in progress
may be forcibly interrupted. When a violation is completed, the offender must
be forced to restore the victim’s means. For instance, a thief must be compelled
to return stolen property or, if this is no longer possible, to pay monetary compen-
sation. Now, limits on our freedom are not reciprocal if rights are enforced
privately, since some people are stronger, smarter, or more willing to confront
19 Ripstein (2009; 2004). 20 Ripstein (2004: 8).

21 Ripstein (2009: 15, 36).
others and thus more able to enforce their rights effectively. So for our freedom to
be subject to reciprocal limits, the government must enforce the rules that specify
the limits on our freedom.22 In particular, it must provide a coercive system by
which those whose rights are violated can have their means restored to them.
It would not be enough to have a non-coercive court system that hears disputes
about alleged rights-violations and makes judgments about whether compensation
is owed, and in what amount. There needs to be a coercive mechanism for ensuring
that court-ordered damages are paid. On Ripstein’s view, such a system is needed
even in a society in which people are seldom tempted to wrong each other, whether
because they have little to gain from wrongdoing or because they have strong moral
commitments.
The difficulty with this argument is that it is unclear what sort of reciprocal limits
on freedom we need if we are not to be subject to each others’ wills. The limits we
need could be moral limits: I enjoy external freedom only if you normatively may
not interfere with my use of my means and I am entitled to compensation if you do
interfere. Alternatively, the limits we need could be empirical limits: I enjoy
external freedom only if it is empirically not possible for you to interfere with my
use of my means without providing compensation. If the limits we need to enjoy
external freedom are purely normative limits, then at least in ideal social circum-
stances, external freedom does not require coercive enforcement. No coercion is
needed for people to have moral entitlements to use resources without others’
interference. No coercion is needed for people to be morally entitled to compen-
sation if others do interfere with their means: if one person in a society of very good
people takes another’s coat by mistake, the accidental thief is morally required to
return it when she learns of her mistake. Moreover, no coercion is needed for these
moral entitlements to be socially and politically recognized.23 Providing coercive
remedies for wrongdoing is one way of publicly recognizing an entitlement, but it is
not the only way. In a society of very good people, entitlements under the property
system are publicly recognized in the rules of the code and in documents such as
wills and deeds that the system recognizes as having legal force. So if coercive
enforcement of law is needed even in ideal social circumstances, as Ripstein claims,
this must be so because the limits on freedom that need to be established are at least
partly empirical limits, not merely normative limits.
If external freedom requires that it be empirically impossible for anyone to
interfere with other people’s use of their means without being forced to restore
them, then external freedom is unattainable. Even under the best legal systems,
22 Though Ripstein’s main argument is for the coercive enforcement of individual rights, it can be
extended to an argument for coercive enforcement of other laws insofar as these laws are required for
the establishment of a rightful condition in which limits on freedom are reciprocally enforced. For
example, the state must redistribute wealth by means of coercively enforced taxation in order to prevent
people in extreme need from becoming dependent on the will of other private individuals. (2009: 267–
86); (2004: 33–5).
23 Social and political recognition of rights matters; it is what the slave with the benevolent master
lacks. Though this slave has a moral entitlement to decide for himself what to do, and though he can be
confident that his master will never actually interfere, the slave’s entitlement to make his own decisions
is not socially recognized.
people can sometimes get away with violating others’ rights without providing
compensation because there is insufficient evidence to support a judgment against
the wrongdoer. Moreover, if it is a contingent matter whether private citizens will
conform to their legal duties, it is also a contingent matter whether rights enforcers
will perform theirs. Judges can be corrupt, incompetent, prejudiced, or negligent.
Juries can find for a sympathetic defendant even if the evidence supports a finding
for the plaintiff. For both of these reasons, there is no way to eliminate the
possibility that one person could interfere with another’s use of her means. In a
society of morally flawed people, a good legal system can make unjustifiable
violations of rights less common, and it can make it more likely that people
whose rights are violated will receive compensation, but it cannot ensure that
every unjustifiable violation of a right will be compensated. In a society of morally
very good people, uncompensated violations of rights are always possible, even if
they never actually occur. Coercive enforcement cannot remove this possibility.
So in a society of morally very good people, a coercive enforcement mechanism
would not give people anything that they did not already have. Even in the absence
of coercive enforcement, people could have socially and politically recognized rights
that others are morally required to respect. Because of their strong moral commit-
ments, violations of these rights would not occur, except in cases in which the
obligation to respect these rights is legitimately controversial—and in these cases,
coercive protection of rights is unjustified. It remains possible that others could
violate their rights without providing compensation, but no coercive enforcement
mechanism can eliminate this possibility. Thus, in a society of morally very good
people, a coercive enforcement mechanism would serve no good purpose. It is
necessary neither to provide citizens with a coercive assurance mechanism nor to
address the problem of reasonable moral disagreement. In a society of morally very
good people, then, there is no way to overcome the pro tanto objection to the
establishment of a coercive enforcement mechanism—that it necessarily places
burdens on the freedom of the innocent. Therefore, the legislature in such a society
would have an entitlement to make law, but it would have no entitlement to
enforce law coercively or to authorize others to enforce law coercively.
One might be tempted to characterize the powers of the legislature in a society of
very good people differently: the legislature has an entitlement to coerce (or to
authorize coercion), but under current social circumstances, it should not exercise
that entitlement. This position does not make sense. The fact that coercive
enforcement mechanisms interfere with the freedom of the innocent is no more
contingent than the fact that coercive enforcement mechanisms interfere with
wrongdoing. The interests of potential victims of government abuse, error, and
intrusiveness are no less important than the interests of potential victims of private
wrongs. To show that a legal system has an entitlement to coerce, it is necessary to
show that its benefits are more important than any moral objections to its funda-
mental features. In a society of very good people, this cannot be done. Perhaps if
social circumstances were to change, law-making bodies would acquire an entitle-
ment to coerce. In the happy circumstances people find themselves in, neither the
legislature nor any other party is in any sense entitled to enforce law coercively.
4. The entitlement to coerce in non-ideal societies
Actual human societies, of course, are not made up entirely of morally very good
people. I do not wish to deny that in a large society made up of morally flawed
human beings, government sometimes has an entitlement to enforce law coercively,
and that it sometimes has good reason to make use of this entitlement. But there is
no justification for a general presumption that any government in a non-ideal
society is entitled to coercively enforce any law it is entitled to make. That
government has an entitlement to coercively enforce a legal requirement on
conduct does not follow from the fact that it has the power and the entitlement
to impose this requirement together with abstract truisms about human nature (e.g.
that human altruism is limited and that humans are vulnerable to weakness of
will).24 To show that government is entitled to coercively enforce a requirement on
conduct, it is necessary to do an empirical investigation that goes beyond truisms
and examines the extent and character of citizens’ moral failings in an actual society.
It is also necessary to ask the moral question whether the need to address violations
of this particular legal requirement justifies establishing a coercive enforcement
mechanism that inevitably puts the innocent at risk of being injured by the
government or its agents. Though there may be no actual human societies in
which government lacks the entitlement to coerce altogether, there are imaginable
social circumstances in which government’s justified power to impose legal require-
ments on conduct is considerably more extensive than its entitlement to enforce
these requirements on conduct coercively.
There are two reasons one might have for believing, to the contrary, that in any
society not made up entirely of superhuman angels or saints, government will be
entitled to enforce whatever requirements on conduct it is entitled to impose. First,
one might think that in a society not made entirely of morally very good people, it
would not be possible to establish laws that are morally binding. The existence of
unpunished free-riders would undermine others’ obligation to obey the law.
Second, one might think that the presence of people who are inclined to break
the law unjustifiably will necessarily give the government an entitlement to address
lawbreaking coercively. In any imaginable human society, the threat of unjustifiable
lawbreaking will present a problem that justifies burdening the innocent with the
risk of being mistakenly or maliciously subjected to legal sanctions or governmental
force. I will address these arguments in turn.
4.1. Free-riding and the moral force of law

The concern about free-riding begins with the thought that people have a moral
obligation to do their part in a cooperative activity only if a cooperative activity is
actually taking place. For a cooperative activity to take place, enough people need to
24 Hart identifies five such truisms: “human vulnerability,” “approximate equality,” “limited
altruism,” “limited resources,” and “limited understanding and strength of will” (1994: 193–200).
be doing their part in that activity. So for people to have a moral obligation to do
their part in a particular system of property rules, for instance, there need to be
enough people who actually follow the system’s rules. If there are not enough
people following the rules of a particular system of property, then nobody has an
obligation to follow that particular system, though everyone has a duty to try to
bring about the existence of a system that is widely followed. Now, perhaps a
system of property that lacked coercive enforcement might be widely obeyed on an
island occupied only by very good people. In a society of flawed human beings,
however, many people are less than fully committed to treating each other fairly.
Thus, a non-coercive system of property could not retain sufficient obedience for it
to count as the established system. For rules of property to remain morally binding
over time in a society of flawed human beings, there must be a coercive assurance
that people will generally continue to comply. Or so one might argue.25
The force of this argument depends on how severe people’s moral failings
actually are. If these moral failings are mild enough that a non-coercive system of
property could secure widespread compliance, though not universal compliance,
that system could be morally binding. Imagine a society of people who fall short of
sainthood but are, as a group, fairly good. A small number of them do not
understand the moral reasons for respecting a system of property. The rest are
committed in principle to respecting each others’ private property and to respecting
limits on the use of public property, but on rare occasions they suffer from
weakness of will. Unjustifiable intentional violations of the rules of property will
occur in this society, but they will not be common. For such people, moderately
flawed though they are, a morally binding non-coercive system of property would
be possible. That people occasionally fail to do their part in a cooperative activity
and that a few people persistently free-ride does not entail that no cooperative
activity is taking place, nor does it undermine people’s obligation to do their part in
that activity. To determine whether morally binding rules of property can be made
non-coercively, it is necessary to have empirical evidence about how much compli-
ance could be achieved without the use or threat of coercion. The abstract truism
that human beings are sometimes inclined to break the law does not entail that law
cannot be morally binding without coercive enforcement.
One may object to this point with a further argument. There are some forms of
cooperation for which the existence of even a few free-riders undermines everyone
else’s obligation to comply. Imagine, for instance, that businesses in a highly
competitive market are told to pay a tax in support of important public goods.
Suppose that the tax is not coercively enforced. If a small number of business
owners, caring little for the public good, opt not to pay the tax, these businesses will
have a competitive advantage. The owners of other businesses in this market may
25 Rawls argues that even if everybody had a shared sense of justice and wanted to adhere to the
existing system of cooperation, the legal system would have to threaten coercive sanctions (possibly
mild sanctions) in order to secure stable cooperation over time. Without sanctions, people would be
tempted not to fulfill their obligations because they would suspect that others are not doing their part
(1971: 240).
very much want to pay the tax, but paying the tax may put their businesses at risk,
depending on how burdensome the tax is and on the economic circumstances these
businesses face. To borrow Hobbes’s words, these businesses fear that by paying the
tax they will make themselves “prey to others.”26 Suppose that economic circum-
stances are such that businesses will, in fact, be putting themselves at risk if they pay
the tax and some of their competitors do not. Then it would be unreasonable to
demand that a business pay the tax unless business owners can be reasonably
confident that other businesses are paying. In a society of saintly people, business
owners could perhaps be confident that their competitors voluntarily comply with
their tax requirements. In a society of flawed human beings, we cannot be so
confident that competitors are voluntarily complying with the rules. So in order for
businesses to have a morally binding obligation to pay this tax, there needs to be an
enforcement mechanism to ensure that payment is compatible with a business’s
ability to compete on fair terms.
There are two problems with this argument. First, coercive enforcement is not
the only way of providing businesses with an assurance that competitors are
complying with a tax. The government could call for businesses’ taxes to be paid
in regular installments and publish information about businesses’ tax payments.
This would provide businesses with information about which of their competitors
pay their taxes and which, if any, evade it. The second problem with the argument
is that it only applies to free-riding by businesses in a competitive market. By way of
contrast, suppose that a tax paid by individual consumers is not coercively enforced,
and that 90 percent of individuals make a good faith effort to pay the amount they
owe. They would not thereby make themselves prey to the non-compliant 10
percent. Even in lean times, most consumers do not compete with each other for
survival. Having more money is good, but it does not make the difference between
survival and death. That 10 percent of consumers shirk their obligation to pay a tax
does not undermine other taxpayers’ moral obligation to pay.27 Many of the rules
of property are more like taxes on individual consumers than they are like taxes on
businesses in a competitive market. Though very widespread non-compliance with
a rule would perhaps undermine the obligation to obey it, a moderate amount of
non-compliance would not. If everyone regarded themselves as perfectly free to ride
any bicycle found on the street (even if doing so involved picking a lock), perhaps
nobody would be under moral obligation not to do so. Bicycles would, in effect, be
common property. If only a small but substantial minority has this attitude toward
26 Hobbes (1994: 99, II.xv.36).

27 Another possible objection to a voluntary tax system is that it would put excessive demands on
the will. Nagel argues that it is reasonable to impose a compulsory tax for the support of the poor but
unreasonable to demand that people make voluntary financial contributions in the same amount. “The
latter is an excessively demanding moral position,” he writes, “because it requires voluntary decisions
that are quite difficult to make” (1975: 145). The problem Nagel describes would be a real problem in
a non-coercive tax system, but it could be mitigated by structuring tax collection to make tax payment
psychologically easier. For instance, it would probably be helpful to make use of automatic tax
withholding from paychecks. I suspect that for most people it is psychologically harder to write a
check than it is to refrain from demanding a refund.
bicycles, however, the moral obligation to treat bicycles as private property is not
undermined. That some people steal bicycles with impunity does not give me
justification for engaging in theft myself, nor does it undermine any of my other
obligations with respect to the property system.28
In order to answer the question whether non-coercive rules of property can be
morally binding, it will be necessary to examine different parts of the property
system separately. Some rules or sets of related rules will be morally binding if they
are widely followed; the obligation to follow them is not undermined if the rules are
not universally followed or if there is no effective enforcement mechanism. Other
rules or sets of related rules will be morally binding only if they are universally
followed or if they are enforced (coercively or otherwise) in such a way that
compliance is appropriate even in the face of non-compliance by competitors. In
order to determine which category a rule or a set of related rules falls in, it is
necessary to address both the empirical question what competitive pressures agents
face as market participants and the moral question what risks a market participant
can be reasonably asked to assume. Having determined what level of compliance is
required for a rule or set of related rules to be morally binding, there will be a
further empirical question whether this level of compliance can be achieved without
coercive enforcement.
I cannot pretend to have enough knowledge of psychology or sociology to
answer definitively the question how much obedience non-coercive rules could
secure, but I will mention some motives that could lead people to comply with
law voluntarily. First, though there are often powerful incentives of self-interest
to violate laws, sometimes violating a law is not in one’s self-interest even if legal
and social sanctions are not taken into account. For example, it is typically not in
a radio broadcaster’s interest to transmit on frequencies that others in the area
have an exclusive right to use. If a broadcaster transmits on a frequency that
someone else is legally entitled to use, and that person is making use of her
entitlement, neither party’s use of the frequency is likely to be a success. Second,
when violations of a rule are visible to the public, concern for one’s reputation
can provide an incentive to follow the rule, and this incentive is not always a
coercive incentive. No doubt some informal social responses to perceived wrong-
doing are coercive, but it is not clear that all of them are. (If prospective
customers avoid businesses they regard as dishonest, not to punish those busi-
nesses but to avoid being cheated, is the resulting pressure coercive?) Further-
more, concern for one’s reputation is not only concern about social sanctions or
rewards that may result from perceived good or bad behavior; people care about
other’s good opinion for its own sake.29 Finally, it is important not to underesti-
mate the significance of motives other than narrow self-interest. Even people who
are not of saintly character often obey the law either out of the belief that there is
28 Furthermore, it is not unreasonable to demand that people refrain from stealing bicycles. It does
not place excessive demands on the will to ask people not to steal even when they could get away with it.
29 (Aristotle 2002: 101, 1097b3).
a moral obligation to obey or because they believe that there are independent
moral reasons to act as the law requires.30
In light of these considerations, I think it is likely that a large human society
could have a system of property rules in which all the rules are morally binding and
at least some of the rules are not coercively enforced. The more important point,
though, is that a substantial amount of empirical evidence is needed to settle the
question whether and when non-coercive rule making could be morally binding in
a large human society. Abstract truisms about human nature do not settle the
question.
4.2. Responding to unjustified lawbreaking

Given that government can make a morally binding law without coercively enfor-
cing it, there is a further question about whether the government has an entitlement
to enforce it coercively. One might think that that the presence of citizens who are
disposed to break the law unjustifiably will always present a problem that provides a
justification for establishing a coercive enforcement mechanism. To show that
government has an entitlement to coerce, however, it is not enough to show that
coercive enforcement would solve a problem. It is necessary to show that coercive
enforcement would solve a problem that matters more than the burdens a coercive
enforcement mechanism necessarily places on the innocent.
Showing that the value of coercive law enforcement justifies the burdens it
imposes on the innocent is not straightforward either normatively or empirically.
It is not straightforward normatively because the question must not be addressed as
a utilitarian calculation. One cannot simply add up the harms a coercive enforce-
ment mechanism could be expected to inflict, according to some measure of total
harm, and compare this with the harm coercive enforcement can be expected to
prevent.31 Even if such a comparison were possible, it would be inappropriate for
three reasons. First, it is necessary to consider the way in which the burdens of
coercive enforcement would be distributed. If innocent people of a particular class
or social group would be disproportionately likely to be punished or subjected to
compulsive force, this must be taken into account. Second, the injuries a coercive
enforcement mechanism inflicts on the innocent, unlike most injuries perpetrated
by lawbreaking individuals, come with the stigma of the government’s judgment
that one is guilty or likely guilty of an offense.32 The stigma associated with a
criminal conviction is most serious, but there is also a negative governmental
judgment associated with other coercive sanctions, such as fines and punitive
damages, and with direct applications of governmental force (e.g. being dragged
away by the police). Finally, it is at least arguably more difficult in general to justify
30 For evidence of this, see Tom Tyler’s study of several hundred Chicago residents’ compliance
with commonly broken laws (1990).

31 I do not mean to suggest here that retributive justifications of punishment should be off the table.
Larry Alexander discusses retributivist approaches to tradeoffs between punishing the innocent and
letting the guilty go unpunished (1983).
32 Dworkin (1985).
doing injury than it is to justify merely allowing harm or injury to occur. Some
defenders of a robust coercive criminal justice system question the relevance of the
doing-allowing distinction to the unintentional punishment of the innocent.
Instituting a practice of punishment knowing that some innocent people will be
mistakenly punished is not morally comparable to intentionally punishing particu-
lar people known to be innocent.33 That said, there may well be a moral difference
between doing something that risks injuring innocents without intending injury to
anyone (e.g. driving a vehicle) and intentionally injuring people, believing each of
these people to be guilty but knowing some of them will in fact be innocent.
Showing that coercive law enforcement presents a greater threat to the innocent
than would its absence is also a complex matter empirically. The relative magni-
tudes of these two threats depend on precisely what moral failings people in a given
society have. Imagine a decidedly non-angelic society deeply divided on ethnic
lines. Because of the prejudices that infect this society, many people in the majority
group believe that members of minority groups break the law far more often than
they actually do. As a result, police officers of the majority group are far too quick to
use force against members of minority groups. Criminal or civil juries dominated
by the majority group find against members of minority groups on evidence that an
unprejudiced juror would consider flimsy at best. Juries that are not dominated by
members of one ethnic group tend to hang even when the evidence clearly points
toward the defendant’s guilt, since most people are inclined to view others of their
own ethnicity charitably. In this society, the coercive legal apparatus would inflict
extensive wrongs on innocent people, and it would be largely ineffective at
preventing, compensating, or justly punishing wrongdoing by private parties. So
it would be a mistake to think that in any society that includes morally flawed
people, a well-designed coercive executive would present less of a threat than would
individual lawbreaking unchecked by a coercive executive. Which threat is greater
depends on which moral flaws people have. If people’s flaws make misuse of public
coercive power a greater threat than a given form of wrongdoing by private parties,
government will lack an entitlement to address that form of wrongdoing coercively.
Even in a society of morally flawed people, then, the entitlement to make law
does not straightforwardly entail an entitlement to enforce. No doubt actual
governments sometimes do have an entitlement to enforce law coercively, but
this entitlement may or may not extend to all of the laws governments are entitled
to make. How extensive this entitlement is depends on many empirical facts,
among them what non-coercive incentives there are to comply with various laws,
what pressures people feel that could justify non-compliance in the absence of
coercive enforcement, what temptations people feel most acutely, what temptations
people are able to resist, and what prejudices and vices afflict people acting as agents
of the state. The extent of the entitlement to enforce law coercively may wax and
wane as social conditions change.
33 Lippke (2010: 471); Wertheimer (1977). Though these arguments concern the justification of
punishment, they apply equally to other forms of governmental coercion.

5. Consequences
That the entitlement to make law does not entail an entitlement to enforce law
coercively has two important consequences for the way political argument should be
conducted. First, this conclusion shows that a common form of argument in both
political philosophy and public political discourse is dubious at best. Both economic
libertarians and civil libertarians often criticize proposed legal requirements by
arguing that coercive enforcement of these requirements would be objectionable.
Such arguments presuppose that the government is entitled to impose these legal
requirements only if the government is entitled to enforce them coercively. This
assumption is not true in general. Sophisticated libertarians need to consider
whether they should advocate limits on the reach of law or whether they should
instead advocate more non-coercive exercises of government’s law-making power.
Note that the libertarian argument does not turn on the claim that every direct legal
requirement on conduct must have a coercive remedy for non-compliance. It
depends instead on the weaker claim that government is entitled to use coercion
to enforce every requirement on conduct it is entitled to impose—and, conversely,
that government is not entitled to impose any requirement that it is not entitled to
enforce coercively. The conclusion that the entitlement to make law does not entail
an entitlement to coerce is thus more significant for political discourse than the
weaker conclusion that laws lacking coercive enforcement are possible.
Awareness that the entitlement to make law does not always come with an
entitlement to coerce can also help us to make more refined diagnoses of injustice.
As an example, many economic egalitarians argue that it is unjust to have a coercively
enforced property system unless that system satisfies certain requirements of dis-
tributive justice.34 If such systems are indeed unjust, we can think more clearly
about the nature of this injustice if we recognize that the justification of property law
does not entail the justification of coercive enforcement of property law. We may
find that the injustice of a coercively enforced property system lies solely in the
wrongful use of governmental force. We may find instead that the injustice of an
inegalitarian property system lies not only in what government forces people to do,
but also in what it asks of them. A government that claims to issue morally binding
laws and orders has a duty to ensure that those laws and orders reflect equal concern
for all citizens, whether or not the laws are backed by the threat of force.
References
Alexander, L. (1983). “Retributivism and Inadvertent Punishment of the Innocent,” Law

and Philosophy, 2: 233–46.
Aristotle (2002). Nicomachean Ethics. Translated by S. Broadie and C. Rowe. Oxford:
34 Dworkin (2002: 2); Blake (2001).

Blake, M. (2001). “Distributive Justice, State Coercion, and Autonomy,” Philosophy &
Public Affairs, 30: 257–96.
Dworkin, R. (1985). “Principle, Policy, Procedure,” in A Matter of Principle. Oxford:
Dworkin, R. (2002). Sovereign Virtue: The Theory and Practice of Equality. Cambridge, MA:
Harvard University Press.
Dworkin, R. (2006). Justice in Robes. Cambridge, MA: Harvard University Press.
Hart, H. L. A. (1994). The Concept of Law. Second edition. Oxford: Oxford University
Press.
Hobbes, T. (1994). Leviathan. Edited by E. Curley. Indianapolis: Hackett Publishing
Company.
Kant, I. (1996). The Metaphysics of Morals, reprinted in M. Gregor (trans.), Practical
Philosophy. Cambridge: Cambridge University Press.
Lamond, G. (2001). “Coercion and the Nature of Law,” Legal Theory, 7: 35–57.
Lippke, R. (2010). “Punishing the Guilty, Not Punishing the Innocent,” Journal of Moral
Philosophy, 7: 462–88.
Nagel, T. (1975). “Libertarianism without Foundations,” Yale Law Journal, 86: 136–49.
Nozick, R. (1974). Anarchy, State, and Utopia. New York: Basic Books.
Rawls, J. (1971). A Theory of Justice. Cambridge, MA: Harvard University Press.
Raz, J. (1999). Practical Reason and Norms. Second edition. Oxford: Oxford University
Press.
Ripstein, A. (2004). “Authority and Coercion,” Philosophy & Public Affairs, 32: 2–35.
Ripstein, A. (2006). “Beyond the Harm Principle,” Philosophy & Public Affairs, 34: 216–46.
Ripstein, A. (2009). Force and Freedom: Kant’s Legal and Political Philosophy. Cambridge,
MA: Harvard University Press.
Sager, L. (1978). “Fair Measure: The Legal Status of Underenforced Constitutional
Norms,” Harvard Law Review, 91: 1212–64.
Simmons, A. J. (1987). “The Anarchist Position: A Reply to Klosko and Senor,” Philosophy
& Public Affairs, 16: 269–79.
Simmons, A. J. (February 2009). “Philosophical Anarchism.” Available at SSRN: <http://
ssrn.com/abstract=1344425> or <http://dx.doi.org/10.2139/ssrn.1344425> (accessed
10 August 2012).
Tyler, T. (1990). Why People Obey the Law. Princeton: Princeton University Press.
Waldron, J. (1988). The Right to Private Property. Oxford: Oxford University Press.
Wertheimer, A. (1977). “Punishing the Innocent—Unintentionally,” Inquiry, 20: 45–65.
PART III
CONCEPTUAL ANALYSIS
9
Farewell to Conceptual Analysis
(in Jurisprudence)*
Andrei Marmor
Analytical legal philosophy is not an exercise in conceptual analysis. This is the

main thesis I hope to establish in this article. Historically, conceptual analysis may
have played some role in H. L. A. Hart’s jurisprudence, but it was rather tangential
to the main project of his theory. Hart’s main objective in The Concept of Law was
not essentially different from that of Austin, namely, to provide a reductionist
theory of law. The main purpose of Hart’s theory was to offer an explanation of law
in terms of something more foundational in nature, that is, in terms of social facts,
which, in turn, can be explained by reference to people’s actual conduct, beliefs,
and attitudes. Hart’s objection to Austin’s reductionism was not to the idea of
reduction but to the particular building blocks that Austin used in articulating his
theory. And I think that this is generally the case: My claim is going to be that the
main methodological thrust of legal positivism is reductionism, not conceptual
analysis. And the main objections to legal positivism are best seen as a denial of the
possibility of such a reduction.
1. Conceptual analysis
The title of Hart’s seminal book, The Concept of Law, gave the impression that the
concept of law stands for something of great importance, and that it is somehow
different from an ordinary understanding of what the word “law” means, yet
awaiting philosophical elucidation. Of course the word “law,” at least in English,
has many uses, such as in “laws of nature,” “the second law of thermodynamics”
and the like, that have nothing, or almost nothing, to do with law in its juridical
sense.1 But if we focus our attention on the use of “law” in the juridical context, it is
* I am greatly indebted to Nicos Stavropoulos and Wil Waluchow for helpful comments on earlier
drafts of the chapter.
1 The etiology of “law” is rather complex, of course. Distinctions and very different uses we take for
granted today may not have emerged for centuries.

210 Andrei Marmor
very doubtful that there is something that the concept of law stands for, distinct
from what people ordinarily mean by the term.
Let us begin with the obvious question: What is a concept? And what is it that we
try to do when we purport to explain or analyze the concept of X? Gilbert Ryle gave
us a pretty good sense of what he meant by “concept” and what conceptual analysis
consists in:
. . . concepts are not things, as words are, but rather the functionings of words, as keeping
wicket is the functioning of the wicket-keeper. Very much as the functioning of the wicket-
keeper interlocks with the functioning of the bowler, the batsman and the rest, so the
functioning of a word interlocks with the functioning of the other members of the team for
which that word is playing. One word may have two or more functions; but one of its
functions cannot change places with another.2
Clearly enough, according to Ryle, concepts are not abstract objects; they are not
things of any kind. A concept designates the myriad ways in which a word is used
by competent speakers of the relevant language in a given language game—that is,
ways in which it plays specific roles in making moves within an interlocking set of
other concepts and arguments. Some concepts are more regimented than others,
depending on the particular discipline or setting in which they are used.3 But even
with regimented or scientific concepts, Ryle thought, it would be hopeless to look
for a set of necessary and sufficient conditions that would provide the set of
functionings a concept has in a language game. Rather we are looking at piecemeal
examination of families of conceptual connections, ways in which the functioning
of a word is dependent on another. Furthermore, Ryle, like other philosophers of
the ordinary language analysis school at the time, believed that these conceptual
connections are epistemically transparent. A careful examination of the ways in
which we use language at particular settings is obvious to any competent user of the
relevant language. Because we know, and necessarily so, the meaning of the words
we use as competent members of the relevant linguistic community, we will always
be struck with the undeniable correctness of any genuine conceptual connection
whenever it is presented to us. How much and how far we can generalize from these
connections, and how deep the solution they may provide for various philosophical
questions, is not entirely clear. Ryle, like Wittgenstein, clearly thought that most
philosophical problems are conceptual problems.
Still, the relation between the meaning of a word and the concept it designates is
somewhat unclear. Words can mean different things, depending on the contexts or
settings in which they are used. Ryle thought that one and the same word can label
different concepts. The concept designated by the word “law” for example, would
be quite different when used in a juridical context from the same word used, say, in
the context of a theory in physics. Quite right, but this does not answer our
question: Once we identify the concept—that is, the general functioning(s) of a
word in a given type of setting, say, “law” in the juridical sense—is the concept
different from what the word means in that general setting? To use the familiar
2 Ryle (1954: 32). 3 Ryle (1954: 33–4).

Farewell to Conceptual Analysis 211
Fregeian terminology: How is concept different, if it is, from sense? I am not trying to
suggest that a concept is, after all, some abstract object or entity (as Frege thought
about sense4). The question is more practical, as it were: When we try to elucidate
or analyze a concept, is there anything else to it than figuring out what the word, in
its relevant settings, means in the language in question? It is difficult to see how it
would be different. Take the word “chair,” for example. We use this word in
English to designate a certain type of artifact, those manufactured for people to sit
on, with a certain typical shape or form, etc. In short, the word “chair” in English
has a fairly clear meaning or sense, designating a certain type of artifact. Now we
can call this a concept, if we like; we can say that all reasonably competent users of
English have a concept of chairs. But it is difficult to see how the concept of chairs is
different from whatever this word means in its standard use in English expressions.
If I asked you to tell me what “chairs” are, would you feel compelled to ask me, “Do
you mean the concept of chairs, or just the meaning of the word in English?”
Giving me an answer to the first would also give me an answer to the second.
Now you might think that there is a difference when less mundane words are at
stake, words such as democracy, courage, or knowledge. As we know, people tend
to have different conceptions of what democracy is, and they may have somewhat
different conceptions of what constitutes courage or what knowledge really is. And
thus it has become common to distinguish the concept from its possible concep-
tions. We all understand the concept of democracy, it is assumed, while we may
disagree about its favored conception. I have no objection to putting things this
way. But let us focus on the concept. People have a concept of democracy insofar as
they understand what the word means, know what it stands for, and, by and large,
use it correctly in their expressions. And thus, by elucidating what the word
democracy stands for, you have given its meaning; at the same time, you have
explicated what the concept is. Suppose, to illustrate, that someone who grew up in,
say, North Korea, has absolutely no idea what democracy is; let us assume that he
has never heard the word. Now, having escaped from North Korea, he hears the
word for the first time and asks you what it means. Let us assume that you give this
person a correct answer. You explain everything that would enable him to use the
word correctly—that is, you give him the correct meaning of the word in English.
Is there anything left for him to know in order to grasp the concept of democracy?
Is there any point at which you can say, “You know perfectly well what ‘democracy’
means, but you have not yet quite grasped the concept”?5 I fail to see what this
residual piece of information might be.
One might object to this by pointing out how the mastery of numerous concept
words in a natural language requires not simply a form of knowing-that, but also
some form of knowing-how; mastery of a concept is often a matter of habituation,
4 Frege’s view about the idea of sense was, curiously, a form of Platonic realism. On the significance
of this stance, see Soames (2010: 12–16).

5 Or does it work the other way around? Is there a point at which we can say, “Well, you may have
the concept of democracy, but you do not quite know what the word means”? That I find even more
mysterious.
212 Andrei Marmor
like a skill we acquire, which takes time, practice, etc. This is absolutely true, and
those of us who had to learn a second language know exactly how frustrating and
difficult this process of habituation often is.6 But this would not show that there is a
gap between mastering the meaning of a word in a natural language and acquiring
its concept. You do not really know what a word means without an adequate grasp
of what it stands for and how to use it in different contexts. If it takes habituation
and practice to acquire this kind of knowledge, that is what it takes to grasp the
meaning of the word and be able to use it correctly. Once you have acquired that
competence, you have acquired the concept, and vice versa. No residue is left here,
in either direction.
Perhaps in one sense this conclusion is too strong. Philosophers sometimes argue
for conceptual claims that are explicitly acknowledged as revisionist; such concepts
are not meant to reflect an agreement in judgments about the concept’s application
to all its standard cases. Therefore, counterexamples to the conceptual claim, based
on prevailing linguistic intuitions or common usage, would not necessarily refute a
revisionist concept. I am not denying here the possibility of such claims; it is
certainly possible to define a concept that is revisionist in some sense. But then the
concept would need to be supported by theoretical arguments about its usefulness
or theoretical purpose, not by anything traditionally called conceptual analysis. An
analysis of concepts is an analysis of the ways in which words function in our actual
language games, which must be based on observation of linguistic practices and
prevailing linguistic intuitions. Conceptual analysis cannot be revisionist. Concepts
can be revisionist in various ways if they are properly defined and serve some useful
theoretical purpose, but that is quite a different matter.
Thus, if we assume that reasonably competent speakers of a natural language
know the meaning of the word law (in its juridical sense), then they understand the
concept of law.7 And, crucially, there is nothing more to the concept than that
which is designated by the word. But now you might be tempted to conclude that
this line of thought vindicates Dworkin’s famous “semantic sting” argument.
Dworkin claimed that conceptual theories of law, like Hart’s, are, essentially,
about the meaning of the word “law.” He argued that this quest for the analysis
of meaning is misguided because it flies in the face of essential aspects of legal
practice, most importantly, the fact that the main participants in the practice have
theoretical arguments about its concept—that is, about what law really is.8 One
standard reply to Dworkin has been that Hart was seeking to elucidate the concept
of law, not the ordinary meaning of the word “law.” People (myself included) have
pointed out that Hart explicitly rejected the idea that his theory is about the
meaning of the word “law,” or that such an attempt to define the meaning of law
6 See Marmor (2009: 68–9).

7 Knowing what a word means, and consequently, understanding its concept, admit of degrees; in
many cases it makes perfect sense to say that A has a better understanding of the meaning of X than B;
or that A’s understanding of the concept of X is more complete than that of B. Nothing that I say in the
text should be taken to imply otherwise.
8 Dworkin (1986: ch. 1).
would be theoretically fruitful.9 If I am right, however, that there is no significant

difference between understanding the meaning of a word and understanding its
concept, then this line of response is taking the wrong tack. In other words, if
you think (I do not) that Hart’s theory of law is basically an attempt to elucidate
the concept of law, you would have to agree with Dworkin that Hart’s theory is
an attempt to articulate what the word “law” means; so it is a “semantic theory”
after all.
I think that Dworkin is right, up to a point. He was right to assume that
conceptual analysis is, essentially, a linguistic inquiry. At least it has been so
conceived by the ordinary language analysis school of Wittgenstein, Ryle, and
Austin, to which Hart (half-heartedly, I think) belonged and which he was certainly
influenced by. As we saw in Ryle’s quotation, the idea of a concept stands for the
“functionings of words” in their settings, and ways in which these functionings are
intertwined with the functionings of related concepts with which they interact in a
given setting. The idea that conceptual analysis is transparent and reveals something
that can be recognized upon reflection as obvious is only made plausible by the
assumption that it is our own linguistic practices that we seek to elucidate here.
Conceptual connections are transparent because they are constituted by our
language, and language is public and knowable to every competent user. Concep-
tual analysis is, essentially, about language. I do not mean to suggest that philo-
sophical inquiry into the conceptual connections of language use is not interesting
or worth pursuing—far from it. The ways we use language typically reveal a great
deal about the ways we think. Thought, however—even if it is collective and
public, as revealed by language use—is never a guarantee of truth. The fact that
we collectively think about something in a certain way does not mean that things
are that way. Systematic errors are not unheard of, even if they are deeply
entrenched in our language use.
None of this is meant to vindicate Dworkin’s “semantic sting” argument. One
main weakness of that argument, as pointed out by several critics,10 consists in
Dworkin’s assumption that criterial semantics, which he attributes to Hart (not
unreasonably, given Wittgenstein’s influence), is incapable of explaining the kind
of “theoretical disagreements” in law that he focuses on. Dworkin’s main assump-
tion, that knowledge of the meaning of a word is necessarily tantamount to
agreement in judgments about the word’s application to all (non-borderline)
cases, is questionable indeed. But this is not the line of response I want to explore
here. In fact, as I indicated above, I share Dworkin’s view that conceptual analysis
is not distinguishable from an analysis of the meaning of words in a given setting.
The question is whether Hart’s theory of law is a form of conceptual analysis at all
and, more generally, whether analytical jurisprudence is best seen as a form of
conceptual analysis. I want to give a negative answer to both of these questions.
9 Hart (1961: 204); see also Marmor (2005: 6–8) and Raz (2009a: 58–66).
10 See e.g. Raz (2009a: 62–76).
214 Andrei Marmor
2. The possibility of reduction
Many courses in analytical jurisprudence begin with Austin’s command theory of

law, and for good pedagogical reasons. Austin’s theory seems very intuitive, almost
compelling at first sight, and thus, by revealing its weakness (usually on the basis of
Hart’s critique), we can motivate a discussion of the hard questions about the
nature of law, opening the way to more sophisticated philosophical accounts
presented to us by philosophers of the twentieth century. But there is a pervasive
danger in this pedagogical approach: By systematically demolishing Austin’s con-
struction, we may be throwing out the baby with the bathwater. And this baby is
the methodological aspiration that motivated Austin: to explain the law in terms of
something else, more foundational in nature. In short, we recognize that Austin’s
theory of law purported to provide a reduction of law to facts of a social kind,
explicable in sociological terms. And this reduction strikes us as crude and ultim-
ately unsuccessful. Many think that Hart must have given up on this reductionist
project, since he demonstrated many of its failures. But this assumption is seriously
misguided. Hart’s own theory of law is as reductionist as Austin’s. Hart clearly
shared Austin’s view that a theoretical explanation of the nature of law should
explain what the law is in terms of social facts, facts that can be explained by more
foundational truths about how people behave, the kind of beliefs they have about
their conduct, and the kind of attitudes and dispositions that tend to accompany
those shared beliefs. In other words, the hallmark of Hart’s theory is the idea that
social rules are at the foundations of law, and that social rules, in turn, can be
explained reductively in terms of people’s actual conduct, beliefs, and attitudes.
All of this requires careful demonstration, of course, but since I have done this
elsewhere,11 I will not try to reiterate the whole argument here—with one excep-
tion that will prove useful for the argument in the sequel. I venture to guess that
Hart’s reductionist project has been blurred by his own cryptic remarks on the
importance of the internal point of view. The standard lesson that people draw from
his remarks is the exact opposite of what Hart actually says in those pages.12
According to the common misunderstanding, Hart’s suggestion that an adequate
account of the normativity of law must explicate the ways in which participants in
the practice regard legal requirements as reasons for their action, is taken to indicate
that Hart shared Kelsen’s anti-reductionist critique of Austin. Hart must have
meant, people assume, that there is something misguided about an explanation
of law’s normativity in purely sociological terms. But in fact, it is the other way
around. Hart’s discussion of the internal point of view is actually a critique of
Kelsen, not so much of Austin. There was not much of a need, after Kelsen, to
11 The first three chapters of Marmor 2011 are devoted to articulating the reductionist project that
was in dispute between Austin, Kelsen, and Hart. Nevertheless, a reviewer (see Patterson’s review in
NDPR 07/17/11) complained that my book presupposes the soundness of conceptual analysis and
invited me to say more in defense of this method; as I try to explain in this chapter, this is an invitation
I have to decline.
12 Hart (1961: 87).
remind us that an account of law as a normative system must include the recogni-
tion that the law is taken to constitute reasons for action for those who regard it as
binding. To regard a norm as binding is to see it as providing reasons for action.
The point Hart makes about the internal point of view is that it can be explained
reductively, by observing people’s actual practices, beliefs, and attitudes. There is
no need, he claimed, to postulate a presupposition of a basic norm, as Kelsen
argued; what we need is only to recognize that, where there is a functioning legal
system in place, most key participants regard legal norms as binding—which is to
say, as something that gives them reasons for action. And, crucially, we can do all
this by reporting on the kind of beliefs and attitudes that the insiders share. We do
not have to endorse their point of view, or talk about it as if we do, only to recognize
that a social rule exists iff the relevant population shares certain beliefs and
attitudes:
For the observer may, without accepting the rules himself, assert that the group accepts the
rules, and thus may from the outside refer to the way in which they are concerned with them
from the internal point of view.13
And perhaps this is even clearer when Hart explains the difference between the rules
of recognition and Kelsen’s basic norm:
First, a person who seriously asserts the validity of some given rule of law . . . himself makes
use of a rule of recognition which he accepts as appropriate for identifying the law.
Secondly . . . If the truth of his presupposition were doubted, it could be established by
reference to actual practice: to the way in which courts identify what is to count as law, and
to the general acceptance or acquiescence in these identifications.14
In short, Hart’s disagreement with Kelsen is precisely about the possibility of
reduction. They both shared the view that legality can only be explained on the
basis of some normative framework that grants certain actions and events in the
world the legal significance that they have. While Kelsen thought that this norma-
tive framework must, ultimately, be postulated or presupposed, Hart strove to show
that there is no need for such a presupposition; the norms that confer legal
significance on acts and events are social norms that can be “established by reference
to actual practice,” as he put it.15 And by practice Hart means the actual patterns of
conduct people exhibit, the kind of beliefs they share about it, and the attitudes that
accompany those beliefs.
Let us get back on track now. If, as I claim, the main methodological thrust
of Hart’s theory of law, like that of Austin’s, consists in a reductionist project,
where does that leave conceptual analysis? Is there a sense in which reduction of
some type of discourse or class of statements to another type or class is a form
of conceptual analysis? That depends on the kind of reduction in play. In some
13 Hart (1961: 87). 14 Hart (1961: 105).

15 In Marmor (2011: ch. 1), I argue that Kelsen’s anti-reductionist argument fails on its own terms;
Kelsen’s argument about the basic norm cannot withstand the pressure for a reductionist account, one
that Hart provided a few years later.
216 Andrei Marmor
contexts, philosophers strive to offer a semantic reduction of one type of discourse to
another. A semantic reduction would have to satisfy the condition that the entire
vocabulary of a given type of discourse or theory, say, D1, can be fully expressed in
terms of the axioms and vocabulary of a different type or class of statements, say
D2. If this condition can be met, we have a semantic reduction of D1 to D2.
I am not sure that there is a very good example of a successful semantic reduction in
any domain,16 but I am willing to assume that, if such a reduction is possible, it
might constitute a form of analysis that is, essentially, about language. We could
then say that the concepts ordinarily used in one type of discourse actually mean
something else, fully explicable in the terms of some other type of discourse or
semantics.17
The type of reduction discussed in legal philosophy, however, is not of this
semantic kind.18 Rather, it is a metaphysical or constitutive form of reduction. The
idea of a metaphysical reduction is to show that a distinct type of phenomenon is
actually constituted by, and fully reducible to, some other, more foundational type
of phenomenon. In our case, the idea is to show that law is constituted by social
practices that can be fully explained by the way people actually behave, the kind of
beliefs they share about their behavior, and the attitudes and dispositions that they
exhibit in the relevant contexts. In short, we try to reduce one type of facts to some
other, more foundational type of facts.
Constitutive or metaphysical reduction is not about concepts. It is about the
nature of things—that is, about the actual properties of objects or phenomena and
their constitutive elements. Consider, for example, the debates about the reduction
of biology to chemistry and physics. The question here is whether we can fully
explain the phenomena manifest in the biological sphere in terms of chemical
reactions and physical properties. This is not a philosophical debate about the
meaning of concepts deployed in the biological sciences. The success of a reduc-
tionist explanation is not determined by how successful the theory is in explaining
our use of language. The criteria of success are determined by the question of how
complete an explanation of biological phenomena in terms of chemical and physical
theories can be. The explanation we seek is of the observable phenomena, not of the
use of concepts in a given setting.
One might object to the example: Arguably, most concepts employed in biology
are of the natural kind—that is, they are concepts purporting to refer to some set of
objects or phenomena, whatever their real nature turns out to be. So there is a sense
in which the use of such concepts already assumes, as it were, that our grasp of their
reference might be partial and incomplete. I certainly do not assume that law is a
16 Some versions of expressivism in metaethics come very close to attempting a semantic reduction
of ethical discourse to expressions of desires, preferences, and the like. How successful this project is,
needless to say, remains highly controversial.
17 I am not suggesting that this would be a form of conceptual analysis in the traditional sense, but
somewhere in the neighborhood.

18 It is possible to construe Austin’s legal theory as attempting a semantic reduction as well. If this is
the case, then Hart certainly did not share the objective; nothing in Hart’s work suggests that he strove
to provide a semantic reduction or that he thought such a project to be feasible.
natural kind of concept. But it still remains the case that an explanation of law is an
explanation of observable phenomena, even if the phenomena are of a social kind,
constituted by the collective behavior, shared beliefs, and social interactions of
human beings. Hart’s reductive theory of law is not an attempt to explain how
people use the concept of law; it is an attempt to explain what law is in terms of
observable social behavior of human beings. More specifically, it is an attempt to
explain what constitutes legal practices and institutions, what makes it the case that
people regard some such practices as legal while others they do not. The fact that
the reference of “law” is a social construction does not entail that a theory of this
reference is a theory about concepts.19
None of this means, of course, that the social nature of the object of a reductive
theory does not impose some constraints on what would count as a successful
explanation. In particular, we must examine the question of how people’s self-
understandings of the social practices they engage in are part of what needs to be
explained, and what kind of constraints this imposes on the nature of the relevant
explanations. I will get to this indirectly, however, first considering the challenge of
interpretivism.
3. The allure of interpretivism
Presumably, nobody denies Raz’s suggestion that “In large measure what we study
when we study the nature of law is the nature of our own self-understandings. . . . it
is part of the self-consciousness of our society to see certain institutions as legal. And
that consciousness is part of what we study when we inquire into the nature of
law.”20 Yet some philosophers argue that it is precisely this truism that should lead
us to realize that an inquiry into the nature of law is essentially interpretative in
nature. And, they argue, interpretation of this collective self-consciousness is partly,
but essentially, a matter of making moral sense of the concepts and ideas we
associate with the practice. Interpretation, in other words, is partly a matter of
evaluative judgments, a matter of figuring out the best moral justifications we have
for the practices in question.
The plausibility of interpretivism in legal philosophy comes from two main
sources. One is the assumption, sometimes explicitly stated, that the alternative to
it, or rather, its object of critique, is conceptual analysis.21 The second consider-
ation lending support to interpretivism comes from analogies to other areas in
which it seems rather compelling, such as self-understandings in the realm of the
19 By way of illustration, consider a Marxist explanation of religion. Marxism clearly assumes that
religion is a social construct, if anything is. But surely a Marxist explanation of religion is not a
conceptual analysis; it purports to explain the relevant social constructions reductively in terms of the
material forces and means by which class struggles generate false consciousness, etc.
20 Raz (2009a: 31). Needless to say, philosophers may disagree about the question of how robust
“our” self-understandings are, and how much of a pre-theoretical consensus or agreement in judgments
we can assume here. I will dwell on this in greater detail below.
21 See Dworkin (1986); Stavropoulos (2012); and Greenberg (2011).
218 Andrei Marmor
arts, or various social practices, such as making promises or conventions of civility.
In these areas it seems clearer, perhaps, that a philosophical explanation of what the
practices are is very closely tied to the kind of values we find in them, and the ways
in which we articulate what makes them valuable and worth paying attention to.
There is, however, a subtle difference between two lines of thought here. According
to one line of thought, the need for an interpretative theory comes from the limits
of conceptual analysis—that is, from the dead end we reach when we focus our
attention on the non-normative facts allegedly constituting legal practice.
According to a different line of thought, the debate between traditional legal
philosophy and interpretivism is about the order of explanation. Instead of begin-
ning with our shared conceptual understandings of what law is, based on the kind
of claims law allegedly makes on us, and then perhaps asking a moral question
about the legitimacy of those claims, interpretivism asks us to reverse the order; it
poses the moral question first, asking what makes certain kinds of institutional
demands legitimate, and then it purports to extrapolate some constraints on what
those institutional or political institutions have to be, and what kind of claims
they can make, and how, in order to satisfy the favored conception of moral
legitimacy.22
One of the main problems with interpretivism is that it takes conceptual analysis
to be its main rival, its main focus of criticism. As I argue, however, conceptual
analysis is really not so central to the mainstream philosophical enterprise of
analytical legal philosophy, or to the “orthodox” view (or “standard picture”), as
intepretivists typically label the main traditions of twentieth-century legal philoso-
phy, particularly the legal positivist tradition. The orthodox view is mostly about
the possibility of reduction.23 And the type of reduction that is sought by the
orthodox view is not about concepts; it is about observable phenomena, including,
of course, social phenomena exhibited by the collective actions and shared beliefs of
a population. So the question we need to examine here is whether interpretivism is
the kind of view that can engage with, and form a genuine critique of, the
reductionism that forms the main object of traditional legal positivism. My answer
is going to be that interpretivism engages with this methodological debate if, and
only to the extent that, it denies the possibility of reduction.
Let me start with the question about the order of explanation.24 There is, of
course, nothing to prevent anyone from starting with any kind of philosophical
question they like. If you want to start with a moral question about the legitimacy
of state coercion or the legitimacy of political authority or whatever, that is fine.
22 I am not suggesting that different philosophers adhere to one or the other of these two versions of
interpretivism; on the contrary, Dworkin, Stavropoulos, and Greenberg (if I understand him correctly)
advance both of these versions concomitantly, shifting from one to the other as if they are one and the
same, or at least very closely linked. At least Stavropoulos makes it quite clear that he does not assume
that the demands in question are necessarily institutional or somehow pre-theoretically “legal” in any
sense.
23 Scott Shapiro’s more recent “planning theory of law” is also quite clearly a reductionist project,
aiming to reduce law to activities of planning, and certain types of collective intentions (based on
Bratman’s theory that is, itself, reductionist in nature). See Shapiro (2011).
24 This argument is most clearly stated in Stavropoulos (2012).
It has been done for centuries and is mostly referred to as political philosophy. The
question is, of course, whether anything about the nature of legal philosophy forces
this order of explanation on us. Interpretivism is committed to the thesis that,
indeed, something forces on us the primacy of the moral question. The order of
explanation is forced on us, the argument seems to be, because the opposite order
takes too much for granted.25 If you start with an observation about the non-
normative facts allegedly constituting the practice you strive to explain, you
necessarily privilege certain prevailing concepts and self-understandings as defining
the very subject matter of what you strive to explain. In other words, even if the
method of legal philosophy is not necessarily conceptual analysis, you must start
with some concepts that define the subject matter, concepts, or conceptions that
determine what is in need of explanation.26 But, interpretivism tells us, nothing
about the prevailing concepts of a practice or an institution really justifies this
privilege of determining what is the true nature of the subject matter that calls for
theorizing. Concepts, prevailing and entrenched as they may be, can get things
totally wrong, even if they are not contested at the time. Just as astronomy would be
wrongheaded to take for granted that stars are holes in the sky, as was once widely
believed, and take it from there, so legal philosophy would be wrongheaded to
assume that law is, say, what political authorities decree, and just take it from there.
The analogy is misleading, of course; we can start with a theory that stars are
holes in the sky and see how far it takes us. As it turns out, it did not take us far
enough, so we had to revise the assumption about what stars are in light of the
relevant evidence. Philosophy, however, does not have the luxury of the scientific
methods allowing hypotheses to be tested empirically.27 So the danger is that if we
take concepts for granted, we might be stuck with trying to explain how stars are
holes in the sky.
If we do not want to make this argument into an indictment of philosophy as
such—which would be self-defeating, of course—we must limit the concern to the
place where it belongs. In other words, I doubt that interpretivism purports to
challenge the very possibility of any kind of non-normative philosophy. But then, if
non-normative philosophy is possible, what makes philosophy of law different? The
answer might be that law is different from other philosophical subjects because it is,
in itself, a normative practice—that is, the kind of practice that makes moral
demands on us. But at this point we are already beginning to lose sight of the
argument. The indictment of orthodoxy is based on the claim that it is bound to
privilege concepts and self-understandings of law in a way that predetermines the
issues; we cannot assume, this argument goes, that our concepts, and the ways in
26 Raz comes very close to admitting as much when he claims that what we strive to explain in legal
philosophy is what the concept of law, as we now have it in contemporary Western societies, stands for,
even if it is also true that, in different societies and at different times, people entertained rather different
concepts of law (or none) (2009a: 40).
27 Brian Leiter, and others who share an empiricist view of philosophy, would disagree. This is not
the place, however, to discuss the controversial idea of “naturalizing” jurisprudence and other forms of
philosophical empiricism.
220 Andrei Marmor
which legal practice is understood by those whose practice it, correctly identify
certain central features of law as the appropriate objects of philosophical inquiry.
At best, such self-understandings form a possible hypothesis that competes with
others we can come up with—but a hypothesis of what? If we do not make any
assumptions about the object of our inquiry, hardly any philosophical explanation
can get off the ground. The answer we are given is that the alternative to privileging
prevailing concepts is to take certain paradigmatic examples as tentatively defining
the subject matter and potentially revising those paradigms as we go along.28
The main problem here is that the philosophical method that interpretivism
recommends is vulnerable to the same problem that it accuses orthodoxy of. If your
starting point is a moral question about the law, and then you assume that the
important aspects of law are those that figure in, or follow from, your moral answer,
you may have also assumed too much. Consider, for example, the moral question
that Dworkin (and, it seems, Stavropoulos) takes to be central—namely, the
question about the justification of the use of collective force or state coercion.
Quite clearly, the question itself presupposes that coercion is an essential aspect of
law; if it is not, why focus on coercion? It may turn out that many of law’s functions
in society have very little to do with the use of force or coercion. Are those non-
coercive functions not essential or not characteristic of law? Interpretative, partly
moral, questions about the law are just as much about the law as any non-normative
questions. They must start somewhere, and not just anywhere; they must start with
some understanding about what law is and which aspects of it are more central to it
than others. (I will return to this point later.)
Now, true enough, Dworkin sometimes gives the impression that he actually
cares little about the question of whether his interpretative-normative theory
captures some pre-theoretical essential aspects of law or not. But then we are
back to the problem that not just about any normative question in political
philosophy has something interesting to tell us about the nature of law. In other
words, unless it is assumed that coercive force (or whatever else one deems to be the
focus of the relevant moral question) is an essential or central aspect of law, the
theory of legitimacy offered would not necessarily have any bearing on the nature of
law. Political philosophy is not legal philosophy by default.
At this point, the second line of thought that leads to interpretivism may be
brought in. What is special about the law, interpretivists seem to claim, is its very
normative nature, the fact that it makes certain normative, including moral,
demands on us, and that we regard at least some of these demands as morally
warranted or binding. By itself, however, this observation, sound as it is, does not
lead to the conclusion that any philosophical theory about a normative practice
must begin with a moral question, or that it must include moral arguments as part
of the theory. The argument has to be more subtle than this. Perhaps something
like the following analogy might do the work: Consider, for example, the realm of
art. We regard certain human creations as works of art, and countless others as
28 See Dworkin (1986: 91–2) and Stavropoulos (2012).

having no artistic merit whatsoever. So we can ask: What makes anything a work of
art? Quite plausibly, we cannot give an answer to this question without having
some views about where value lies in art, or what makes art have the kind of values
we associate with it. Interpretivism in the realm of art, or philosophy of art, if
you like, seems rather compelling. Why would it not be equally compelling in
philosophy of law?
There are several reasons to doubt the analogy. First, interpretation is central to
art because works of art are created as objects of interpretation; they are created
(partly) as an invitation to appreciate their aesthetic and artistic features, to
appreciate the kinds of achievement they manifest, and the like. Law, however, is
not created to become an object of interpretation—law is created to guide human
conduct; it has practical purposes. Making a law is not an invitation to the public to
offer interpretations of it or to appreciate its legal qualities. Second, there is a clear
sense in which art is an essentially contested concept. That is, the word “art” stands
for a certain form of human achievement, an aspiration for excellence that is
essentially contested, inviting different views about what kind of achievement it
is and what its standards of excellence are. Each work of art contributes to this
cultural debate, as it were, making an implicit statement about its conception of art,
what the creator values in it, etc. And that is why the question of whether a given
artifact is a work of art or not, and the question of what we value about art, are very
closely linked.
Legality, however, is not an essentially contested concept. We do not regard
legality as a form of human achievement, inviting, as it were, different conceptions
of what makes it an achievement, what it is an achievement of and the standards of
excellence we associate with it.29 The making of good law—morally, politically,
economically, or otherwise—is of course a form of achievement. We may have
different conceptions of what would be a good law in this or that domain, but the
relevant issue that is essentially contested here is not the legality of the law-making
but the moral or other evaluative dimension of it—that is, the contested element
here is the “good,” not the “law.”30
None of this shows interpretivism to be misguided as a philosophical method. It
only shows that we have not yet found a way in which interpretivism actually
engages with, not to say refutes, the reductionist project that forms part of the
“orthodox” view. It seems, so far, just an alternative philosophical interest in law, or
something in that neighborhood, guided by moral concerns about the potential
impact of certain types of demands on our reasons for action. In order to engage
with the philosophical method of the orthodox view, interpretivism has to argue
that reduction of law and legality to some social facts is not attainable. In other
words, interpretivism is a challenge to legal positivism iff it amounts to the denial of
29 The rule of law can be seen as a form of achievement that clearly admits of degrees of success or
failure. It may well be an achievement of a given society and its government that it adheres to the rule of
law, and societies or regimes achieve these goals to a greater or lesser degree. But this is an entirely
different matter. I have elaborated on this distinction in Marmor (2011: 132–4).
30 This is not meant to be an argument against a teleological conception of law, such as John
Finnis’s, maintaining that law is essentially concerned with the common good.
222 Andrei Marmor
the possibility of reduction. I think that it may amount to that, but I will explain
this a bit later. First, I want to consider the challenge posed by interpretivism to the
kind of essentialism that has informed contemporary legal positivism.
4. Essentialism in perspective
You cannot offer a reduction of a certain type or class of phenomena to some other
type or class of facts unless you have a fairly robust sense of what constitutes the
main features of the reduced class. Furthermore, the success of a reductionist theory
depends on how complete the explanation it offers is, and we cannot have any sense
of that without knowing what needs to be explained. If we want to offer a
reductionist theory of law, we must have a pretty good idea of what is essential to
law and what is only incidental and contingent. Otherwise, we would lack any
criteria of success for the reduction offered. And this is where, interpretivists claim,
we get into trouble. Assumptions about essential aspects of law are profoundly
suspect, they claim, because they privilege certain conceptual assumptions or self-
understandings of law that might be as erroneous or misguided as the idea that stars
are holes in the sky. In other words, the idea here is that any assumption about the
essential aspects of law is bound to be contestable, morally or otherwise, and hence
cannot form a benchmark against which we can test the plausibility of a reduction-
ist theory.
The debate typically focuses on one prominent example, offered by Joseph Raz,
claiming that it is an essential aspect of law that it purports, and necessarily so, to
claim legitimate authority over its subjects.31 Law’s authoritative nature, Raz
famously argued, is an essential aspect of law, and thus any theory about the nature
of law must give an explanation of how law can be authoritative and how it can
make the kind of claims that potentially meet the conditions of constituting
legitimate authoritative directives. Interpretivism challenges this assumption not
by denying that law is authoritative, but by challenging the essentialism that Raz’s
position assumes here. Simply put, the question is: How can we tell that anything is
essential or necessary to law? If it is a conceptual claim—as most commentators
understand Raz to be assuming—we are back to the difficulties of conceptual
analysis and the unwarranted assumption that prevailing concepts are somehow
privileged in defining the subject matter of the inquiry. And if it is not a conceptual
claim, what kind of claim is it?
This is a fair question. Before I try to answer it, however, let me reiterate a point
I made earlier—that I do not see how interpretivism avoids the same problem.
Suppose you do not make any assumptions about essential aspects of law and
instead ask yourself what justifies the use of collective force, or any other similar
question that puts a moral concern about law at the forefront. We would still need
31 This is not the only essential feature of law that Raz argued for; law also essentially claims
normative superiority to any other normative system, and it claims the authority to regulate any aspect
of human conduct. See Raz (2009b: 149–54).
to know that your moral concern is one about the law, and, more importantly, that
the conclusions you reach are such that they tell us something of interest about the
nature of law. I just do not see how any of this can be done without making certain
assumptions about aspects of law that we deem more essential to it, or more
characteristic of it, than others. What makes Dworkin’s question about the legit-
imacy of the use of collective force more relevant to law, than, say, a question about
the moral legitimacy of progressive taxes? The answer has to be this: We can have
law and a functioning legal system in place without a progressive tax system, while,
presumably, we cannot have law without some coercive element at work. Which
is to say that we have already assumed that coercion is somehow more essential to
law, more central to what law is, than a progressive system of taxation. And I hope it
is clear that the example generalizes to all cases.
Dworkin might respond that the choice of the moral question is itself morally
determined; we need to focus on law’s coercive aspect because it is the morally
pressing issue here (and, of course, it may not be the only moral question in the
neighborhood; others might be just as relevant). That may be true; it is possible that
there are very good moral reasons to be concerned with law’s coercive aspect. But
again, none of this would guarantee that the answer to the moral question you
come up with tells us something about the nature of law. Unless you assume that
coercion is central or essential to what the law is, a moral theory about the
legitimacy of coercion would be just that—a moral theory about the legitimacy
of coercion.
But I do not want to dodge the question: On what grounds can we claim that F is
an essential feature of X? To begin with, it would be a mistake to equate essential
properties of X with a priori necessary conditions. That “F is an essential property of
X” is not necessarily an a priori truth. And it is not always the case that if F is an
essential property of X, then F is a necessary condition for something to be an
X. There are many essential properties of objects or categories that we learn from
experience. We learn from experience, for example, that the ability to use language
is an essential aspect of homo sapiens Linguistic competence is partly, but quite
essentially, what defines humans as a distinct biological species. This is hardly an a
priori truth, though. In fact, it rests on an empirical generalization that could turn
out to be inaccurate.32 Second, to continue with the example, when we say that
linguistic competence is an essential feature of humans, we are not committed to
the idea that a person with serious brain damage that causes her to lose the ability to
use language renders that person non-human. Essential features are not necessary
conditions. To take another example, it makes sense to suggest that “manufactured
for humans to sit on” is an essential feature of chairs; it is what chairs are. Yet a chair
designed for an art exhibit, not made for people to sit on, might still be a chair.
32 We may come to learn, for example, that dolphins’ linguistic competence is much closer to
humans’ than we think—perhaps even superior to ours! And then, of course, we may need to revise our
views about how distinct a characteristic linguistic competence is, and how it defines our categorization
of species. But it would not undermine the claim that linguistic competence is essential to humans; it
may also be essential to dolphins.
224 Andrei Marmor
Essential features of an X might be necessary conditions for central cases to be an X,
but what counts as central cases admits of some vagueness, with borderline cases
along the way to marginal and deviant cases.
Be this as it may, what constitutes the idea of an essential property of something
is bound to vary between different domains or realms of inquiry. Essential features
in number theory are probably different from essential features in biology. Let us
focus on essential properties of social practices. (Now you know what is coming.)
We can say that it is an essential aspect of chess that it is a competitive game.
Perhaps we need to qualify this statement: It is an essential aspect of chess, as we
know it, that it is a competitive game. For all we know, there might be Martians
who “play” chess not as a competitive game but as a religious ritual; perhaps the
winner becomes the high priest.33 So it is not really a game on Mars. Whether
Mars-chess is, in some sense, chess or not, is a difficult question to answer, but it is
not what is needed here. What is needed is to realize that the characterization of
chess as a competitive game is not a statement about concepts or, generally, about
language. It is a claim about the nature of a certain type of human activity. And
whatever we know about chess is not something we learn by mastering a language,
but by observation, learning, and participation in this and similar activities. If you
do not know, for example, that chess is a competitive game, that players must aim to
win the game, and that what counts as winning is determined by the rules, you will
not be able to play chess, nor will you understand what is going on when observing
others playing it. True enough, you will also not know what the word “chess”
means, thus we can say that you do not quite have the concept of chess. But
learning that chess is a competitive game is not something that comes about by
learning to speak English; it comes about by learning about chess, what it is and
how to play it (or, at least, how to understand what others are doing and aiming for
when playing it).
Admittedly, I have only responded to half of the challenge; I suggested that
essential properties of social practices are properties or features of observable and
learnable activities, not of language or concepts. But I have not said anything about
what makes a feature essential and how we can come to know it. I think we can
safely say this: A feature of a social practice is essential to it if without it the practice
would either not have existed at all or would have been radically different from what
it is. At least, this is how we normally go about trying to determine whether a
suggested feature is essential or not. If you suggest that some feature F is essential to
law, we should ask whether a practice that is similar in all other respects, but lacks F,
is still close enough to what we recognize as law in our society. This would easily
show, for example, that progressive taxation is not essential to law; we would clearly
recognize as a legal system one that has a flat tax rate, and even one that imposes no
taxes at all. Some cases are harder to determine, of course. The question of whether
coercion is essential to law is not so easy to answer. Thinking about a practice or
institution that would be similar to law in all other respects, but for the fact that
33 The example is from Schwyzer (1969).

there is no coercion involved whatsoever, might stretch our imagination beyond

our ability to come up with a determinate answer. It may be so, although I am not
suggesting that this is the case. If it is, we may not have a determinate answer to the
question of whether coercion is an essential aspect of law. Since essential features are
neither a priori nor necessary, the possibility of indeterminacy about some cases is
not philosophically problematic. I see no reason to deny that what is essential to a
social practice is vague, allowing for some borderline cases.
Now let us assume that Raz’s thesis that law is essentially an authoritative
institution is on a par with the suggestion that chess is a competitive game. First,
we can conclude that this is not a conceptual claim, at least not in the sense that it is
a proposition about concepts. It is a proposition about an essential feature of a
certain type of human activity, a social practice. Second, we can have a sense of how
to go about determining whether Raz’s suggestion is true or not: Imagine a practice
or institution that is somehow similar to law in all other respects but for the fact
that it makes no claim to be authoritative. How close would it be to what we regard
as legal practice in our world? I hope the answer is evident, but that is not what
I need to substantiate here. Finally, it does not follow from the essential authorita-
tive nature of law that each and every law must make a claim to be authoritative;
some deviant cases are possible.34
Authority was just an example; other essential features of law have been identi-
fied by Hart (that modern legal systems combine primary and secondary rules), by
Kelsen (that laws must come in systems, with some hierarchical structure) and, if
I may, by myself (that law has conventional foundations). And these features, and
others, are not mutually exclusive, of course. Needless to say, the list of features that
philosophers identify as essential to law is not randomly selected. The features serve
some theoretical purposes, highlighting an aspect of law that calls for theoretical
explanation and has some significance for the kind of theory offered. But there is
nothing question-begging about that. Theoretical questions and interests always
come up against the background of previous theories and the difficulties they give
rise to; our interest in particular questions and the ways in which we phrase them is
path dependent, shaped by traditions constituting the relevant discipline. This is
quite universal and applies to moral theories as well.
5. The limits of reductionism
I mentioned earlier that interpretivism can critically engage with what it calls the
orthodox view iff it denies the possibility of reduction. In his earlier criticism of
Hart’s jurisprudence, this is precisely the path that Dworkin followed. Though not
phrased explicitly as an anti-reductionist critique, Dworkin’s argument about legal
principles aimed to show that Hart’s rules of recognition cannot provide a complete
34 One example I have in mind are cases where legislatures enact laws that have no prescriptive
action-guiding content. In the US, for example, states often legislate that X is the “state bird” or
something like that, which is a symbolic act without any authoritative directive included.
226 Andrei Marmor
answer to what constitutes legal validity, because legal validity is partly deduced by
moral reasoning. This is quite clearly an argument that purports to show that
legality cannot be reduced to social facts.35 Unfortunately, the interpretivist turn in
Dworkin’s theory obscured this anti-reductionist argument, and led many to
suspect, including Hart himself, that interpretivism and legal positivism talk past
each other. Legal positivism, as Hart clearly saw it, aims to provide the essential
features of law quite universally—that is, wherever law exists—whereas the impres-
sion that Dworkin left about interpretivism was that it articulates a theory of
adjudication for Anglo-American law.36 I think that Hart was not quite right
about this. There is a clear sense in which Dworkin’s interpretative theory chal-
lenges the reductionism inherent in Hart’s legal positivism. It aims to show that
legality cannot be reduced to facts about people’s conduct, beliefs, and attitudes
because legality is partly, but necessarily, constituted by truths about moral issues.
If truths about morality necessarily figure in the determination of what counts as a
legal requirement, then it is clearly not the case that legality admits of reduction to
social facts. And, of course, Dworkin does not need to show that this is the case in
every legal system we are familiar with; in order to refute the possibility of reduction
it is sufficient to show that reduction is not possible at least in some cases. Since
reductionism is a universal claim, it can be refuted by counterexamples.
This is not the place to evaluate Dworkin’s anti-reductionist argument; I have
done that elsewhere.37 I want to focus, instead, on a question we have left open so
far, about the kind of constraints imposed on any theory about the nature of law
stemming from the fact that part of what we aim to explain is people’s self-
understandings of what law is. Dworkin maintains that a theory about the nature
of law must provide the kind of explanation that participants in the practice can
recognize, at least upon reflection, as something that rationalizes for them the ways
in which they understand the practice and its normative requirements. And he
clearly thinks that this constraint is general, one that applies to philosophical
explanations, viz., constructive interpretations, of social practices in general. Call
this the rationalization constraint. Now, of course, if you hold the view that any
philosophical explanation of a social practice is necessarily a form of constructive
interpretation, the rationalization constraint makes a lot of sense. Constructive
interpretation, as Dworkin understands this idea, is the kind of explanation that is
aimed at those whose practice you interpret; it aims to make sense for the
participants what they should regard as valuable about the practice and how they
should rationalize their own understanding of it.
A reductionist account of a social practice, however, is not committed to the
rationalization constraint. In some clear cases, reductionism in social explanations
clearly violates this constraint. A Marxist explanation of religion, for example,
cannot offer the kind of explanation that religious people could come to recognize
as their own, so to speak, at least not as long as they still deem themselves religious.
35 See Dworkin (1977); I explain this in greater detail in Marmor (2011: ch. 4).
36 See Hart (1994). 37 Marmor (2011: ch. 4).
In other words, a reductionist explanation of a social practice is not expected to

provide a kind of “aha!” moment for the subjects, providing them with an insight
that would rationalize the practice for them.
This does not mean, however, that the social nature of the object of explanation
of a reductive theory does not impose some constraints on the kind of explanations
we can offer. It is difficult, if not impossible, to explain anything in the social realm
without invoking people’s beliefs (and attitudes shaped by those beliefs, etc.). Yet
most (not to say all) of our beliefs are parasitic on natural language; there are
countless beliefs that people cannot have unless they speak a natural language.
Semantics, and perhaps more generally, linguistic competence, play a crucial role in
belief formation, the individuation of the content of beliefs, and so on and so forth.
Therefore, social explanations that rely on the articulation of people’s beliefs,
attitudes, and the like, are inevitably constrained by the constitutive role that
language plays in making those beliefs and attitudes possible and shaping their
content.
Let me demonstrate with an example from Hart’s account of social rules. For a
social rule to exist, Hart maintained, certain things must be true: It must be the case
that there is some regularity of behavior manifest in the overt conduct of a certain
population, and this conduct must be accompanied by certain types of beliefs and
attitudes that most individuals in that population share about it. For this kind of
reduction to be successful, it has to be shown that the explanation is complete, and
that the elements offered are sufficient to distinguish social rules from other, similar
type of phenomena.
Critics have long argued, however, that Hart’s explanation is not complete.38 In
particular, it cannot distinguish between a case of following a social rule and one of
complying with a generally or widely accepted reason. People often exhibit a
regularity of behavior because they believe that there are some reasons that apply
to the circumstances at hand, and thus they comply with the reason and exhibit the
same kind of attitudes that Hart associates with the “acceptance” of a rule. But,
nevertheless, it is not a rule they follow. The textbook example given here concerns
strategies followed in games: Players are often expected to know, as they do, that
there are certain things not done in a game, for example, because it would be a bad
strategy. So, by and large, they do not do them, and would criticize others who do,
etc. However, such widely recognized strategies do not form part of the rules of the
game. If you fail to comply, you might be criticized for being foolish or amateurish,
but not for violating a rule.39
What Hart may have missed here (perhaps, I am not claiming that he did) is the
fact that following a rule requires the attribution of a reflexive type of belief to those
whose rule it is. The difference between complying with a widely believed strategy
and complying with a rule consists in the fact that participants regard the existence
of the rule as a reason for their action—which is to say that we cannot account for
38 Myself included, see Marmor (2001: 2–4), and reference there.

39 The game example is taken from Warnock (1971: 43–6).
228 Andrei Marmor
the relevant type of belief people have without invoking the very concept that we
try to explain.
Is this as serious a worry as it seems? Let’s see where we stand. To account for the
existence of a social rule, we must attribute certain beliefs to those whose rule it is.
One of those beliefs must be a belief in the existence of a rule, which means that
participants must understand the concept of a rule to have the relevant kind of
belief. As I argued earlier, however, to understand the concept of a rule is not
different from knowing what the word (in the relevant context) means and how to
use it correctly. Now, this would be a serious problem for Hart’s account if we
thought that his theory about the nature of social rules was meant to be a
conceptual definition. Then we would have to admit that he was trying to define
what the concept of a rule is in a way that presupposes that people understand the
concept. But I do not quite see it that way. What Hart was trying to do is to provide
an account of social rules, not a conceptual definition of what rules are. This
account purports to identify the kind of beliefs and attitudes people must have
when following a social rule. One of those beliefs may well require that participants
know, at least roughly, what the word “rule” means in their natural language. In
other words, I do not think that it was part of Hart’s project to provide a conceptual
analysis of “rules,” or “social rules” for that matter. The project was to show what
facts constitute the existence of a social rule in a given population, and to show that
we do not need to have more in our toolbox for this purpose than the elements he
identified, namely, regularities of behavior coupled with certain types of beliefs and
attitudes. The fact that beliefs and attitudes depend on natural language may well
show that the kind of reductive explanation Hart offered cannot go all the way
down; at some point it hits the bedrock of semantic competence shaped by the
natural language we speak.
It is crucial to note, however, that the semantic competence we must assume here
does not commit us to agree with the views, no doubt often partial and incomplete,
that people have about the nature of the things the relevant words stand for. People
who have reasonable knowledge of English use the word “rule” correctly on the
majority of occasions, though few would be able to define with any precision what
the word means or stands for; even philosophers find that task difficult. Further-
more, we often use words correctly even if we know (or believe, if you prefer) that
the word stands for something that does not exist, or that it stands for a confused
idea. (I find the idea of God profoundly confused and incoherent, but I am
confident that I know what the word means and I can certainly use it correctly.
You may have other examples in mind.)
In other words, the constraints imposed by the need to rely on semantics do not
entail that we have to take all that much for granted. It only means that we must be
aware of the fact that the semantics of a natural language play a very important role
in belief formation. Language, in other words, imposes certain constraints on social
explanations, but it is not the main object of them. It certainly does not mean that
philosophical problems are conceptual problems, or that conceptual analysis forms
the relevant methodological foundations. Every reductive explanation presupposes
that the class reduced to is, in some relevant sense, more foundational than the
reduced class; but it does not have to assume that it is as foundational as it gets.
References
Dworkin, R. M. (1977). “The Model of Rules I,” in Taking Rights Seriously. London:
Duckworth, ch. 2.
Dworkin, R. M. (1986). Law’s Empire. Cambridge, MA: Harvard University Press.
Greenberg, M. (2011). “The Standard Picture and its Discontents,” in L. Green and
B. Leiter (eds.), Oxford Studies in Philosophy of Law. Oxford: Oxford University Press.
Hart, H. L. A. (1961). The Concept of Law. Oxford: Oxford University Press.
Hart, H. L. A. (1994). The Concept of Law: Postscript in second edition. Oxford: Oxford
University Press.
Marmor, A. (2001). Positive Law & Objective Value. Oxford: Oxford University Press.
Marmor, A. (2005). Interpretation and Legal Theory. Revised second edition. Oxford: Hart
Publishing.
Marmor, A. (2009). Social Conventions: From Language to Law. Princeton: Princeton
University Press.
Marmor, A. (2011). Philosophy of Law. Princeton: Princeton University Press.
Raz, J. (2009a). Between Authority and Interpretation. Oxford: Oxford University Press.
Raz, J. (2009b). Practical Reason and Norms. Second edition. Princeton: Princeton
University Press.
Ryle, G. (1954). Dilemmas. Cambridge: Cambridge University Press.
Schwyzer, H. (1969). “Rules and Practices,” Philosophical Review, 78: 451.
Soames, S. (2010). Philosophy of Language. Princeton: Princeton University Press.
Stavropoulos N. (2012). “Obligations, Interpretivism, and the Legal Point of View,” in
A. Marmor (ed.), Routledge Companion to Philosophy of Law, 76–92.
Warnock, G. (1971). The Object of Morality. London: Methuen, 43–6.
10
What Do We Want Law to Be?
Philosophical Analysis and the
Concept of Law*
Natalie Stoljar
It is a mistake . . . for those engaged in conceptual analysis to dismiss inquiries

into operative and target concepts, with the thought that only the conceptual
project can discover “our” concept. For example, if we discover that we are
tracking something that is worthwhile to track in using our [legal] vocabulary,
then even if this is not what we originally “had in mind,” it still may be what
we have been and should continue to be talking about.1
We must avoid, if we can, the arid wastes of inappropriate definition, for, in
relation to a concept as many-sided and vague as that of a legal system, disputes
about the “essential” character, or necessity to the whole, of any single element
soon begin to look like disputes about whether chess could be “chess” if played
without pawns.2
1. Introduction
Philosophical analysis in answer to a question “What is X?” is often thought of as an a

priori exercise—undertaken from the philosopher’s armchair—in marshaling and
organizing our intuitions about the nature of X or about the necessary and sufficient
conditions for the correct application of the term “X.” Philosophical analysis in
* The title of this chapter expresses my indebtedness to the work of Sally Haslanger, in particular to
her articles “Gender and Race (What) Are They?; (What) Do We Want Them to Be?” (2000) and
“What Are We Talking About? The Semantics and Politics of Social Kinds” (2005). For helpful
comments, I am grateful to Imer Flores, Liam Murphy, Veronica Rodriguez-Blanco, Fabio Shecaira,
Juan Vega Gomez, and especially Stefan Sciaraffa. I am also indebted to comments from audiences at
the Analytic Legal Philosophy Conference, Law School, University of San Diego, April 2011, the
McMaster Conference on The Nature of Law: Contemporary Perspectives, McMaster University, May
2011 and the Problema Seminar, Legal Research Centre, UNAM, Mexico City, May 2012. I am aware
that I have been unable to respond adequately to all the constructive suggestions that I have received.
Finally, thank you to Lyndon Entwistle for his assistance in preparing the final version.
1 Haslanger (2005: 21). 2 Hart (1958: 622).
What Do We Want Law to Be? 231
answer to the question “What is law?” is no exception to this general methodology.

One plausible example of the methodology at work is John Austin’s analysis of law in
The Province of Jurisprudence Determined. He claims to have “endeavored to resolve a
law (taken with the largest signification which can be given to that term properly)
into the necessary and essential elements of which it is composed.”3
This traditional picture of conceptual analysis has held sway in philosophy at
least since Socrates’ investigation of the concept of knowledge in the Theaetetus. It is
the approach to answering “What is X?” questions that many of us were taught in
philosophy classes. It is alive and well in philosophy journals. Of course, sometimes
it does not attempt to resolve a concept into necessary and sufficient conditions
preferring rather to identify a cluster of core features.4 And sometimes it need not
correspond to a wholly a priori project of reflection from a philosopher’s armchair.
The process of discovering the initial data to be subjected to philosophical analysis
may just as easily be undertaken a posteriori by looking in a dictionary or
conducting surveys of what people think they mean by their words or concepts.
Despite the possible variations, traditional conceptual analysis of some form is a
prominent method of answering “What is X?” questions in philosophy.
Contemporary debates about the analysis of the concept of law also accept the
traditional picture. Stephen Perry for instance advocates conceptual analysis for the
term “law,” namely the project of “[making] explicit what the theorist claims is in
some sense already implicit in our common understanding.”5 Those who reject
conceptual analysis also seem to presuppose the traditional picture. Brian Leiter
thinks that the Quinean critique of the analytic-synthetic distinction undermines the
possibility of conceptual analysis.6 Liam Murphy argues that an investigation into
common usage cannot deliver a determinate answer to contested questions about the
content of the concept—such as whether there a necessary connection between law
and morality—and therefore conceptual analysis should be abandoned.7 Even those
proposing that there is a plurality of inquiries accept the traditional picture and claim
that the plurality obtains within it. Oberdiek and Patterson argue that the “mistake
that Quine and, by extension, Leiter make is to treat conceptual analysis as one
particular type of philosophical work, and one, moreover, that depends upon a
robust conception of the analytic/synthetic distinction . . . there are at least four
distinct, overlapping methods of conceptual analysis.”8 The four methods they
identify correspond to four versions of the traditional conceptual mode of inquiry.
One task of this chapter is to suggest that traditional conceptual analysis is still
useful for understanding aspects of the concept of law. The role of the conceptual
3 Austin (1995: 117). Brian Bix points out that John Austin was the first “to approach the theory of
law analytically (as contrasted with approaches to law more grounded in history or sociology, or
arguments about law that were secondary to more general moral and political theories). Analytical
jurisprudence emphasizes the analysis of key concepts, including ‘law,’ ‘(legal) right,’ ‘(legal) duty,’ and
‘legal validity’ ” (2010).
4 For example, see Nussbaum (1995).
5 Perry (2000: 333).
6 For example, see Leiter (2003).
7 For example, see Murphy (2005).
8 Oberdiek and Patterson (2007: 72).
232 Natalie Stoljar
should not be overstated, however. I agree with Oberdiek and Patterson that there
is a plurality of projects within conceptual analysis. However, drawing on recent
work by Sally Haslanger, I argue that only one of these projects corresponds to
traditional conceptual analysis. Haslanger calls the traditional picture a “concep-
tual” inquiry, the aim of which is to uncover our manifest concept of the X in
question—the concept of X that we believe ourselves to have. She identifies two
further projects of conceptual analysis: a descriptive inquiry that investigates the
phenomena in the world that we use our concept to track, with the aim of
providing an analysis of what Haslanger calls an operative concept; and an ameli-
orative inquiry that theorizes about our “legitimate purposes” in using a concept
and yields a target concept.9 For concepts that refer to social kinds or objective
social types, all three projects can be employed to provide a fuller or more resolved
account of the concept. If there is no convergence on the manifest concept, the
theorist might turn to an operative or a target concept. If there is an incompatibility
between the manifest and the operative concepts, a target concept might be
employed to decide between them, and so forth.
A more specific aim of the chapter is to apply this pluralistic methodology to
some of the debates within the literature on methodologies of law. There is a
bewildering array of different positions. They include: the position that the concept
of law is an interpretive or hermeneutic concept and hence that conceptual analysis
of law must employ moral considerations;10 the claim that theorizing about a
hermeneutic concept can be “descriptive” and does not have to employ moral
considerations;11 the argument that conceptual analysis corresponds to an investi-
gation of common usage and, since there is no convergence on common usage,
conceptual analysis is pointless;12 and the position that theorizing about the nature
of law is evaluative at a metatheoretical level but still “descriptive/explanatory” and
non-moral.13 The methodology offered by Haslanger can perhaps shed light on the
logical space of legal methodologies and help us resolve some of the contested
questions.
A final aim is to take a side in the debate about the conceptual analysis of the
concept of law. I take seriously the special features of concepts that refer to social
practices, of which law is one. As Haslanger puts it: “our concepts and our social
practices are deeply intertwined. Concepts not only enable us to describe but also
help structure social practices, and our evolving practices affect our concepts.”14
Although both the conceptual and the descriptive modes of inquiry are legitimate
ways to undertake conceptual analysis, I argue that these approaches are limited in
9 See e.g. Haslanger (2005).

10 For example, see Dworkin (1986) and Perry (2000).
11 Dickson (2001) and Leiter (2003).
12 For example, see Murphy (2008).
13 For example, see Dickson (2001).
14 Haslanger (2005: 13). Frederick Schauer (2005) argues for a parallel position: he emphasizes that
because law is a socially constructed concept that refers to an evolving social kind, conceptual change
and revision are possible, even if it is “slow, incremental and uncertain.” He notes that theorists’
conceptual analysis can potentially effect conceptual change. Schauer calls this approach “prescriptive”
conceptual analysis.
answering questions of importance about the nature of the social type we call “law.”
Answering the question “What is law?” will have to turn to what Haslanger calls the
ameliorative project. We must ask questions such as: What are the legitimate
purposes that we want this concept to serve in our practice? What do we want
the concept to do for us? How do we characterize the concept in order to enable it
to better serve our purposes? Answering these questions will sometimes employ
moral evaluation. It is obvious therefore that I am in broad agreement with the
positions taken by Dworkin, Finnis, and Perry each in a rather different way, and
perhaps by Murphy in yet a different way again, namely that to offer a fully fleshed-
out answer to the question “What is law?” we need to engage in substantive moral
and political argument.
The plan of the chapter is as follows. Section 2 identifies three possible strategies
of conceptual analysis for concepts that refer to social kinds. I point out that there
are exemplars of all three strategies in the literature on legal methodology. Section 3
elaborates two (non-legal) examples of concepts that refer to social kinds to show
the plausibility of the claim that the analysis of such concepts employs a plurality
of strategies. Section 4 applies the first two strategies—the conceptual and the
descriptive—to law and shows how the application of each has helped to illuminate
aspects of the concept of law. As an example of a descriptive approach, I focus on
the descriptive/explanatory theory offered by Julie Dickson and suggest that it
presupposes a descriptive strategy yielding an operative concept. I argue that it
is helpful to characterize H. L. A. Hart’s much-discussed methodology as both
descriptive and genuine conceptual analysis on the model of a descriptive strategy
delivering an operative concept of law. Section 5 defends the position that the
ameliorative strategy is an important component in conceptual analysis of law.
I argue that, despite the logical possibility of applying each of the conceptual and
descriptive strategies to law, as well as their usefulness in answering questions about
certain aspects of the nature of law, ultimately they are limited. When the manifest
and the operative concept conflict, or when the operative concept is itself indeter-
minate, an ameliorative strategy should be invoked.
2. Three projects of conceptual analysis

Haslanger identifies three possible conceptual inquiries. Although the outcomes of
these inquiries may coincide, it is also possible that each will lead respectively to
different concepts of the X under investigation. I briefly show how different legal
theorists have employed each of these different approaches in their accounts of the
methodologies of law. Haslanger’s taxonomy is useful therefore to illuminate the
logical space of legal methodologies.
2.1. The conceptual project

How do we answer questions of the form “What is X?:” “What is knowledge?”
“What is gender?” “What is democracy?” “What is marriage?”? The traditional
234 Natalie Stoljar
project of conceptual analysis “looks to a priori methods such as introspection for
an answer. Taking into account intuitions about cases and principles, one hopes
eventually to reach a reflective equilibrium.”15 Reflection on intuitions yields what
Haslanger calls a manifest concept. Nicos Stavropoulos, drawing on Frank Jackson’s
work, articulates this first approach:
. . . conceptual analysis seems committed to elucidating the understanding inherent in the
ordinary, day-to-day use of the term and reflection about the things it designates. It seems
targeted . . . to explication of the ordinary conception of the . . . concept, the “folk theory” of
the nature of the things that fall under the concept. And this is achieved by reflection on
possible cases designed to elicit intuitions that reflect that theory.16
In “Hart’s Methodological Positivism,” Stephen Perry says something similar:
Typically, the philosophical analysis of a concept attempts to make explicit what the theorist
claims is in some sense already implicit in our common understanding. This can take the
form of drawing attention to propositions that the theorist argues are either implicitly
presupposed or necessarily entailed . . . [or] of an attempt to show that the concept is
equivalent . . . to some other concept . . . [or] . . . it will amount to a more ambitious attempt
to reduce one concept to a logical configuration of others . . . 17
Both authors characterize conceptual analysis as aiming to reveal and elucidate what
is implicit in actual usage. Stavropoulos writes that it “aims at recovering some,
perhaps idealized, common understanding, in the sense that it articulates but can
never transcend the understanding implicit in ordinary use and reflection.”18 On
Perry’s account, the common understanding reflected in actual usage must be
understood as a kind of moral understanding. He argues that conceptual analysis
of normative concepts like authority and legal obligation, as well the concept of law
itself, necessarily employs moral evaluation. Participants in legal practices are
practical reasoners who expect the practice to give them reasons for action that
they would not otherwise have: “the idea is to make moral sense of the practice by
showing people why and under what circumstances they might have reason to
comply with it.”19 Elucidating a concept with this aim in mind requires providing a
moral justification of the concept. If there is no moral justification of the concept of
legal obligation, for instance, participants in legal practice will have no reason to
obey the law.20
15 Haslanger (2005: 12).

16 Stavropoulos (2000: 70).
17 Perry (2000: 333).
18 Stavropoulos (2000: 71). Note that Stavrapoulos does not adopt the conceptual strategy but
rather a version of the descriptive strategy (see section 2.2). He articulates the former in order to
attribute it to H. L. A. Hart. Stavropoulos (2000).
19 Perry (2000: 350).
20 Perry could be classified as adopting traditional conceptual analysis that in the case of law
necessarily employs moral evaluation. However, since traditional conceptual analysis is usually thought
to be non-moral (because it is “semantic” or analytic), and also because of the affinity between Perry’s
view and that of Ronald Dworkin, it may be neater to classify Perry as employing the third
“ameliorative” strategy rather than the first “conceptual” one.
There is a crucial distinction between Haslanger’s first category of conceptual

analysis and the second and third categories that I outline below. The conceptual
inquiry delivers a manifest concept, namely: the concept that we (in some sense)
have in mind. As we will see, outcomes of the descriptive and ameliorative inquiries
may, as Stavropoulos puts it, transcend actual usage. Thus, the concept revealed by
an investigation into actual usage may not coincide with the concept yielded by the
descriptive or ameliorative inquiries.
2.2. The descriptive project

In the second type of inquiry, we ask “what kinds (if any) our . . . vocabulary tracks.
The task is to develop potentially more accurate concepts through careful consider-
ation of the phenomena, usually relying on empirical or quasi-empirical
methods.”21 This strategy has its roots in the natural kind externalism first elabor-
ated by Hilary Putnam.22 The meanings of natural kind terms, such as “gold” or
“water,” are not given by the intuitions we associate with the concepts when we
reflect on them—intuitions such as that water is a colorless, odorless, potable liquid
that is found in rivers and lakes—but rather by features of the physical stuff that we
use the term “water” to track. Haslanger comments that “scientific essentialists and
naturalizers . . . start by identifying paradigm cases . . . and then draw on empirical
(or quasi-empirical) research to explicate the relevant kind to which the paradigms
belong.”23 It is important however that the externalist model is not limited to
articulating concepts that pick out natural or biological kinds: “Externalism is an
option whenever there are relatively objective types.”24 For instance, the social
externalism defended by Tyler Burge extends externalism to concepts whose
contents are individuated by features of the social environment. In Burge’s famous
example, the content of the concept “arthritis” corresponds to whatever arthritis
really is according to relevant medical classifications.25
For social externalism, as for natural kind externalism, conceptual analysis
proceeds by first picking out paradigms or canonical referents of the type, and
then analyzing the nature of the canonical referent to work out what would
constitute being-in-the-same-kind as the canonical referent. Adèle Mercier,
following Putnam and Burge, emphasizes that being-the-same-kind-as is different
from being-thought-by-ordinary-users-of-a-word-as-being-the-same-kind-as. Further,
“not just any user of the language . . . can extrapolate from canonical referents to
whatever bears the same-kind-of-thing relation to them.”26 We find out about the
nature of the kind, and hence about the meaning of the concept, when a commu-
nity of experts investigates the nature of the canonical referent. Chemists tell us that
the molecular structure of water is H2O and hence the content of the concept of
water is H2O and not XYZ. In the same way, rheumatologists tell us that arthritis is
a condition of the joints and not a generalized condition, and hence “arthritis”
implies “condition of the joints.” Thus, on the descriptive strategy, the content of
21 Haslanger (2005: 12). 22 Putnam (1973). 23Haslanger (2005: 12).

24 Haslanger (2005: 18). 25 Burge (1979). 26 Mercier (2007: 18).
236 Natalie Stoljar
the concept is delivered by the nature of the kind to which the concept refers not by
conventional usage. (One could say for short that meaning is fixed by the linguistic
usage of experts. This is true only because the linguistic usage of experts corres-
ponds to what is taken to be the nature of the kind.)27 On this descriptive account,
conceptual analysis therefore often transcends what we have in mind and what
could be articulated through an examination of common understanding.28
Haslanger introduces a third conception of externalism that she calls “Objective
type externalism:”
Terms or concepts pick out an objective type, whether or not we can state conditions for
membership in the type, by virtue of the fact that their meaning is determined by ostension
of paradigms (or other means of reference fixing) together with an implicit extension to
things of the same type as the paradigms.29
Objective type externalism is different from social externalism because there can be
cases that “fall within objective type externalism because at least there are some
paradigms that fix the reference in question; [although] they don’t seem to qualify
for social externalism because there isn’t yet a standard linguistic usage.”30 Stavro-
poulos focuses precisely on this point when he presents an alternative view of
conceptual analysis to the “criterial semantics” that he attributes to Hart.31 Criterial
semantics is a version of traditional conceptual inquiry. Stavropoulos elaborates an
alternative to the traditional approach that employs “deep” concepts for which
analysis does not attempt to elucidate actual usage at all: “the standard to which
actual usage is responsible is given by a projection beyond actual usage itself.”32
Hence, in the case of law, there may be no standard linguistic usage agreed on
among experts, yet still the descriptive project investigating our operative concept
is a possible one. Its starting point is the paradigm that fixes (by ostension) the
reference of the term “law.”33
27 Coleman and Simchen make an interesting argument that I cannot examine in detail here. They
propose that Putnam’s externalism can be extended to all common nouns like “chair” and “pencil.” On
their view, “law” is analogous to “chair” or “pencil” not to natural kind terms like “gold” or “water.”
The key difference between “chair” and “pencil” on the one hand and natural kind terms on the other
is that the former are not “linguistically deferential”—i.e. they do not “exhibit a division of linguistic
labor” which means that users of terms like “chair” and “pencil” do not need to defer to experts to tell
them what counts as being in the extension of the terms. Coleman and Simchen argue that “law” is not
linguistically deferential due in part to “a perceived lack of agreement among jurisprudents [which
constitutes] a key factor as to why the extension of ‘law’ is not fixed by reliance on jurisprudential
expertise” (2003: 22).
28 Another example: “Being a lawyer is different from being thought to be a lawyer . . . Ordinary
speakers are competent with the word ‘lawyer’ because most of the people whom we think of as lawyers
actually are. But none of what an ordinary speaker need know to use ‘lawyer’ competently determines
the individuation conditions for being a lawyer; those are determined by Bar exams, as these are
determined by those most informed about what one must know to be a lawyer” (Mercier (2007: 17)).
29 Haslanger (2005: 18).
30 Haslanger (2005: 25, fn. 9).
31 Stavropoulos (2000: 81–5).
32 Stavropoulos (2000: 81). Both Haslanger and Stavropoulos draw on Christopher Peacocke’s
example of the mathematical concept of a limit. Peacocke (1998).

33 It should be noted that a prominent form of descriptive approach within the methodology of
law—the descriptive/explanatory approach defended by Julie Dickson (2001) and (2004), does not
2.3. The ameliorative project

The third possible category of inquiry within conceptual analysis asks:
“What is the point of having the concept in question . . . What concept (if any) would do the
work best?” In the limit case, a theoretical concept is introduced by stipulating the meaning
of a new term, and its content is determined entirely by the role it plays in the theory . . .
those pursuing an ameliorative approach might reasonably represent themselves as provid-
ing an account of our concept or perhaps the concept we are reaching for by enhancing our
conceptual resources to serve our (critically examined) purposes.34
Haslanger calls the concept that does the work best the target concept. The
ameliorative strategy has some affinities with Rudolf Carnap’s notion of explicative
definition: “The task of making more exact a vague or not quite exact concept used
in everyday life . . . or rather of replacing it by a newly constructed, more exact
concept, belongs among the most important tasks of logical analysis and logical
construction. We call this the task of explicating, or of giving an explication for, the
earlier concept.”35
Murphy considers the possibility of resolving contested issues about the concept
of law using a Carnapian explicative definition. He writes that “if there is a concept
of law that ‘we all share’, it is indeterminate or partly ambiguous.”36 An explicative
definition, which Murphy characterizes as a kind of instrumentalism, might be
needed to disambiguate the concept. This is a plausible line of argument; however,
there are different ways to articulate instrumentalism that Murphy does not
distinguish: (i) instrumentalism might require adopting a particular disambiguation
of a concept because doing so will have “beneficial moral consequences.”37 Schauer
terms this “prescriptive conceptual analysis;”38 (ii) instrumentalism might corres-
pond to positing or stipulating a particular disambiguation because it is entailed by
a substantive theory of law; or (iii) instrumentalism might require engaging in
moral or other evaluation of the purpose of the social practice to which the concept
refers to yield a more precise or newly constructed concept which better serves the
purpose. Alternative (i) is Murphy’s own position.39 An illustration of alternative
correspond directly to Haslanger’s descriptive strategy. It is, however, closely related to it. I consider
this approach in detail in section 4.
34 Haslanger (2005: 12–13).
35 Carnap (1956: 7–8). Following Carnap, W. V. O. Quine comments that an explicative
definition aims to “improve upon the definiendum by refining or supplementing its meaning”
(Quine (1963: 24–5), quoted in Haslanger (2005: 23–4, fn. 1). Haslanger comments that “on my
view, whether or not an analysis is an improvement on existing meanings depends on the purposes of
the inquiry” (2005: 23–4, n. 1).
36 Murphy (2008: 1093).
37 I adopt Dickson’s useful terminology. She distinguishes three ways in which theorizing about law
might employ moral considerations: the beneficial moral consequences view; the moral evaluation
view; and the moral justification view (2001: 9). For the purposes of this chapter, I will not distinguish
systematically between moral evaluation and moral justification.
38 Schauer (2005).
39 In earlier work Murphy endorsed a beneficial consequences argument to decide the question of
which concept of law to adopt. Like Schauer, he proposed a kind of prescriptivism about the concept of
238 Natalie Stoljar
(ii)—of “stipulative definition” mentioned by Haslanger—may be offered by
“ethical” positivism on which substantive positivism is defended for moral reasons.
The theory that is adopted for moral reasons, ethical positivism, entails a particular
disambiguation of the concept of law.40
The idea of an ameliorative strategy as I will understand it here corresponds to
alternative (iii). Carnap mentions that explicative definition yields “a newly con-
structed, more exact concept.”41 To those familiar with legal theory, this suggests
Dworkin’s notion of constructive interpretation, in which the theorist proposes an
account of the purpose of social practices such as law.42 The first step in the process
of constructive interpretation is analogous to the identification of paradigms on the
descriptive approach. Dworkin proposes that, at the preinterpretive stage, “we have
no difficulty identifying collectively the practices that count as legal practices in our
own culture. We have legislatures and courts and administrative agencies and
bodies and the decisions these institutions make are reported in a canonical
way.”43 At the second, interpretive, stage, substantive answers to questions about
the purpose of the social practice are articulated and defended. Dworkin describes
the “interpretive attitude” of participants in rule-governed social practices: first, the
practice “does not simply exist but has value . . . it serves some interest or purpose or
enforces some principle”; secondly, the rules of the practice are taken by the
participants as “sensitive to its point . . . : People now try to impose meaning on
the institution—to see it in its best light—and then to restructure it in the light
of that meaning.”44 Dworkin’s concept of law—“law as integrity”—is the result
neither of a priori reflection on intuitions nor of descriptive theorizing about the
nature of a paradigm. Rather, it is a theoretical posit that is introduced because it
promotes what Dworkin takes to be the purpose of the social practice of law,
namely, to provide a moral justification of coercive legal institutions. On Dworkin’s
account, therefore, an ameliorative strategy yields a target (or interpretive) concept:
the concept of law as integrity.
The first step in Dworkin’s ameliorative project is to agree on a paradigm or
pretheoretic account of the practice in the world that the term “law” tracks. Recall
that in characterizing the descriptive project Haslanger said that “the task is to
develop potentially more accurate concepts through careful consideration of the
law. Murphy (2000) and Schauer (2005). More recently, however, he has declared that instrument-
alism is “hopeless” because there is also intractable disagreement over the arguments of political
morality themselves that might be invoked in the consequentialist argument. Murphy (2008). He
suggests the possibility of adopting a kind of “eliminativism” about law, i.e. that we stop using the
terms “law” or “the law” and talk about something else such as legal materials and legal decisions (2008:
1106). These are interesting suggestions that I can’t explore further here. I think however that Murphy
exaggerates the negative effects of disagreement at the level of theory or conceptual analysis. Conse-
quentialists and Kantians might disagree over many theoretical matters of morality including how we
conceive of morality itself. Does this mean that we should adopt eliminativism about morality? Or
should we continue to attempt to make the best and most persuasive case possible for the view of
morality that we endorse?
40 Campbell (2005). 41 Carnap (1956: 8).
42 Dworkin (1986). 43 Dworkin (1986: 91).
44 Dworkin (1986: 47).
phenomena, usually relying on empirical or quasi-empirical methods.” The task of

the ameliorative project is also to develop more accurate concepts, but rather than
relying on empirical methods, the theorist evaluates the phenomenon using a
normative and justificatory strategy to characterize its purpose. Thus, as with the
descriptive strategy, the ameliorative project does not engage in an attempt to
characterize actual (or even idealized) common usage. The outcome of an ameli-
orative investigation, like the outcome of a descriptive investigation, may yield a
concept that diverges from the one that we thought we had in mind because an
interpretation of the purpose may for example “better articulate the standard to
which participants were antecedently committed.”45 Stavropoulos is right therefore
to propose that target or “interpretive” concepts have two important characteristics:
they are “deep” in the same way that operative concepts may be deep, and they are
also evaluative.46
3. Conceptual analysis pluralism

I have identified three strategies of conceptual analysis for concepts that refer to
social kinds. These different strategies help to categorize the logical space of legal
methodologies, because proponents of different legal methodologies adopt one or
other of the three strategies. In this section, I elaborate a kind of pluralism about
conceptual analysis for a particular subset of social kind concepts, those that refer to
social practices. Social kind concepts comprise a diverse set, and not all are
sufficiently analogous to law to make adequate illustrations for our purposes.
“Arthritis” refers to an objective social type but not one that is relevantly similar
to law. A distinctive feature of the social kinds that concern us here is that they are
comprised of practices that are used by the participants in the practices “to
understand themselves:” Joseph Raz writes that “it is a major task of legal theory
to advance our understanding of society by helping us to understand how people
understand themselves.”47 Social practices are comprised of explicit and tacit rules
that organize and guide human social behavior. Theorists of these practices are
engaged in an attempt to “advance our understanding of ourselves” by advancing
our understanding of the relevant aspects of human behavior. For example,
theorists of religion explicate the set of explicit and tacit rules that govern human
religious behavior.48
Conceptual analysis pluralism has two components. The first is that there are
different (and incompatible) yet logically possible concepts of law. One way of
46 I have not been able to find a published account of Stavropoulos’s analysis of interpretive
concepts as both deep and evaluative, so I take this point from Dickson (2004: 153).
47 Raz (1994: 237).
48 There is widespread agreement that the task of legal theory is to advance our understanding of
how we understand our own legal practice. Although Perry and Dickson adopt very different
methodologies of law, the passage from Raz is quoted approvingly by both authors. Perry (2000:
348) and Dickson (2001: 40).
240 Natalie Stoljar
characterizing the three strategies identified in the last section is to say that they are
merely epistemological inquiries or decision procedures, that is, that they are
different approaches a theorist might employ to find out about the (correct) concept
of law. Conceptual analysis pluralism is not merely epistemological, however. It
claims that each strategy is logically possible for the analysis of concepts that refer to
social practices. None should be adopted as a matter of necessity or excluded from
logical space as a matter of necessity. All give reasonable answers to the question:
“What is our concept of law?” The manifest concept corresponds to the concept we
take ourselves to have; the operative concept picks out the practices in the world that
our vocabulary tracks, and the ameliorative concept corresponds to the normative
standard to which our practice is implicitly committed. When the different inquiries
deliver different and incompatible concepts of law, each is a logically defensible
concept. It follows therefore that there are several different concepts of law and
genuine pluralism—not merely epistemological pluralism—obtains.
A second feature of pluralism is that because all three inquiries are logically
possible ways of answering the question, “What is our concept of law?,” conflicts
can arise both within the different inquiries and between them. As many have
pointed out, the manifest concept may be indeterminate or it may conflict with the
operative or target concepts. The operative concept may also be indeterminate or
may conflict with the target concept, and so forth. When such indeterminacies or
conflicts arise, it will be desirable or even normatively required to employ a different
strategy.49 For instance, reflection about the operative concept may be desirable to
disambiguate the manifest concept; or reflection on the target concept may be required
to resolve conflicts between the manifest and the operative concepts. This process may
not fully resolve indeterminacies but it is a valuable method of answering questions
we care about and providing a more fully fleshed-out account of the concept.
Pluralism is incompatible with many methodological positions in legal theory.
For example, Stavropoulos employs a parallel with the natural kind externalism
described in the previous section. He argues that determinacy or objectivity can be
secured for legal propositions because the semantics of legal concepts parallels that
of the semantics of terms like “arthritis.”50 Objectivity will be delivered, however,
only if the semantics I described above as the descriptive approach is the correct
semantics for legal terms. In other words, it appears that on Stavropoulos’ view the
approach delivering the manifest concept is incorrect whereas the approach deliver-
ing the operative concept is correct.51 Other theorists—for example Dworkin and
Finnis—propose that theorizing about social practices like law necessarily requires
49 Veronica Rodriguez-Blanco suggested the possibility that a strategy may be not only desirable but
normatively required. Although I do not have space in this chapter to pursue an argument along these
lines, it is plausible that in some cases a particular conceptual inquiry may have normative priority over
others.
51 I am grateful to Stefan Sciaraffa for helping me clarify these points and for pointing out that on
many views, including those derived from Putnam’s semantics for natural kinds, conflict between the
manifest and the operative concepts is not possible because e.g. the manifest concept is just an incorrect
concept of law.
moral evaluation, and relevant moral theorizing delivers the correct concept of the
practice.52 These approaches therefore treat the ameliorative or target concept as the
correct concept of law and hence are also incompatible with conceptual analysis
pluralism.
In the remainder of the section, I illustrate conceptual analysis pluralism by
elaborating two examples of concepts that refer to social practices: “tardy” and
“marriage.” Although there is a difference of degree in the significance of the
practices these concepts track—“tardy” tracks a practice that is circumscribed in
ways that marriage clearly is not—both track practices that structure human
behavior and hence are used by participants in the practices to understand them-
selves. Since they have this feature in common with law, analysis of these concepts
will help to illuminate the conceptual analysis of law.
3.1. Tardy
Haslanger describes the case of the concept “tardy” as it is used in local school
districts. (I present a modified version here.) Suppose that a school district has a set
of rules about tardiness including defining as tardy any child who arrives at school
after the bell rings at 7.50 a.m. School officials who are charged with implementing
the policy report that the meaning of “tardy” corresponds to the definition; hence
their manifest concept “tardy” corresponds to the definition. Despite the rulebook,
however, the way in which “tardy” is used on a daily basis—the actual behavior that
the term “tardy” is taken to track—may vary relative to the school or even relative
to particular classrooms or teachers. Suppose in a particular school there is a five-
minute morning lineup after the bell rings but before students have to go to their
classrooms, and suppose in this school teachers only mark children tardy if they
arrive in class late. In this school, the operative concept of tardy diverges from the
manifest concept under which children are tardy if they arrive after the bell at 7.50.
Now suppose a school inspector arrives and asks why children who arrive after the
bell are not marked tardy. The reply might be: “coming a few minutes late for
lineup but getting to class on time is not really tardy.” As Haslanger points out, a
reply such as this implies that there is “a further way of thinking about what
tardiness ‘really is’ that should take us into normative questions: Should we have
the category of ‘tardy’ in our school district? If so, how should it be defined? . . . [T]he
situation . . . is ripe for an ameliorative inquiry that would have us consider what the
point is of a practice of marking students tardy, and what definition (and corres-
ponding policy) would best achieve the legitimate purposes.”53 The ameliorative
inquiry evaluates the standards of tardiness to which we are implicitly committed,
and hence is desirable in providing a richer account of what tardiness really is.
52 Julie Dickson notes that both Dworkin and Finnis “are insistent that their methodological
approach is the only way to construct an explanatorily adequate account of the nature of law” (2004:
148, my italics).
53 Haslanger (2005: 12).
242 Natalie Stoljar
3.2. Marriage
What is marriage? For example, is same sex marriage really marriage? Or does the
meaning of “marriage” imply that the term “same sex marriage” is a misnomer or
even a contradiction? Reflections on the intuitions implicit in our common
understanding or day-to-day usage of “marriage” may reveal that “marriage” applies
only to opposite sex unions, never to same sex unions. Consider an argument to this
effect that I adapt from a brief to Canadian courts.54 Our common understanding
of the concept of marriage is informed by the history of the institution of marriage.
The relevant history—the history that informs our concept—is Judeo-Christian
religious history in which marriage is necessarily an opposite sex union. Since this
history is implicit in actual usage and common understanding, it is also implicit
that a core or necessary feature of our manifest concept of marriage is that it is an
opposite sex union.
This argument, though it may have some initial plausibility, does not settle the
question of what marriage is or of how the concept of marriage is best analyzed.
Mercier observes that there is a cluster of “historical semantic markers” of the term
“marriage,” namely religious, heterosexual, woman bartering, unequal, non-consensual,
procreative, non-dissolvable.55 She points out that some of these markers are no
longer associated with the concept: e.g. religious, woman-bartering, non-consensual,
non-dissolvable. What justifies the argument that heterosexuality is a core or neces-
sary component of the manifest concept of marriage when these others are not?
Further, Mercier adopts Burge’s social externalism to argue that the meaning of
marriage is fixed not by the results of a priori reflection on our intuitions and their
historical provenance, but rather by the nature of the kind to which paradigms of
marriage belong: “We all agree on canonical referents of ‘marriage’: our moms and
dads, Charles and Diana, Henry VIII and Catherine of Aragon, Socrates and
Xanthippe, Ike and Tina Turner, Edward and Mrs. Simpson . . . But what consti-
tutes being-in-the-same-kind-of-relationship-as-that?”56 The answer to the question
will differ according to which experts on marriage we consult: Roman Catholic
theologians may have one answer, and Islamic religious authorities another. Hence
there are different operative concepts of marriage. None of the religious operative
concepts however correspond to the concept we are seeking: in a secular society, “our
question . . . is what counts as a correct extrapolation when it comes to the Canadian
legal [civil] concept of marriage.”57 In 2011, there is a standard linguistic usage of
“civil marriage [Canada]” that fixes the operative concept and arguably the sense of
“marriage.” But before the 2005 Civil Marriage Act, there was no standard usage.
How were the relevant experts—legislators and courts—to decide then?
In the absence of a standard usage of experts in the case of a social kind concept
like “marriage” we (or parliaments and courts) need to employ an ameliorative
strategy. What is the (secular) point of the institution of marriage? What is the
54 Stainton (2001). See Mercier (2007: 4) for a summary of Stainton’s argument.

55 Mercier (2007: 23). 56 Mercier (2007: 18).
57 Mercier (2007: 20).
purpose of talking about certain kinds of social arrangements as marriages? Is

marriage about promoting procreation and “traditional” family values, or alterna-
tively is it about respect for persons’ rights to identify their intimate relationships as
marriages and have them recognized as such by civil society? An answer to the
ameliorative inquiry is important because ( pace Stainton) the manifest concept will
change over time and there will be more than one operative concept. Moreover “in
practice it is difficult to keep the three strategies entirely distinct.”58 Has the
manifest concept of marriage changed in Canada due in part to the word “mar-
riage” being used operatively since 2005 to refer to same-sex couples? Does our
secular or religious operative concept implicitly import an account of the purpose of
talking about marriage as a particular type of social arrangement?
The examples of “tardy” and “marriage” lend support to both components of
conceptual analysis pluralism. The three modes of inquiry are in play at the same
time; none is excluded from the relevant logical space. Moreover, engaging in
conceptual analysis of “tardy” or “marriage”—that is, trying to work out what
“tardy” really means in our school district or what “marriage” really means in our
secular society—reveals that the different concepts conflict. To answer the ques-
tions we care about, especially about the standards to which we are implicitly
committed, we may have to employ an ameliorative inquiry.
4. The conceptual and descriptive strategies applied to law

Like “tardy” and “marriage,” “law” refers to a social kind through which we
understand ourselves though one that is more complex and internally variegated.
It is “many-sided and vague.”59 In this section, I examine more specifically the ways
in which the conceptual and the descriptive strategies have been or might be
applied to answer the question “What is law?” Both the conceptual and descriptive
strategies correspond to what Perry terms “methodological positivism:” both imply
that we can answer the question “What is law?” without engaging in moral
evaluation or justification.60 I suggest here that both strategies can help to illumin-
ate the concept of law.
Conceptual analysis on the traditional model is, broadly speaking, “committed
to elucidating the understanding inherent in the ordinary, day-to-day use of the
term and reflection about the things it designates.”61 Some have argued that
traditional conceptual analysis is a pointless exercise because it is committed to a
mistaken analytic-synthetic distinction, or because there is intractable disagreement
over common usage. Others have claimed that conceptual analysis of law under-
58 Haslanger (2005: 13). 59 Hart (1958: 622).

60 It should be noted that methodological positivism—although called “positivism”—does not entail
a substantive positivist position. Ethical positivists adopt a moral methodology to argue for substantive
positivism; and non-moral methodology (that is, methodological positivism) could endorse a substan-
tive non-positivist position. Perry (2000: 311). See n. 20 for comments on how to classify Perry’s own
methodological position.
61 Stavropoulos (2000: 70).
244 Natalie Stoljar
mines methodological positivism because conceptual analysis of normative con-
cepts like law of necessity employs moral evaluation. I suggest that all of these
positions are too strong, and that there continues to be room within traditional
(non-moral) conceptual analysis to answer questions about the concept of law.
There is a clear example in the literature of significant progress having been made
in our understanding of the concept of law by purely conceptual argument. Austin
engaged in what appears to be a traditional conceptual inquiry. He offered an a
priori analysis of the concept of law that was generated by reflecting on intuitions
about the distinction between positive law and other types of commands, organiz-
ing them and finally offering necessary and sufficient conditions for the application
of the term “positive law.” Austin’s analysis of laws as “commands with a threat of
sanction attached issued by a political superior to political inferiors” included an
account of legal obligation as a habit of obedience of the inferiors to the political
superior. Although Hart rejects the position that the aim of conceptual analysis
is to provide a definition or resolve the concept into necessary and sufficient
conditions,62 he nevertheless employs a conceptual strategy to rebut many of the
elements of Austin’s account. Hart’s critique of the proposal that laws are com-
mands employs the strategy of reflection on counterexamples. He acknowledges
first that “order backed up by threat” is a plausible account of some examples of law
such as criminal law; crimes are prohibitions issued by parliament and enforced by
sanctions. However there are examples of laws that do not fit this pattern: “there are
other varieties of law notably those conferring legal powers to adjudicate or
legislate . . . or to create or vary legal relations . . . which cannot be construed,
without absurdity, as orders backed up by threats.”63 Hart is reflecting here on
the extension of the term “law,” on the things that “law” designates. The examples
show that the analysis of a law as an “order backed up by threat” is too narrow.
Further, as Perry observes, Hart adopted traditional conceptual analysis when
he “[demolished] Austin’s reductive analysis of legal obligation.”64 Chapter IV of
The Concept of Law contains an extended description of an imaginary case of an
Austinian sovereign, Rex, to which subjects have a habit of obedience. Through a
detailed exploration of features of the imaginary case, and the clarification of the
important distinction between a habit of behavior and a rule that has an “internal
aspect,” Hart argues that the Austinian account does not capture our notion of legal
obligation—the notion that is familiar to citizens of a modern municipal legal
system. On an initial reading, therefore, it is plausible to characterize Hart’s critique
of Austin as an illustration of familiar conceptual analysis yielding an elucidated
account of the concept of law. This conceptual investigation appears also to have
been successful in promoting convergence in common usage at least among
Anglo-American lawyers and theorists. Laws of modern municipal legal systems
are now thought to be social rules not commands. Thus, there is room for a purely
conceptual (and non-moral) strategy to be applied in considering the nature of
62 see Stavropoulos (2000) on this point.

63 Hart (1994: 79). 64 Perry (2000: 333).
normative concepts such as law and legal obligation, even if, as Murphy alleges, this
form of conceptual analysis will not decide some deeply contested questions.65
Let me now turn to the descriptive strategy applied to law. Julie Dickson and
others may prefer to characterize Hart as theorizing about the empirical data of a
modern municipal legal system. And indeed, as many commentators have pointed
out, Hart is committed to the importance of the descriptive. He famously said that
his own project was an exercise in “descriptive sociology” and that it offers a
“morally neutral” analysis of law.66 But there is a puzzling question: how does
Hart both engage in conceptual analysis and maintain his commitment to a
descriptive methodology? Perry argues that Hart’s position is unstable because
there is no way to reconcile an external descriptive strategy which yields an account
of law that “would not necessarily correspond to what ‘we’ . . . ha[d] in mind” with
conceptual analysis.67 The descriptive strategy yielding an operative concept recon-
ciles precisely these features. Indeed a “rational reconstruction” of Hart’s project
(that is, a reconstruction that makes no claim to historical accuracy) might be to say
that he implicitly employs the descriptive strategy.68 Hart identified a paradigm or
“standard case”—the modern municipal legal system—that formed the subject of
his investigation.69 His analysis yielded the conclusion that the paradigm has
certain essential features: it contains primary rules, secondary rules, and a funda-
mental rule of recognition the acceptance of which is manifested in the practice of
the system’s officials. Further, the identification of the essential features of the
standard or local case allowed him to extrapolate to the general case, that is, to the
kind or type picked out by the term “law.” Attributing the descriptive strategy to
Hart allows us to say that what he calls descriptive sociology is an example of
conceptual analysis rather than in tension with it.
Dickson defends the possibility of a descriptive methodology in its own right and
also attributes it to Hart.70 Her account of the descriptive methodology, however,
does not correspond precisely to the one I have identified. Rather, the task of the
theorist is to offer an “explanatorily adequate” account of the empirical data of the
65 There may consequences of this line of thought that I cannot develop here. These kinds of
examples call into question Dworkin’s “semantic sting” argument. Dworkin seems to allege that
disagreements such as these are not genuine but rather count as “talking past each other” because
each side employs a different criterion for what counts as law. It seems obvious to me that genuine
disagreement is possible among positivists who are employing a conceptual strategy, that is, within the
conceptual strategy. Perhaps then only a particular version of conceptual analysis—“criterial semantics,”
the approach attributed to Hart in Stavropoulos (2000)—is subject to the semantic sting argument.
Coleman and Simchen (2003) have convincingly shown, however, that the attribution of criterialism to
Hart or to positivism in general is misguided.
66 Hart (1994: 240).
67 Perry (2000: 314).
68 Of course, Hart could not have explicitly had in mind the later natural kind or social externalism
advocated by Putnam and Burge, but there is some evidence that he was influenced by or himself
influenced R. M. Hare’s notion of “descriptive meaning” in which the features of the things falling in
the extension of terms were elaborated to help with an elucidation of the term’s sense. See e.g. Hare
(1952: 54). (I am grateful to my colleague Andrew Reisner for noticing Hare’s reference to Hart.) See
Rodriguez-Blanco (2003) for another argument.
69 Hart (1994: 3–4).
70 Cf. Dickson (2004) and Waluchow (1994).
246 Natalie Stoljar
legal system under consideration. For Dickson, the legal theorist is more like a
chemist who theorizes about the nature of the colorless, odorless liquid flowing in
rivers and lakes, or an anthropologist who attempts to provide an explanatory
account of the religious practices of a foreign culture. My contention is that this
kind of explanatory theorizing presupposes the descriptive project of conceptual
analysis that I introduced in section 2. Recall that, when a concept refers to an
objective type, the “meaning is determined by ostension of paradigms . . . together
with an implicit extension to things of the same type as the paradigm.”71 The
descriptive/explanatory account must assume that there is some pretheoretic con-
sensus about what we are theorizing about. It must assume a paradigm, standard
case or canonical referent of “law.” Only once a paradigm has been identified can
we engage in descriptive/explanatory theorizing about the empirical data of the
paradigm to determine what would count as being the same kind of thing as that. If
Dickson is right that Hart offers a descriptive/explanatory theory of law,72 and I am
right that the latter presupposes the descriptive strategy of conceptual analysis, we
can indeed attribute the descriptive strategy of conceptual analysis to Hart and
provide an answer to the puzzle about how descriptivism and conceptual analysis
can be reconciled.
What then is the descriptive/explanatory project that Dickson thinks is applic-
able to theorizing about law? There are two dimensions: the first is an account of
what would be required to provide a descriptive/explanatory account of any
phenomenon; the second is what would be required for a descriptive/explanatory
account of hermeneutic practices of which law is one. Dickson points out that value
judgments must be employed in all descriptive/explanatory theories. There is no
such thing as “pure” descriptivism:
[A]ll theorists, no matter the subject matter of their theories, must make value judgments of
a certain kind and that these value judgments are required simply in virtue of the nature of
theoretical accounts; namely, that they attempt to construct cogent and structured explan-
ations that can assist others in understanding as fully as possible the phenomena under
consideration . . . I term these kinds of value judgments “purely metatheoretical” value
judgments and include simplicity, clarity, elegance, comprehensiveness, and coherence
among the virtues that any successful theory attempts to live up to.73
Dickson is concerned to distinguish an implausible conception of a descriptive/
explanatory approach from her plausible conception. The implausible conception
conceives of a theorist as someone who “merely passively record[s] and reproduce[s]
the passing legal scene, hence not providing an elucidation or analysis of aspects of
71 Haslanger (2005: 18).

72 Note that Perry does not think it plausible to attribute this strategy to Hart. Perry (2000: 321).
For more discussion of this point, see section 5.
73 Dickson (2004: 135). The metatheoretical values identified by Dickson seem to correspond to
what Leiter calls epistemic values: Leiter (2003). He points out that although descriptive theorizing is in
some sense evaluative because it employs epistemic values, it need not employ moral values and hence
it is not morally evaluative. For a discussion of the similarities and differences between Dickson’s and
Leiter’s accounts of descriptive theorizing, see Dickson (2004: 133–41).
law at all.”74 Dickson is right that a descriptive/explanatory theory must provide

more in the way of explanation and elucidation than a merely passive recording of
lists of empirical facts. As she puts it, it must employ metatheoretical values in an
attempt to “construct cogent and structured explanations.” Hence it is evaluative in
some sense. However, for Dickson, this process is not entirely sufficient. Her
descriptive approach is evaluative also because law is a hermeneutic concept that
(following Raz) she claims is “used by people to understand themselves and their
social world.”75 Legal theorists therefore must make what she terms “indirectly
evaluative” judgments about which features of law are significant for the purposes
of explanation in a manner that is sensitive to the beliefs of the participants in legal
practice, including their beliefs about the moral features of law. Here we reach the
crucial distinction in the argument. The key question is whether legal theorists can
avoid introducing a morally evaluative component. Dickson is at pains to maintain
that a successful legal theory must both “take adequate account of how law is
understood by those living under it”76 while at the same time it “need not take a
stance on whether the participants are correct in their ascriptions of . . . moral
value.”77 In the same way, an agnostic studying the Roman Catholic Mass or an
anthropologist studying a foreign culture need not herself morally evaluate the
Mass or the commitments of that culture, even to make judgments about which
features are significant or important to investigate.
5. An argument for the importance of the ameliorative inquiry
Dickson is arguing against those who claim that moral evaluations by the theorist
are necessary for an explanatorily adequate account of law. For example, one might
argue that theorizing about social practices in general requires moral evaluation, or
that moral evaluation is required for the specific case of law. Dworkin adopts the
former approach. He proposes a theory of interpretation of social practices that is
based on a participant viewpoint and which requires reflection by the participant/
theorist about the (moral) purpose of the practice.78 Perry is closer to the latter
approach. He thinks that for distinctively legal concepts, such as obligation,
authority, and the concept of law itself, conceptual analysis necessarily employs
moral evaluation.79 Dickson’s elaboration of the descriptive and indirectly evalu-
ative has convincingly shown that moral evaluation is not necessary for theorizing
about law for either the former or the latter reason. Thus, Dickson’s argument helps
74 Dickson (2004: 132). Dickson suggests that Perry attributes this implausible conception of
descriptivism to Hart, and hence that in critiquing Hart, he is attacking a “straw man” (Dickson (2004:
133)). It is true that Perry does not think Hart can plausibly be said to be engaging in a descriptive/
explanatory enterprise. But it seems to me this is not because he attributes the implausible conception
to Hart. Rather, according to Perry, there is no evidence that Hart does in fact theorize about the legal
data employing the metatheoretical values that Dickson advocates. Perry (2000: 321).
75 Dickson (2004: 137). 76 Dickson (2001: 44).
77 Dickson (2001: 69). 78 Dworkin (1986).
79 Perry (2000).
248 Natalie Stoljar
to solidify the first component of conceptual analysis pluralism because she has
established that the descriptive (non-moral) strategy is a possible mode of inquiry.
A question remains however: under what circumstances should we leave the
conceptual and/or descriptive strategies and move to the ameliorative? Under what
circumstances is the ameliorative strategy desirable?80 I sketch the beginnings of
three arguments showing that there are answers to important questions about our
concept of law that only an ameliorative strategy can provide. The first two
arguments point out that for law, as for the concepts of tardy and marriage,
conflicts potentially arise between the manifest and the operative concept or within
the operative concept itself, and thus an articulation of a target concept will provide
a more fully fleshed-out account. The third argument focuses on the role of
the theorist for practices like law. I claim that it is artificial to relegate a theorist
of her own practice always to the position of observing the practice from the
outside. Theorists are often participants in the practice, and hence they potentially
play a structuring role such as refining and making precise the concept under
consideration.
5.1. Conflict between manifest and operative concepts of law

In the last section I argued that traditional conceptual analysis was successful in
showing that laws are social rules not commands, and that there is convergence in
common usage among lawyers, judges, and Anglo-American legal theorists to that
effect. Let us call the position that laws are social rules (not commands) an aspect of
the manifest concept of law. On Dickson’s descriptive/explanatory account, how-
ever, whether or not laws are commands or rules is a question of the outcome of our
best descriptive/explanatory theorizing about the legal data, a process that is
informed by metatheoretical values such as simplicity, elegance, and coherence.
For the purposes of illustration, suppose the empirical data are compatible with
both theories. The question of what theory to adopt will have to be resolved by the
application of metatheoretical values. Which theory is simpler, more unified, and so
forth? Hart often mentions the simplicity of Austin’s theory of law.81 Finnis
observes that an attractive feature of “the notions of command, political superior
and habit of obedience was precisely their simplicity and definiteness.”82 This
suggests that employing the value of simplicity to break the tie would yield
Austin’s rather than Hart’s version of positivism. Perry also observes that, on the
descriptive/explanatory approach, distinguishing between the two theories would
require employing metatheoretical values to argue that one has superior explanatory
80 Veronica Rodriguez-Blanco pointed out that there is an additional normative question, namely
whether the ameliorative strategy has normative priority over others. I think the claim that the
ameliorative strategy takes normative priority is plausible and worth pursuing. For the moment
however, I limit myself to the weaker position that it is desirable to employ the ameliorative strategy
in cases in which it will answer questions we care about.
81 For example, Hart (1994: 51).
82 Finnis (1980: 5).
power.83 Hence, the operative concept delivered by descriptive/explanatory theor-

izing will be determined by which metatheoretical values have priority over others.
If the operative concept corresponds to the command notion and hence conflicts
with the manifest concept delivered by the conceptual inquiry, an ameliorative
strategy may be employed to resolve the conflict. An ameliorative strategy will
evaluate which notion of law—command or rule—best serves our theoretical and
moral purposes.
5.2. The operative concept may be indeterminate

Murphy says that there is “insufficient agreement in the intuitions that are the data
for any philosophical conceptual analysis.”84 He claims in effect that the manifest
concept is indeterminate. Similarly, it is likely that the operative concept that tracks
the social kind picked out by the term “law” is indeterminate. For example, Finnis
points out that Hart leaves open the question of whether officials of a legal system
adopt the internal point of view for prudential or moral reasons. So there may be no
determinate account within descriptive/explanatory theory of the notion of “legal
viewpoint.”85 A theorist qua anthropologist will have to conclude something like
“It is indeterminate whether officials in modern municipal legal systems are
motivated by moral or prudential reasons,” and indeed it might be sufficient for
certain purposes to leave it at that. Finnis’s discussion suggests, however, that the
question of which viewpoint law requires of officials is a question of theoretical
significance that we care about. If we want an answer to the question within our
theory, we will have to adopt an ameliorative inquiry that examines the purposes of
legal practice.
The selection of paradigms and the problem of extrapolation might also intro-
duce indeterminacy into the operative concept. I suggested that a good candidate
for canonical referent of the term “law” is Hart’s modern municipal legal system.
This suggestion is not uncontested in the literature. In The Model of Rules I,
Dworkin in effect selected another paradigm.86 For Dworkin, the canonical refer-
ent of our term “law” is the Anglo-American common law legal system not the
broader modern municipal system. Since judicial decisions are such a significant
component of the common law system, any descriptive/explanatory account of this
paradigm requires us to evaluate data about adjudication. Hence, Dworkin was not,
as some commentators claim, simply changing the subject from a theory of law to
a theory of adjudication. Rather, at this early stage in his thought, Dworkin like
Hart was employing the first stage of a descriptive strategy. The next stage of the
descriptive strategy is to move by extrapolation from the local to the general; that is,
83 Indeed it is precisely because Hart did not engage in an argument over which theory best
exemplifies superior predictive power that Perry thinks it implausible to attribute descriptive/explana-
tory theorizing to Hart. He says that Hart does not claim that a theory such as Austin’s which describes
“social phenomena in purely behavioristic terms and treat[s] the internal point of view as epiphenom-
enal at best is deficient in the scientific sense of failing to have predictive power” (Perry (2000: 321)).
84 Murphy (2005: 7). 85 Finnis (1980: 13). 86 Dworkin (1977).
250 Natalie Stoljar
to work out what constitutes being in the same kind or type as the paradigm.
In Dworkin’s theory, the moral principles employed by common law judges in
adjudication were among the key features of the canonical referent of “law.” Thus,
broadly speaking, extrapolation from the paradigm yielded the result that moral
considerations can constitute criteria of legal validity. I am suggesting that the
difference between Hart and Dworkin (at this early stage in the latter’s thought) is
not necessarily a difference in their methodologies but rather a difference in choice
of paradigms and hence in the identification of key features that the theorist uses to
extrapolate from the local to the general.87 In other words, Hart and Dworkin had
different although overlapping operative concepts of law. One might attempt to
employ further descriptive/explanatory theorizing to decide which operative con-
cept to adopt (a strategy perhaps adopted by inclusive positivists). It may be more
profitable to adopt an ameliorative strategy in which questions about the purposes
of our legal practice can be directly addressed.
5.3. The role of the theorist

Dickson’s descriptive strategy requires that the theorizing be “descriptively pure,”
not in the sense that it is non-evaluative, but rather in the sense that the theorist
avoid direct moral evaluation of the practice. She brings this out using an example
of an agnostic’s descriptive/explanatory account of a Roman Catholic Mass. Be-
cause she is agnostic, she does not seek to take on the perspective of the participants
in Catholic religious practices to morally evaluate their commitments. So, as
mentioned above for the case of legal viewpoint, if there are questions that are
indeterminate from the perspective of the participants in the Roman Catholic
religion, the theorist will be limited to making descriptive statements of the
following sort: “For the participants in the practice, the answer to questions A, B,
C, etc are indeterminate.”
Dickson implies that the theorist not only can but should always insulate herself
from the task of moral evaluation, and should always act qua anthropologist and
never qua participant in the practice. Is it desirable for the theorist always to be
limited to the position of agnostic? Is it desirable for a legal theorist never to adopt
an ameliorative stance in which moral evaluation might be necessary? It seems that
the answer is “no.” It is artificial to claim that someone reflecting on her own
community’s secular notion of marriage, for example, always must be agnostic
about the moral purposes of marriage for herself and her community. When a
theorist herself is a participant in a practice, her stance is analogous, not to that of
anthropologists or agnostics, but rather to that of Catholic theologians who theorize
about their own practice by adopting an ameliorative strategy to work out the
meaning of Catholic doctrine.
Moreover, theorists who always avoid the ameliorative inquiry will be unable to
answer significant substantive questions. Consider again Dickson’s defense of
87 As Leiter (2003) points out, Finnis argues that identifying the “central case” requires moral
evaluation or considerations of practical reasonableness. I am making a similar point here.

methodological positivism. The import of her argument is that it potentially clears

the way for a non-moral argument for substantive positivism. For example, she
implies that from mere description (albeit of an indirectly evaluative sort) of the
behavior and the internal states of the participants in our legal practice, we can draw
substantive conclusions about the conditions of legal validity. She writes that it
follows from Hart’s descriptive/explanatory theory that “officials in a legal system
must regard themselves as bound in common by a rule that is manifest in their
official practice and by means of which they identify what counts as valid law—that
must be present in order for a legal system to exist.”88 This seems to be intended as
a statement of a substantive positivist thesis that allegedly follows from a descriptive
methodology. However, how do we get from statements like “officials regard
themselves as bound in common by a rule” and “officials use the rule to identify
what counts as valid law” to the conclusion that the rule of recognition is the only
source of legal validity? In other words, how do we get from a descriptive claim,
namely “our legal vocabulary tracks a practice in which participants think the only
source of legal validity is the rule of recognition” to the substantive theoretical
conclusion that “the rule of recognition is the only source of legal validity”? I am
suggesting that answers to these important questions will be delivered only by
adopting an ameliorative strategy that directly addresses questions about the
legitimate sources of legal validity. That is, they will be delivered only when the
theorist acts qua theologian rather than qua agnostic or anthropologist.
6. Conclusion
“Law” is a concept that refers to an evolving social kind. There is a tendency however
to get stuck in debates over whether a particular way of approaching the conceptual
analysis of “law” is the right way as if the content of the concept is fixed and static and
completely distinct from the structuring effects of theorists and other participants in
the practice.89 Theorists and other participants in legal practice must take up the
challenge offered by the nature of law as a social kind and embrace not the “arid
wastes of definition” but the question of what kind of thing we want law to be.
References
Austin, J. (1995). The Province of Jurisprudence Determined. Edited by W. E. Rumble.

Bix, B. (2010). “John Austin,” in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy
(Spring 2010 Edition): <http://plato.stanford.edu/archives/spr2010/entries/austin-john>.
Burge, T. (1979). “Individualism and the Mental,” Studies in Epistemology: Midwest Studies
in Philosophy, 4: 73–121.
88 Dickson (2004: 126).

89 The last two sentences recap some ideas expressed in Schauer (2005) and Haslanger (2005).
252 Natalie Stoljar
Campbell, T. (2005). “Prescriptive Conceptualism: Comments on Liam Murphy,
‘Concepts of Law’,” Australian Journal of Legal Philosophy, 30: 20–9.
Carnap, R. (1956). Meaning and Necessity: A Study in Semantics and Modal Logic. Chicago:
University of Chicago Press. (1st edn. 1947.)
Coleman, J. and Simchen, O. (2003). “Law,” Legal Theory, 9: 1–41.
Dickson, J. (2004). “Methodology in Jurisprudence: A Critical Survey,” Legal Theory, 10:
117–56.
Dworkin, R. (1977). “The Model of Rules I,” in Taking Rights Seriously. Cambridge, MA:
Finnis, J. (1980). Natural Law and Natural Rights. Oxford: Clarendon Press.
Hare, R. M. (1952). The Language of Morals. Oxford: Clarendon Press.
Hart, H. L. A. (1958). “Positivism and the Separation of Law and Morals,” Harvard Law
Review, 71: 593–629.
Haslanger, S. (2000). “Gender and Race: (What) Are They? (What) Do We Want Them to
Be?,” Noûs, 34: 31–55.
Haslanger, S. (2005). “What Are We Talking About? The Semantics and Politics of Social
Kinds,” Hypatia, 20: 10–26.
Haslanger, S. (2006). “What Good Are Our Intuitions?,” Aristotelian Society Supplementary
Volume, 80: 89–118.
Leiter, B. (2003). “Beyond the Hart/Dworkin Debate: The Methodology Problem in
Jurisprudence,” American Journal of Jurisprudence, 48: 17–51.
Mercier, A. (2007). “Meaning and Necessity: Can Semantics Stop Same-Sex Marriage?,”
Essays in Philosophy, 8: article 14.
Murphy, L. (2000). “The Political Question of the Concept of Law,” in J. L. Coleman (ed.),
Hart’s Postscript: Essays on the Postscript to The Concept of Law. Oxford: Oxford
University Press.
Murphy, L. (2005). “Concepts of Law.” Australian Journal of Legal Philosophy, 30: 1–19.
Murphy, L. (2008). “Better to See Law This Way,” New York University Law Review, 83:
1088–108.
Nussbaum, M. (1995). “Objectification,” Philosophy and Public Affairs, 24: 249–91.
Oberdiek, J. and Patterson, D. (2007). “Moral Evaluation and Conceptual Analysis in
Jurisprudential Methodology,” in R. Harrison (ed.), Current Legal Issues: Law and
Philosophy. Oxford: Oxford University Press.
Peacocke, C. (1998). “Implicit Conceptions, Understanding, and Rationality,” Philosophical
Issues, 9: 43–88.
Perry, S. (2000). “Hart’s Methodological Positivism,” in J. L. Coleman (ed.), Hart’s
Postscript: Essays on the Postscript to The Concept of Law. Oxford: Oxford University
Press.
Putnam, H. (1973). “Meaning and Reference,” Journal of Philosophy, 70: 699–711.
Putnam, H. (1975). “The Meaning of ‘Meaning’,” in Philosophical Papers, vol. 2: Mind,
Language and Reality. Cambridge: Cambridge University Press.
Quine, W. V. O. (1963). “Two Dogmas of Empiricism,” in From a Logical Point of View.
New York: Harper and Row.
Raz, J. (1994). “Authority, Law and Morality,” in Ethics in the Public Domain. Oxford:
Clarendon Press.
Rodriguez-Blanco, V. (2003). “A Defence of Hart’s Semantics as Nonambitious
Conceptual Analysis,” Legal Theory, 9: 99–124.
Schauer, F. (2005). “The Social Construction of the Concept of Law: A Reply to Julie
Dickson,” Oxford Journal of Legal Studies, 25: 493–501.
Stainton, R. (2001). Expert witness affidavits in Egale v. Canada (AG), filed in the Supreme
Court of British Columbia, in the matter of Applications for Licenses by Persons of the
Same Sex who Intend to Marry; and in the matter of The Marriage Act and The Judicial
Review Procedure Act (Vancouver Registry No. L001944; L002698; L003197), August
2001; and in Halpern v. Canada (AG), filed in the Ontario Superior Court of Justice
(Court files 684/00, 30/2001), November 2001.
Stavropoulos, N. (1996). Objectivity in Law. Oxford: Oxford University Press.
Stavropoulos, N. (2000). “Hart’s Semantics,” in J. L. Coleman (ed.), Hart’s Postscript: Essays
on the Postscript to The Concept of Law. Oxford: Oxford University Press.
Stavropoulos, N. (2008). “Interpretivist Theories of Law,” in E. N. Zalta (ed.), The Stanford
Encyclopedia of Philosophy (Fall 2008 Edition): <http://plato.stanford.edu/archives/
fall2008/entries/law-interpretivist>.
Waluchow, W. J. (1994). Inclusive Legal Positivism. Oxford: Oxford University Press.
PART IV
NEW DIRECTIONS
11
Legal as a Thick Concept*
David Enoch and Kevin Toh
1. The data: some common observations about legal statements
“A leasehold interest is alienable.” “Corporate insiders are prohibited from making

securities trades based on material non-public information.” “Citizens have the
right of assembly.” These are some examples of legal statements—the statements by
uses of which we characterize the laws of a legal system and communicate with each
other their contents. These legal statements, the point is often made, walk a fine
line between the descriptive and the normative.
On the one hand, such statements seem to be very closely tied to some perfectly
natural facts—e.g. facts like how many people raised their hands in a certain room
when someone had just said “Let us vote”; facts like which criteria are considered by
certain people in gowns as authoritative for themselves and others in their commu-
nity; facts like what conventions of a certain kind commonly prevail among the
members of a community; or some such. The tie between legal statements and such
facts is so close in fact that in many instances, invoking some facts like these is
enough to show that a given legal statement is true or warranted. And this tight
relation between the law and such natural facts is something that a theory in general
jurisprudence should capture or explain.
On the other hand, legal statements seem normative, or at least in the vicinity of
normativity. Talk about “the normativity of law” is unclear and obscure, to be
sure.1 But without committing ourselves to this way of talking, we can list the
following, perhaps related, characteristics as the features of legal discourse, and
perhaps of the law, that a theory in general jurisprudence should also try to
accommodate: (i) legal statements typically involve normative terminology (e.g.
deontic terms “right” and “obligation”); (ii) disagreements about legality seem to
* For helpful comments on earlier drafts, we thank Matti Eklund, Brian Leiter, Andrei Marmor,
Joseph Raz, and Stefan Sciaraffa. The thinking for this chapter started in conversations in the 2009–10
academic year at the University of Texas School of Law in Austin. We are grateful to many members of
the UT Law School faculty—especially Mitch Berman, John Deigh, and Larry Sager—for providing us
with a wonderfully congenial and stimulating environment. David Enoch’s research was supported by
the Israel Science Foundation (grant no. 136/09).
1 See Enoch (2011).
258 David Enoch and Kevin Toh
manifest some of the common characteristics of normative disagreements, and do
not look very much like purely factual or empirical disagreements; (iii) when certain
people—perhaps judges, or perhaps some judges some of the time—make certain
legal judgments, these judgments seem to incorporate an element of acceptance
or endorsement; and (iv) there seems to be some important relation between
legal statements on the one hand and reasons for action on the other—certainly
legal reasons, but perhaps also real or genuine (unmodified) reasons.
Much of what is going on in the long contentious and oftentimes scholastic
debate about the nature of law can be understood, we conjecture, as various
attempts to accommodate both the close relation between the law and some
naturalistic—perhaps social or historical—facts, and the relation between the law
and the normative. At the risk of caricaturizing some of the debate, legal positivists
could be characterized as those who take the relation between law and social facts as
the central phenomenon to explain, and attempt to explain the ties to normativity
in a way that is consistent with that allegedly central phenomenon. And natural law
theorists could be characterized, we believe, as those who are more impressed with
the ties to normativity, and attempt to explain the relation to social facts in a way
that’s consistent with those ties.
One may initially be tempted by the thought that some judicious disambigu-
ation is all that is called for. Perhaps some—we can call them “insiders”—use such
words in order to endorse, recommend, and/or criticize some behavior. Legal
statements, out of their mouths, are normative statements. And perhaps others—
we can call them the “outsiders”—use such words in a way that is normatively
neutral or uncommitted, and that merely picks out some complex social facts.2 But
this does not appear a promising line. For insiders and outsiders can, and often
actually do, engage each other in meaningful discussions about what the law is, or
about which norms are legally valid. And they do not seem to be talking past each
other when they do. So ambiguity does not seem to be the phenomenon usually at
play here.3
In this chapter, we want to draw jurisprudents’ attention to a body of literature
bordering on moral philosophy and the philosophy of language—the literature
discussing the so-called thick terms and concepts. Roughly, these are the terms and
concepts usages or expressions of which combine description and evaluation. We
think that discussions of the nature of thick terms and concepts could help us in
thinking about the nature of legal statements, and of law more generally, perhaps
even pointing us towards some hitherto neglected theoretical avenues and options.
And we also think—though we are not as confident—that reflecting about the
example of legal concepts and statements can also serve to enrich the discussion of
2 This disambiguation reflects one way of understanding Hart’s distinction between internal and
external legal statements, and of the distinction between judgments from the internal point of view and
those from the external point of view. See Hart (1961/94: e.g. 56–7, 88–91, 102–3). See section 5 for
further discussion of this issue.
3 This may be one way of understanding Raz’s criticism of Hart’s distinction between internal and
external legal statements, as we elaborate in section 5. See Raz (1974: 140–3; 1975/90: 171–7; 1977:
153–7; 1980: 234–8).
Legal as a Thick Concept 259
thick concepts more generally. The hypothesis that should be considered quite
seriously, we believe, is that legal statements employ thick terms, and thereby
express thick concepts, and that this feature of legal statements is responsible for
their straddling the line between the descriptive and the normative. We are not
confident that the resulting view or family of views is the way to go in legal
philosophy.4 But we are confident that it is an option, or more precisely constitutes
a family of options, very much worth pursuing.
2. Thick concepts
At least since Bernard Williams (1985: esp. 129–31, 140–5) introduced these
terms, it has been common to distinguish, among normative or evaluative (or
more specifically ethical) concepts, thin and thick ones. The distinction is not
meant to be a categorical or bright-line one, but is instead one of degrees.5 Thin
ethical concepts—e.g. good, bad, right, wrong,6 and such—are, as it were, purely or
predominantly evaluative or normative (which two terms we will treat as inter-
changeable in this chapter) in content, and devoid or nearly devoid of descriptive or
factual content. On the other hand, thick ethical concepts—e.g. courageous, cruel,
disgusting, elegant, and such—while containing an element of evaluation, have a
much greater proportion of descriptive or factual contents than thin concepts. For
example, to say of an act that it is courageous is not merely to evaluate it in some
(usually, or perhaps even necessarily, positive) way. It is also to say of it that it
involves something by way of coping with fear, and perhaps also danger; and it may
even say something about the relation between the evaluation and this particular
descriptive content—say, that the act is worthy of positive evaluation partly because
it involves coping well with fear in the face of perceived danger. To say of an act that
it is cruel is not merely to evaluate it in some (usually, or perhaps even necessarily,
negative) way. It is also to say of it that it involves something by way of indifference
to or even enjoyment of the suffering of others; and it may even say something
about the relation between the evaluation and this particular descriptive content—
perhaps that the act is worthy of negative evaluation partly because it involves
indifference to or enjoyment of the suffering of others. And so on.7
This intuitive, initial way of characterizing the contrast between thin and thick
concepts is all right as far as it goes. But when it comes to the attempt to give a
fuller, more explicit philosophical account of the distinction and its implications,
4 In fact, one of us (Enoch) is rather sympathetic to the more shamelessly positivist, conventionalist,
reductionist, non-normative way of understanding legal statements, while the other (Toh) is much
more impressed by the seemingly normative nature of (at least some central instances of ) legal
statements.
5 See Scheffler (1987).
6 We shall in this chapter use italics to indicate concepts and quotation marks to indicate terms or
expressions. We trust that contexts will make plain when we are using italics to designate concepts and
when we are using italics for emphases.
7 Things are of course much more complicated than what these quick examples indicate, and there
is room for many more options here. See the helpful discussion in e.g. Elstein and Hurka (2009).
things are far from clear or uncontroversial.8 We will return to some of these
controversies a few paragraphs hence. For now, though, and without pretending
that anything here is obvious or settled, let us note the following characteristic of
thick concepts.
The uses of words or terms expressing thick concepts, or thick terms—at least
their paradigmatic, literal uses—in some way involve a commitment to an under-
lying normative or evaluative judgment. Thus, paradigmatic, literal uses of the
word “courageous” (e.g. “It was courageous of her to stand up to her boss in that
way”) in some way commit the speaker not just to the relevant descriptive content
(“Standing up to her boss in that way involved overcoming some fear”), and not
just to an addition of some evaluative “coloring” (“And hurray for that!”), but also
to some more general evaluative judgment—in this case, perhaps something like
that overcoming fear is often the thing to do, or that many acts that involve
overcoming fear are the better for it.9 Importantly, such normative judgments
seem to underlie even some non-assertoric uses of thick terms. If a person says “It
was very courageous of her to stand up to her boss,” and a second person says “No,
it wasn’t courageous at all,” the second person seems just as committed as the first
person is to the underlying normative judgment—namely, that oftentimes over-
coming fear in the face of danger sheds favorable light on an action. The second
person seems to be disagreeing either about the attribution of the property picked
out by the descriptive content of “courageous,” or about its presence counting in
favor of the action on this occasion. Thus, we expect the second person to continue
with something like: “No, it wasn’t courageous at all, as she has tenure, and her boss
has no power over her,” or “No, it wasn’t courageous at all, but rather foolhardy, as
in this case she shouldn’t have taken that unnecessary risk.”
That a commitment to the underlying normative judgment is typically involved
even in negative statements involving thick concepts gives rise to one of the most
important phenomena related to thick concepts—namely, the not-one-of-my-words
response. As the story goes,10 Oscar Wilde was asked during his trial whether he
denied that something he said or did was an instance of blasphemy. And Wilde
responded: “‘Blasphemy’, sir, is not one of my words.” Wilde responded well, for by
either accepting or denying that his act was blasphemous, he would have committed
himself to the normative judgment underlying uses of the word “blasphemous”—
8 For this observation, for a critical evaluation of several common suggestions, and for an original
suggestion (in terms of epistemic analyticity), see Eklund (2011).

9 We are not insisting that the underlying normative judgment is fully general, exception-less,
universal, or anything of this sort. Whether it will be any or all of those things depends on further
considerations—e.g. whether ethical particularism and reason-holism are at the end of the day plausible
theories. For the now-standard defense of such particularist views, see Dancy (2004). For some
criticism see McKeever and Ridge (2006); Schroeder (2009). And for a discussion of the possible
variability in the evaluative or normative element of thick concepts across contexts, and for what
follows from such variability regarding the nature of thick concepts, see Väyrynen (2011).
10 See e.g. Gibbard (2003: 288–9); Heuer (2012: 224 n. 22). Based on Foldy (1997: 8), Eklund
(2011: 33) brings a somewhat different quote from the one that follows in the text: “the word
‘blasphemous’ is not a word of mine.” Williams (1995: 29) also refers to the event, but recounts it
as centering on the word “obscene.”
roughly, that speaking against God, the Church, or some religious tenets and
teachings is to be avoided—and this is precisely the judgment that Wilde was
eager to refrain from committing himself to.11 We will return to the not-one-of-
my-words response later on.
Theorists of thick concepts differ with regard to what can be called (following
Väyrynen 2011: 2) the location of evaluation. Everyone agrees, for instance, that a
certain descriptive content—i.e. having something to do with coping with fear
and danger—is a part of the very meaning of the word “courageous.” But while
everyone further agrees that the word “courageous” is used also to convey some
evaluative or normative message, theorists differ about how exactly it does that. The
main divide here is between seeing this as a semantic matter, and seeing it as a
pragmatic one. If we think that some positive evaluation is a part of the very
meaning of “courageous,” we will have to conclude that the judgment “Her facing
up to her boss was courageous” entails “Her facing up to her boss was good in at
least one respect,” just as it entails “Her facing up to her boss involved in some way
overcoming fear.” Furthermore, if evaluation is a matter of the semantics of thick
terms, then if the evaluative thought conveyed is false, so is the relevant judgment as
a whole. If, for instance, there was nothing good about her facing up to her boss,
then “Her facing up to her boss was courageous” will be false, and this even if it did
involve overcoming fear, and thereby satisfied the descriptive part of the content of
“courageous.” In general, if we go for a semantic answer to the location-of-
evaluation question, we seem to think of the evaluative and the descriptive elements
of thick concepts symmetrically, as equal partners in constituting the relevant
concept.
But this is not the only possible way to go here, and some theorists think that the
evaluative nature of thick concepts is a matter of pragmatics rather than semantics.
Thus, they believe that while the sentence “Her facing up to her boss was
courageous” usually serves, among other things, to convey a positive evaluation
of her facing up to her boss, the sentence does not strictly speaking say or entail that
what she did was good in some respect. Rather, according to these theorists, the
relevant evaluation is conversationally implicated or is presupposed.12 On such
pragmatic accounts, there is an important difference between the descriptive and
the evaluative elements of thick concepts. Thus, saying “Her facing up to her boss
was courageous, but it did not involve in any way overcoming fear” may be strictly-
speaking contradictory, whereas “Her facing up to her boss was courageous, but it
was not in any way good” need not be. The second part of the latter sentence may
11 This is not the only possible explanation of the not-one-of-my-words response. Another is that
the concept blasphemous has separable descriptive and normative parts to its content; that Wilde
accepted that the action he was accused of satisfied the descriptive part of the content of blasphemous,
insisted that it did not satisfy the normative part, and was unable to convey this conjunction using the
word “blasphemous.” Whether this is a good explanation—and indeed, to what extent it differs from
the one in the text—depends on details in the theory of thick concepts, which we cannot get into here.
12 For a quick survey of the semantic and pragmatic answers to the location-of-evaluation question,
and for many references, see Väyrynen (2011: 2). For Väyrynen’s own presuppositional account, see
his (2012).
merely involve a cancellation of a usual implicature or presupposition of what is
being asserted in the first part of the sentence. Notice, by the way, that such
pragmatic approaches are perfectly consistent with the very natural thought that if
one does not understand the evaluative nature of a thick concept, then one is
missing something important about it. One may be missing something about the
way it is often, usually, or even paradigmatically used. It is just that what one is
missing is not something about its semantic value.
For our purposes here, we do not need to take sides regarding this particular
controversy—though we are going to get back to it in subsequent sections. It is,
however, worth noting the following oddity in the literature on thick concepts.
Most of this literature seems to assume that there will be one uniform answer to the
location-of-evaluation question with regard to the entire class of thick concepts and
thick terms. But this assumption is most likely false.13 We are here in the business
of understanding and explaining a huge class of concepts and expressions in a
natural language. And it cannot be ruled out that they do not all behave in the same
way. It is quite possible, for instance, that utterances using racist derogatory words
presuppose (rather than assert) the relevant negative evaluation,14 while “courage-
ous” involves or entails a positive evaluation as a part of its very meaning, and
“lie”15 has a negative evaluation merely as a conversational implicature. Indeed,
with this possibility in mind it becomes tempting, if not irresistible, to think of
natural-language words as involving some evaluative element in any number of
ways, across the pragmatic-semantic divide. Perhaps, for instance, the expression
“physically strenuous”16 is evaluatively neutral in meaning, but in some contexts (in
discussing workouts, say) conveys an evaluative content as a fairly standard conver-
sational implicature. And perhaps the word “fashionable” is almost invariably used
evaluatively—certainly more standardly so than “physically strenuous”—but still
only as a pragmatic matter (as evidenced, perhaps, by the possibility of cancelling
the relevant evaluative suggestion). And perhaps in the case of “blasphemous,” the
evaluative content is already a part of its semantic value. If so, usage of the words
“physically strenuous,” “fashionable,” and “blasphemous” without a commitment
to the relevant evaluation will be, respectively, non-standard in some contexts, non-
standard across contexts, and insincere or confused.17
13 Before delving into a discussion of thick concepts, Williams (1995: 25) more generally disputes
the “fiction” that “there is a homogeneous class of ethical assertions” that merit a uniform treatment or
explanation. For hints of some discomfort regarding a uniformity assumption about thick concepts in
particular, see Elstein and Hurka (2009: 520), where they quickly distinguish between important and
unimportant thick concepts, and (521), where they speak of “thick-ish concepts”; Väyrynen (2011: 5
n. 15). Eklund (2011: 32) is much more explicit here: “A fully general account of thick concepts may
be a chimera.” But these authors do not pursue the point.
14 For a comprehensive discussion of pejoratives, and for many references, see Hom (2010).
15 One of Williams’s original examples. See Williams (1985: 140).
16 This example is due to Väyrynen (2012).
17 As a matter of fact, there are even more options than what our discussion in the text indicates. We
have assumed in the text that some descriptive content is a part of the semantic value of thick concepts,
and have enumerated some semantic and pragmatic options in specifying where the evaluative or
normative content belongs. But bearing in mind that there is a spectrum, rather than a dichotomy, of
thick and thin concepts, we can notice that at least for those concepts at the thinner end of the
This non-uniformity point is an important one—both in general, and for our

subsequent discussion of the role of thick concepts in legal and meta-legal state-
ments—and it is important not to let the possibility of stipulation confuse us here.
People are of course at liberty to stipulate a meaning for the technical term “thick
concepts” as they see fit. Nothing can hold them back from deciding to restrict their
use of the term to just those concepts that possess as a part of their very semantic
value an evaluative element. But here, as elsewhere, such stipulation will not take
the place of a theory—either philosophical, linguistic, or both. Natural-language
expressions do not come wearing a stipulated technical term on their sleeves, and so
given a natural-language expression that seems to have an evaluative “flavor,” we
still have to ask whether the relevant phenomenon reflects a semantic or pragmatic
feature, if pragmatic what kind of pragmatic feature, etc. How we decide to label
the concepts picked out by those natural-language expressions for which the
answers to these questions are of a certain kind is of no real interest.
For this reason, insisting that real thick-concept words have an evaluative
element to their very semantic value seems in general pointless. The most natural
way of understanding such talk is as a stipulation of a meaning for the technical
term “thick concept,” and as such neither objectionable nor interesting. Now, there
may be some ways of understanding such insistence as philosophically substantive.
One such way would be to understand it as a claim about certain paradigmatic
examples: We have a list of paradigmatic thick concepts that we can extract from
the literature, and claiming about them that their semantic value includes an
evaluative component is a substantive (and possibly false) claim. Another possible
way of understanding the insistence is by understanding it as making a possibility
point: True, one may claim, the nature or meaning of each natural-language
expression has to be determined piecemeal; but it is at least possible that some
such terms contain an evaluative element in addition to a descriptive one as a part of
their semantic value. Thus understood, this claim is meant to negate the rather
extreme view that general reasons in semantic theory make it impossible for any
term to contain both these elements as a part of its semantic value.18 But thus
understood it is not a claim about the nature of any specific purportedly thick
concept or thick term. So there is little interest in talk of what it takes for a concept
or term to be really thick.
Now, this is not to deny that some interesting and theoretically significant
distinctions among evaluative concepts and terms could be drawn. Let us here
mention two such distinctions, which will prove important later on. First, we may
spectrum, the evaluative content is clearly a part of the semantic value of the relevant concepts, whereas
there are various semantic and pragmatic options in locating the descriptive content.
18 We are not sure, but at least one way of understanding Williamson (2003) is as putting forward
this general claim. The point in the text here is closely related to the controversy over separablism, or
that between entanglers and disentanglers—namely, the controversy regarding the possibility of
separating the descriptive and the evaluative parts of the content of thick concepts. If such separation
is possible, then presumably thick concepts with an evaluative part to their very content are not in
principle more mysterious than conjunctions one conjunct of which is evaluative. See McDowell
(1981); Williams (1985: 130, 141–5).
want to distinguish between expressions that do and those that do not admit of an
intelligible not-one-of-my-words response. Whereas “fashionable” seems to belong
in the first category, “physically strenuous” obviously belongs in the second.
Second, we may want to delineate all and only those concepts mastering of
which involves understanding their evaluative uses. “Physically strenuous” does
not belong to this special category, whereas “fashionable” seems at least a borderline
case. As a test for thickness, neither distinction is more than a stipulation. While
the phenomena that the two distinctions bring to our attention are interesting, the
relevant stipulations are not.
Be that as it may, when it comes to an attempt to better understand a specific
natural-language expression and the concept it presumably expresses, we cannot
rely on general claims such as that real thick concepts contain an evaluative element
as a part of their very content. Rather, the thing to do is to carefully study the
specific expression and concept, and apply general tests for the distinction between
semantic and pragmatic elements, between what is being asserted and what is being
presupposed or implicated, and so on.
3. A neglected option: legal as a thick concept

Our view is that many of the crucial legal concepts that our legal judgments deploy
are thick concepts. These arguably include concepts like crime, tortious, consti-
tutional, contract, inheritance, infringement, etc. One way of making progress here
would be to discuss these (and many other) legal concepts piecemeal, checking
whether each one is plausibly considered a thick concept. Rather than doing that,
we will concentrate here just on one key concept—that of legality. It is, of course, a
central case, and a more general one than the aforementioned ones, and its nature
may have implications for the nature of those other concepts. Furthermore, it is of
central jurisprudential importance. We are going to argue that legal is illuminatingly
conceived as a thick concept. If our case for this particular concept is persuasive,
then we hope, in view of the representative status of this concept and its centrality,
that the case for conceiving other legal concepts as thick concepts will be seen as
promising and worth pursuing.19
We suggest, then, that “legal” be thought of as a thick term, and the concept legal
as a thick concept. The concept’s descriptive content can then be understood in
terms of representations of some social facts—i.e. the social facts in virtue of which
some act or practice type counts as legal or illegal. These facts would differ from one
jurisdiction to the next, but they may be uniform across jurisdictions on sufficiently
high levels of abstraction. But as with other thick concepts, that it has descriptive
19 In some ways, we would have preferred to concentrate on the concept legally valid rather than on
legal. But as far as we observe, the term “legally valid” rarely occurs in the ordinary first-order legal
discourse. Legal scholars, and especially legal philosophers, often talk of legal validity, and attribute this
status to norms, rather than to acts or practices. But in the ordinary first-order legal discourse, we more
often talk of what is “legal” or “illegal” than of what is “legally valid” and “legally invalid.”
content does not preclude its being evaluative as well.20 Indeed, often, or perhaps
even necessarily, declaring an act legal (or illegal) would involve an expression of
some evaluative or normative commitment. There may be an underlying normative
judgment involved here—perhaps something to the effect that the fact that an act
satisfies the descriptive criteria for legality is a reason for certain officials to permit
them, or perhaps to the effect that the fact that an act fails to satisfy these descriptive
criteria counts strongly against these officials permitting it.
Think again about the common observation we started with—namely, that legal
statements seem to walk a fine line between the descriptive and the normative. If
legal and cognate concepts are thick concepts, then the descriptive contents of legal
statements can be very easily accommodated. The relation between an act’s being
legal and its satisfying the relevant descriptive criterion would then be similar to, for
example: the relation between an act’s being courageous and its involving overcom-
ing fear; the relation between an act’s constituting a lie and its involving the
intentional telling of something believed to be false; the relation between a suit’s
being fashionable and its being favored by a certain group of people who set fashion
trends and to whom others regularly defer in matters of fashion style; etc.
Where things get interesting is in accommodating the more normative (or
normative-ish) features of legal statements. Think, for instance, of the nature of
legal disagreements. As Dworkin (1986: ch. 1) famously argued, at least some legal
disagreements seem to resemble normative disagreements, where reasonable and
unconfused people could differ even after all the non-normative, factual issues—
e.g. about how certain people have voted, what kinds of conventions prevail among
certain people, etc.—have been settled or agreed to. A disagreement between two
judges about the legality of some act seems very similar in these respects to a
disagreement between two people about whether a certain act was courageous,
rude, or blasphemous.21 Think also about the fact that oftentimes an assertion that
an act is legal amounts to a kind of endorsement, and the fact that asserting that a
certain act is illegal amounts to a kind of criticism. Here again, the similarities to the
use made of some thick concepts are quite telling. At least in some circles, an
assertion that a suit is fashionable amounts to an endorsement of sorts; and, perhaps
in some circles, saying that a dress is immodest amounts to a criticism of sorts. And
think about the relation of such assessments to reasons, rights, duties, and the like.
Here too, we believe, similarities are striking. For when someone notes that a suit
you are eyeing for purchase is unfashionable (or that it is “so ’80s”!), that speaker
seems to be also asserting or implying that you have a reason not to buy it—
20 Much of the literature on thick concepts is obsessed with the question whether the descriptive
and the evaluative “components” of the content of thick concepts can be, in some sense, separated. For
one recent discussion, and for many references, see Elstein and Hurka (2009). Fortunately, we think
we can remain neutral on this issue here. When we talk in the text about the descriptive “part” and the
evaluative “part,” we do not mean to suggest that the two can be neatly separated. We briefly get back
to the issue between separabilists and non-separabilists (or entanglers and disentanglers) in section 6.
21 At the same time, the disagreement here seems less open-ended than disagreements about the
application of thin concepts. Indeed, this was a part of Williams’s (1985: 141) original characterization
of the thick.
certainly a fashion reason, but perhaps a real reason as well. And when someone says
that not speaking up will be cowardly, she may very well be saying—or implying, or
implicating, or some such—that shutting up should be avoided, that you have even
a decisive reason to speak up.
4. Jurisprudential payoffs: widening the range of options

The observation that legal is a thick concept will not, all by itself, determine the
conclusion of the debate over the nature of law. But we believe that it can facilitate
that discussion by placing it in a wider philosophical context, and also by showing
that the legal case is not that special or unique after all. (Come to think of it, would
it not be a bit surprising if it were?) According to the standard way of formulating
the jurisprudential debate about the nature of law, roughly, legal positivists are
supposed to say that what the law is ultimately is purely a matter of some facts
about the behavior and psychological attitudes of a group of people; and natural
law theorists are supposed to say that what the law is is ultimately a matter of
evaluative or normative, or more specifically moral, considerations.22 Legal positiv-
ists would naturally be inclined to conceive legal statements as factual statements
that represent some behavioral-cum-psychological states of affairs. Natural law
theorists on the other hand would take the view that legal statements are assess-
ments whose truth or correctness ultimately depends on certain evaluative or
normative considerations.
The first way in which the construal of legal as a thick concept may promote the
discussion about the nature of law is by showing that there are more options here
than we have hitherto thought. Thus, we can view legal as a thick concept where the
evaluative element is thought of as a part of its very semantic content, or as a matter
of pragmatics.23 In the latter case, the relevant evaluative content may be a matter
of presupposition, or of conversational implicature. Or it may be neither of those,
so that the evaluative “flavor” of legal may be a matter of weaker, even more
context-sensitive matters (perhaps like the case of physically strenuous). One of the
focal points of the philosophical debate about thick concepts has been about what
could be called “separability” or the possibility of “disentanglement.” Briefly, this is
the possibility of analyzing the meaning of a thick term into its descriptive and
evaluative components, so that we could identify the term’s extension using only
the descriptive component. Some of the main instigators of the debate about thick
terms have denied such a possibility,24 whereas others allow for it. We here refrain
from taking a position on this thorny issue. The descriptive and evaluative contents
of legal may be separable or disentanglable, or they may not be. And if thick
22 One of us takes issue with this standard way of conceiving the debate. See Toh (MS).
23 As indicated in n. 17, it could be the case that it is the descriptive content that should be
explained pragmatically. There are some indications that Hart was of this opinion, as one of us has
pointed out previously. See Toh (2005: 85–8).
24 See e.g. McDowell (1981: 144); Williams (1985: 130, 141–2, 217–18 n. 7).
concepts incorporate in some way (one of the ones stated above, or perhaps some
other way) an underlying normative judgment, then we can ask what exactly is the
content of the normative judgment underlying legal. In the end, it seems to us, it is
of little consequence which options on this now-richer menu get classified as species
of legal positivism, and which others get classified as natural law positions.
In these ways, then, a better understanding of legal statements may depend on
the true nature of thick concepts. For instance, if separabilists or disentanglers have
it right in general, so that at least in principle it should always be possible to
disentangle the normative and the descriptive elements of thick concepts, then this
will apply to the case of legal as well. And if a presuppositional account of thick
concepts in general wins the day, this too will apply as a particular instance to legal,
and this last result would presumably argue in favor of legal positivism. But let us
remind the reader once more of the non-uniformity point. So if, for instance,
disentanglers are wrong with regard to some thick concepts, the next question to
ask is whether in the specific case of legal, disentangling is possible. And similarly, the
questions that it would be useful to ask with regard to the location-of-evaluation
problem should be those specific to legal, rather than to thick concepts more
generally.
5. Jurisprudential payoffs: domesticating and disciplining

some influential jurisprudential themes
In the foregoing ways, then, conceiving legal as a thick concept seems both natural
and potentially illuminating and theoretically productive.25 But things get more
interesting still when we consider the following set of influential, but also (to many)
somewhat obscure, related themes in twentieth-century analytic jurisprudence:
(i) Hart’s distinction between legal judgments from the internal point of view
and legal judgments from the external point of view, and the corresponding
distinction between internal and external statements; (ii) Dworkin’s claim that
determinations of what the law is necessarily require the determining persons to
take up a view about “the point” or “justifying purpose” of the law; (iii) the
judgments from “the legal point of view” that Raz posited and highlighted; and
(iv) the notion of basic norm that played a crucial role in Kelsen’s jurisprudential
theorizing. By lifting these themes from what we deem the hitherto rather parochial
discussion in legal philosophy, and relocating them in the wider context of the
discussion of thick concepts, we can dispel some of the obscurity surrounding
them, and ultimately domesticate them, so to speak. We can draw—at the very
least, by way of an analogy—on the discussions in the thick-concepts literature, in
order to appreciate what would and what would not be relevant considerations in
25 One may worry that when it comes to the debate about the nature of law, this amounts not so
much to a solution as to giving a new name to the problem. Admittedly, a part of what we are doing
here is to reframe old questions. But, as we hope our discussion shows, such reframing has theoretical
payoffs.
specifying and assessing each of these themes. We stress once again that our main
interest here is not exegesis or the history of ideas. We merely want to note how
deploying the notion of thick concepts could help us to reconstruct—that is,
rationally reconstruct—and discipline, in a theoretically productive way, at least
some things these philosophers say.
Hart famously distinguished between two kinds of legal judgments, and corres-
pondingly two kinds of legal statements.26 The so-called internal legal judgments
are those made from the point of view, which Hart called “the internal point of
view,” of legal officials and others who are similarly committed to the laws of their
legal system. As Hart explained them, these contain an element of endorsement.27
External legal judgments, on the other hand, are the judgments of observers
external to the legal systems that they study (e.g. an anthropologist), who take
acts as legal, without thereby in any way endorsing them, and take certain acts as
illegal without thereby in any way criticizing them. The first thing to note about
this distinction—which, despite all the criticism it has elicited,28 still seems to
capture something in the relevant phenomenology here—is that similar distinctions
could be drawn with regard to judgments deploying many thick concepts. Thus, we
can think of judgments about what is and what is not fashionable made by a
participant (for whom proclaiming a suit fashionable amounts to some kind of
endorsement), and also by a fashion historian (who refrains from endorsing any
standards of fashionableness). The relation between these judgments seems to
resemble the relation between internal and external legal judgments that Hart
posited and deployed.
The two of us are not of one mind about the accuracy of the distinction between
internal and external legal judgments as Hart draws it. One of us (Enoch) is
inclined to view Hart’s distinction as ill-drawn, for it seems to find an ambiguity
where none exists. If someone says that in our legal system ç-ing is legal, we do not,
it seems, need any further information about the desirability or advisability of ç-ing
in order to determine the truth value of what she said. And if a well-known outsider
tells us that ç-ing is legal, we the insiders, it seems, can without mistake report what
he said in indirect speech, by saying “He said that ç-ing is legal.” Furthermore, if we
know that he is reliable on such matters, we can even infer from this “ç-ing is legal,”
which from our mouths presumably is an internal legal statement. No equivocation
seems to be involved in this inference. So it seems difficult to maintain that internal
and external legal statements differ in meaning.29 Still, to repeat, something of
importance seems to be captured by Hart’s distinction, and there seems to be some
important distinction between insiders’ and outsiders’ uses of “legal.” What, then,
are we to say here?
26 Hart (1961/94: e.g. 56–7, 88–91, 102–3). See also Bulygin (1982).
27 For an attempt to take this element of endorsement seriously, and then to read Hart as an
expressivist at least about this element, see Toh (2005).
28 See e.g. Raz (1981); Postema (1982: 170).
29 This, we take it, is the crux of Raz’s criticism of Hart’s distinction.
The other of us (Toh) is somewhat more sanguine about Hart’s distinction

between internal and external legal judgments, and the corresponding distinction
between internal and external legal statements. It seems quite true that people who
are committed to the laws of their legal system and those who are disaffected or
uncommitted may hold a conversation about whether a certain act is legal, without
thereby equivocating and talking past each other. But what this data may motivate
is not so much the conclusion that (what Hart calls) internal and external legal
statements have the same meaning, but instead that people (or more specifically,
participants) may utter internal legal statements and thereby display their endorse-
ments or commitments either sincerely or insincerely.30 If we go with the latter
position, then we would also be opting for the view that while inferences to and
from the contents of internal legal statements, whether sincerely or insincerely
uttered, are perfectly licit, inferences between the contents of internal legal state-
ments and external legal statements would involve illicit equivocation. Such bifur-
cated construal of the legal discourse, it may be thought with some plausibility,
better coheres with the phenomenon of post-factual legal disagreements that we
took notice of before, than does a unitary construal that treats legal statements in
general as having the same meaning that is exhaustively descriptive or factual.31
Whichever approach we take towards Hart’s distinction, it seems, we can take
lessons from the phenomenon of thick concepts. Take a conversation between a
fashion insider and a fashion historian about whether something is fashionable. It is
plausible to think that the two can engage each other in a meaningful conversation
without talking past each other, and this despite the fact that there is something
importantly different in the ways the two are using the term “fashionable.” One
natural way of understanding what is going on here is that the fashion insider is
committed to the underlying normative judgment which renders fashionable a thick
concept—roughly, that the fact that an article of clothing satisfies certain highly
complex popularity criteria among some set of trend-setters counts in favor of
wearing it—whereas the fashion historian does not accept (and may even reject)
this judgment. But, to continue with the conjecture, an acceptance of this judg-
ment is not a part of what is being said or even implied when the fashion insider says
“This suit is so fashionable;” instead, the underlying normative judgment is being
assumed or presupposed.32 And the fashion historian can competently participate in
the discussion about what is and what is not fashionable (here and now) while
refraining from a full or sincere commitment to what is presupposed.
An alternative, also natural, way of understanding what is going on is that a
genuine, meaningful conversation is accomplished by the fashion historian’s tem-
porary and pretended commitment to the relevant normative or evaluation judg-
ment, where this judgment is “located” in what he says or asserts, rather than in
30 This is another lesson one may take away from Raz’s discussion of “detached” legal statements.
31 The bifurcated treatment would not be too different from Hare’s treatment of “good gothic
revival.” A speaker who uses the term in an “inverted commas” sense would actually be talking at cross
purposes with a speaker who uses the term in the fully committed sense. See Hare (1952: 124–6).
32 As Stalnaker (1978: 84) points out, a speaker need not believe what he presupposes; he may
presuppose something that he finds convenient to assume for the sake of a particular conversation.
what he presupposes. In other words, following some theorists of thick concepts,
the evaluative or normative commitment could be treated semantically, rather than
pragmatically as in the explanation of the preceding paragraph. If we opt for this
semantic explanation, then it would mean that a genuine, meaningful conversation
about what is fashionable would require outsiders like fashion historians to pretend
or simulate the relevant normative or evaluative commitment in a more thorough-
going way than what such a conversation requires according to the alternative
pragmatic explanation.33 And unless such a thoroughgoing pretension or simula-
tion by outsiders is involved, the two people would be talking past each other.
In either case, we could say that the fashion outsider would have to conduct the
conversation from the point of view of the fashion insiders. According to either
explanation, there is indeed an important difference between the ways in which the
fashion insider and the fashion outsider deploy the term “fashionable;” but they
would not necessarily talk past each other. According to the first explanation, the
two speakers use the term with the same semantic value. As for the relevant
presupposition, more than one explanatory option is available.34 If presuppositions
do not require belief or some other belief-like commitment,35 the two speakers
would share the relevant presupposition (that a clothing article’s satisfying the
relevant descriptive criteria counts in favor of wearing it, roughly); but whereas
the insider would believe it, the outsider would merely accept it temporarily for the
conversation at hand. Alternatively, if the presupposition here involves a commit-
ment, the insider and the outsider would not share the relevant presupposition; but,
since they use the term “fashionable” with the same semantic value, they could still
converse usefully, at least to an extent.36 According to the second explanation, the
two speakers normally use the relevant term with different semantic values, but the
outsider temporarily adopts the insider’s semantic value when he enters into a
genuine conversation with the insider.
Two analogous explanations, we submit, are available for the concept legal as
well. And each provides a fairly plausible reading of both Dworkin’s claim that
judgments as to what the law is (and more generally what he calls “constructive
interpretations”) require judging persons to attribute some “point” or “justifying
purpose” to the practices and traditions that make up the legal systems whose laws
33 One conception of simulation that is serviceable here is Yablo’s (2001: 196), who explains:
“Someone is simulating belief that S if although things are in relevant respects as if they believed that S,
when they reflect on the matter they find that they do not believe it; or at least are agnostic on the
matter; or at least do not feel the propriety of their stance to depend on their belief that S if they have
one. They do not believe that S except possibly per accidens.” Yablo attributes this notion to Walton
(1997). One significant implication of this conception of simulation, as Yablo notes (197), is that a
person who is simulating need not be mindful of that fact about himself.
34 In order to choose between them, more would have to be said about presuppositions in general—
more than space allows us here, and frankly more than we are competent to say.
35 As Stalnaker (1978: 84), for one, says.
36 If they reach an agreement that, say, some sunglasses are really fashionable, the insider may then
proceed to say “So of course you should buy them!” At this point, the difference in presupposition may
place a burden on how useful their conversation can remain.
they are interpreting,37 and Raz’s claim that legal judgments are judgments that
take “the legal point of view” whether that point of view is assumed sincerely or
insincerely. Both of these claims arose out of the belief that what the law is could
not be characterized accurately from what Hart called the external point of view, or
the belief that there is distinction to be observed between what the members of a
community think that a practice requires and what it “really requires.”38 Relevant
here is the second of the distinguishing marks of some thick concepts we discussed
at the end of section 239—namely, that understanding or mastering the concept
legal necessarily involves understanding its evaluative content. There is no problem
in understanding what, for example, “physically strenuous” means without know-
ing anything about the fact that in some circles, in some contexts, using this term
amounts to an endorsement. But arguably, one just does not understand the word
“courageous” and its role without understanding the relation that many people
accept between overcoming fear and some form of evaluation. Perhaps one can
use the term “courageous” without engaging in such evaluation, but such a use will
be in some important way parasitic on the more evaluatively charged, internal-
point-of-view-ish, use of “courageous.” The same might be said about the word
“fashionable.”
Dworkin and Raz can be seen as viewing “legal” as more like “courageous” and
“fashionable” than “physically strenuous.” They are of the view that a person who is
after what the law really requires needs to “join in” the practice that he is trying to
understand. In other words, in order to understand fully or master “legal,” and be
able to apply it to novel cases, a person needs to understand and even take up its
evaluative content. But the two philosophers differ significantly in the details as to
what such a joining-in amounts to. The attitudes that the two philosophers deem
necessary for joining in—the “point”-attributing interpretive attitudes and “the
legal point of view,” respectively—could be cashed out in terms of the normative
judgment that underlies the uses of the thick term “legal,” and that commitment
could be characterized either pragmatically or semantically. An insider is someone
who is sincerely and fully committed to the relevant normative judgment. The
outsider, on the other hand, does not have the same normative commitment.
Dworkin seems to think that such an absence would handicap an outsider to
some extent from understanding what the law really requires,40 whereas Raz opines
that the outsider can manage accurate understandings and characterizations of what
the law really requires by taking on the commitment temporarily and for the
purpose of the conversation at hand. Again, such a temporary commitment could
be cashed out pragmatically or semantically.
In order to make further substantial progress, we would need not only to notice
that legal is a thick concept, but we would also need to plug in the true theory of
37 See Dworkin (1986: esp. 47). The same or similar claims are made in e.g. Fuller (1964: ch. 4);
Dworkin (1975: 101–5); Finnis (1980: ch. 1); Murphy (2004).

38 See Dworkin (1986: 64); Raz (1975/90: 175; 1980: 235–6).
39 We turn to the first mark—i.e. the amenability to the not-one-of-my-words defense—in section 6.
40 See Dworkin’s rather awkward and baroque treatment of the question of how the law of immoral
legal systems is to be interpreted. Dworkin (1986: 101–13).

thick concepts, the true answer to the location-of-evaluation question, or more
plausibly the true answer to this general question as applied to the specific case of
the concept legal. Recall the non-uniformity lesson from section 2. There is no
reason to suspect that all natural-language expressions that have some evaluative
flavor to them behave exactly the same way. The first of the two previously
discussed explanations of legal assumes that at least with regard to some thick
concepts, a presuppositional account is plausible, and that in particular, this is so
for the specific case of legal and related concepts. And the way to determine whether
this is so, is to use general tests for presupposition and see whether they apply to the
case at hand.
In the context of a general discussion of thick concepts, Pekka Väyrynen (2012)
recently did just this. Relying on some of the tests for presuppositions commonly
employed by linguists and philosophers of language, Väyrynen argues for a pre-
suppositional account of thick concepts. Though his discussion is much too
ambitious in our estimation—attempting a general, uniform theory where more
piecemeal analyses seem to be in order—he does seem to establish a plausible case
for the more modest claim that at least some thick concepts—his lead example is
that of the concept lewd—are best understood as incorporating such presuppos-
itions. We cannot here give the full details of Väyrynen’s argument. For our
purposes, it is sufficient to note that the tests he uses for presuppositions—the
projection test (roughly, noting that presuppositions survive under many embed-
dings which cancel entailments); the attitude verb test (if you hope that p, and
p presupposes q, it need not be the case that q is a part of what you hope); etc.—
lend credibility to the conclusion that legal too is a thick concept that is amenable to
the presuppositional explanation.
Indeed, when Raz introduces the idea of judgments from the legal point of view,
he does so partly in an attempt to offer an interpretation of what Kelsen means by
the “presupposition” of the basic norm.41 According to Kelsen, the basic norm—
which says something like “One ought to obey the prescriptions of this commu-
nity’s constitution”—has to be in some sense assumed or presupposed by those
who state the contents of the laws of a legal system or by their statements.42
A number of legal philosophers have complained that Kelsen’s language of presup-
position is obscure, misleading, or both.43 We are now in a position to suggest one
possible and natural way of understanding what Kelsen may have been after here.
The basic norm could be construed as the underlying normative judgment for
statements using the thick concept legal (or some closely related concept); and
specific legal statements (like specific statements about what is and what is not
fashionable) may presuppose, rather than assert, that normative judgment. It is then
possible to think of Kelsen as suggesting or gesturing towards the conception of
legal as a thick concept, understood presuppositionally.
41 Raz (1974: 141; 1981: 451), quoting from Kelsen (1960/67: 217–18).
42 Kelsen (1960/67: 201–5).
43 See e.g. Hart (1961/94: 293); Marmor (2009: 159).
Such a presuppositional analysis of “legal” is a natural and non-distorting

precisification of the Kelsen and Raz characterization of legal judgments. And it
would help us adjudicate the above-discussed dispute between Dworkin and Raz
about how fully or accurately uncommitted or disaffected members of a commu-
nity could understand and characterize what the law really requires. Depending on
the contents of the presupposed evaluative judgments, a disaffected outsider may or
may not be able fully to understand and characterize what the law really requires.
For it is at least arguable that even if the outsider can engage in the insider’s
conversation, still the former’s competence with the relevant concept is parasitic on
the latter’s. Perhaps, in order to be competent with the concept legal one must—
even if one does not share the relevant normative or evaluative presupposition—at
least see its point. Perhaps, in other words, the concept legal passes one of the quasi-
tests for thickness mentioned at the end of section 2—namely, the explanatory
priority test—so that engaged insiders’ uses of and competence with the concept
legal are explanatorily prior to the outsider’s use and competence of the same
concept. And perhaps this is at least one of the major things Dworkin is eager to
emphasize with his anti-positivism. With the presuppositional analysis of “legal”
then, we may be able to make a plausible case for Dworkin’s take on the kind of
“joining in” that is necessary to determine what the law is.44 But more generally
and importantly, with the presuppositional treatment of “legal” and its cognate
concepts, we may be able to avoid making bald assertions about what uncommitted
or disaffected members can or cannot do, and to see what issues to zero in on.
All of this was rather quick, and much more work undoubtedly needs to be done
here, for instance to examine the linguistic behavior of “legal” and related expres-
sions. But we believe that enough has been said to show that such an inquiry is
called for, and that there is some likelihood that it will prove theoretically product-
ive. Let us also note that even if the presuppositional account turns out not to fit
very well the concept legal, this in no way would show that it is unhelpful to think
of the nature of legal statements in terms of thick concepts. For one thing, the non-
uniformity point may apply even within the class of legal concepts. It is therefore
quite possible that even if a presuppositional account of legal in particular fails, such
an account of some other key legal concepts will be much more successful.
Furthermore, the failure of a presuppositional account would merely indicate that
a different theory is called for—for thick concepts in general, or much more
plausibly, for the concept legal in particular. An alternative theory that locates the
evaluation in what is said or asserted has already been discussed. And there are
pragmatic accounts other than presuppositional accounts. The lesson remains that
the way—at the very least, a way—to make progress on understanding legal
statements and in particular the way in which they incorporate, if they do, a
44 As Stefan Sciaraffa pointed out to us, Dworkin’s case would be further strengthened if philoso-
phers like McDowell and Williams were right in thinking that the separability or disentanglability of
the descriptive and evaluative components of thick concepts is false, or if separability or disentaglability
is false for the concept legal.
normative element, is by broadening the horizons of those interested in jurispru-
dence and thinking about thick concepts more generally.45
6. Jurisprudential payoffs: a new question?

We have so far argued that the conception of legal as a thick concept could enable us
to reframe in a theoretically productive way some of the themes and questions
discussed in the recent legal philosophical debate about the nature of law. We want
to end the chapter by observing how the conception could serve to highlight
another, related, question, that is seldom discussed.
At the end of section 2, we discussed the two ways of distinguishing among the
concepts involving evaluation that may be deemed partial tests for thickness. In the
preceding section, we discussed one of these—namely, the explanatory priority of
the evaluatively committed usages. In this section, we want to discuss the other—
the not-one-of-my-words response. Some words admit such a response whereas
some do not. We can now ask: Does “legal” admit of such a response? Can we
imagine fairly standard conversational contexts in which, say, someone is asked
“But surely you don’t deny that such an action is legal, do you?” and responds with
“‘Legal’, sir, is not one of my words”? “Blasphemous,” as the original Wilde
example shows, admits of such a response, whereas “physically strenuous” probably
does not. If someone recommends a workout routine by noting that it is physically
strenuous, and if I am actually looking for a relaxing, mild workout, the natural
thing for me to say is not that “physically strenuous” is not one of my expressions,
but rather that I understand that it is physically strenuous, but that I am not just
now interested in anything of that sort. Well then, is legal more like blasphemous or
physically strenuous in this respect? Suppose someone recommends to you an action
by noting that it is required by the law, and suppose you want to resist this
recommendation. Also suppose that you do not deny that the relevant action is
required by a norm that satisfies whatever descriptive criteria are necessary for legal
validity. In such a case, it seems to us, it is much more natural for you to respond
with something like “Yeah, this action is legally required, but I don’t care about
legal requirements right now” than with “‘Legal’ (or ‘legally required’) is not one of
my expressions.” If so, it seems that “legal” and its cognates are in this respect more
like “physically strenuous” than “blasphemous.”
We do not want to overstate this last point. First, in some contexts perhaps a not-
one-of-my-words response would make sense even regarding legal. Perhaps, for
instance, out of the mouth of a political anarchist, in some contexts, it would make
45 Eklund (2011) suggests an account of thick concepts in terms of epistemic analyticity. On his
interesting account, and somewhat roughly, the relevant underlying evaluative judgment is one that it
is part of competence with the relevant thick term to be disposed to accept it, absent some special
reason to deny this analyticity (that is, some reason to think of the relevant thick concept as
objectionable). Applied to legal, we get the interesting question—is it a part of competence with the
term “legal” (and other key legal terms) to accept the inference from “The law requires that you ç” to
“You have a reason to ç,” or some such—absent, that is, some special reason to reject this inference?
perfect sense. And second, even if “legal” does not admit (usually) of a not-one-of-
my-words response, it is not clear what exactly follows from that. Not, as we have
emphasized, that legal is not a thick concept, for it is not clear that the availability of
a not-one-of-my-words response is a make-or-break test for thickness. Perhaps what
would follow is at least some pressure against viewing legal as a thick concept whose
evaluative element is a part of its semantic value. But this, to repeat, is consistent
with it being a thick concept, accounted for pragmatically, perhaps along the
presuppositional lines we have sketched in the preceding section.46
In closing, we want to turn to a deeper issue that is implicated by the not-one-of-
my-words response. Notice that this response is appropriate when a speaker
considers the relevant term bad or objectionable. Some thick concepts are bad or
objectionable,47 and this fact can be easily explained in terms that we have been
using. If thick concepts in some (semantic or pragmatic) way incorporate some
underlying normative judgment, as we have been arguing, and if this judgment is
false, then the relevant thick concept is bad or objectionable, and we all should
usually respond to statements employing the relevant thick term with this-is-not-
one-of-my-words responses. Some racist slurs are rather obvious examples here,48
but presumably there are less offensive ones as well. Blasphemous is one if Wilde was
right. Chaste is probably another, insofar as it incorporates a judgment to the effect
that people (in particular, women) should refrain from sexual activities or expres-
sions, and to the extent that this judgment is false. So for any thick concept, we can
(and perhaps often we morally should49) ask whether it is objectionable, whether it
is a bad thick concept.
Noticing this general fact about thick concepts, and assuming now that legal is a
thick concept, we can now ask: Is legal a bad thick concept? This question has, as far
as we know, not been discussed, perhaps because the relation to other thick
concepts has not been appreciated. But it seems to us that this question is
nonetheless an extremely important one. In order to answer it, we have to come
first to a conclusion regarding the exact normative judgment underlying legal. And
we then have to see whether it is true. And the matter may not be clear-cut. For
instance, assume that the underlying normative judgment is that if an act or
practice meets whatever factual criteria constitute the descriptive element of legal,
then citizens have a reason—a real reason, not merely a legal reason—to carry out
the act or take part in the practice. What would it take for this judgment to be true,
and for legal to be in effect a good or unobjectionable thick concept? Would it be
necessary that whenever an act or practice is legal, whatever the relevant descriptive
46 And again, recall that different accounts may be appropriate for different legal concepts.
47 Eklund (2011) repeatedly emphasizes both that this is so, and that accommodating this is a key
desideratum for a theory of thick concepts.
48 Though, as Eklund (2011) notes several times, it is controversial whether slurs should be
considered thick terms. Perhaps one motivation for thinking otherwise is that in the case of racist
slurs we always (or at least typically) have a more neutral term that captures the descriptive content of
the racist slur. Not so with thick concepts, it may be thought. It is not clear, however, how important
this difference is. For a discussion of racist slurs, and for many references, see Hom (2010).
49 This is the crux of Blackburn’s (1998: 102) objection to Williams’s claims about thick concepts
leading to a “self-serving complacency.”

criteria, there is a reason to carry out the act or take part in the practice? It is very
hard to see how such a strong and general claim could be true.50 Would it suffice
that whenever an act or practice is legal in a specific legal system (this one!), or
perhaps in a non-specific legal system satisfying some criterion (that it is not too
unjust, say) there are reasons to carry out the act or take part in the practice? Or
would it suffice that most of the time there are reasons to act as the laws specify?
There are pressures here in different directions. Consider, for comparison, the
concept patriotic. If the normative judgment underlying its uses is that whenever an
action or an attitude favors one’s country this counts in favor of the action or
attitude, then the judgment is obviously false and the concept quite bad. But still,
many of us, including even those cognizant of the falsity of the general judgment,
are perfectly happy employing the concept in a way that is restricted to just those
cases where the underlying normative judgment is true. So perhaps a similarly
tolerant attitude is called for when it comes to the concept legal. On the other hand,
we may want to think of thick concepts as (purportedly) licensing some inferences, so
that, for instance, the concept boche51 licenses the inference from “Hans is
German” to “Hans is boche,” and from “Hans is boche” to “Hans is cruel,” and
so indirectly, from “Hans is German” to “Hans is cruel.” If this is so, then boche is a
bad, objectionable concept, and it should not be employed even when speaking
about Germans who happen to be cruel.52 So, if legal licenses inferences from
“ç-ing satisfies some descriptive criterion D” to “We should ç” (via “ç-ing is legal”)
perhaps legal is a bad, objectionable thick concept after all, and so it should not be
employed even in cases where we should act according to a norm that satisfies the
relevant descriptive criteria, even when we should precisely because it satisfies them.
So it is not clear what exactly to say about the question whether legal—if it is a
thick concept—is an objectionable one. And we should again note the possibility of
non-uniformity. Perhaps some concepts that are often parts of legal statements are
objectionable and others are not. But it is clear, we think, that this is an important
question that raises a host of related issues that deserve more attention than they
have received. In this way too, we think, thinking of legal and related concepts as
thick concepts could be theoretically productive.
7. Conclusion
In this chapter, we have broached, though we have not established with any
thoroughness, many ways in which we could make progress in jurisprudential
matters by conceiving legal and cognate concepts as thick concepts. Even if this
50 See Enoch (2011) for a discussion of this and some related issues.
51 This is Michael Dummett’s influential example. See Dummett (1975: 454).
52 Here the discussion of thick concepts borders on even more general discussion in the philosophy
of language, in the context of some version of conceptual role semantics. For one such discussion—one
that clearly and explicitly notes that some conceptual roles are “bad,” not just in tonk- or boche-like
cases, but also in cases where the licensed inference preserves necessary and a priori truth (and is still
objectionable)—see Boghossian (2003).
particular proposal by itself would not yield jurisprudential solutions, we surmise, it

would facilitate jurisprudential thinking by widening legal philosophers’ set of
options in pursuing solutions and by informing them of the benefits and costs
associated with analogous moves in the thick concepts literature.
References
Blackburn, S. (1998). Ruling Passions. Oxford: Clarendon Press.
Boghossian, P. (2003). “Blind Reasoning,” Proceedings of the Aristotelian Society, suppl. 77:
225–48.
Bulygin, E. (1982). “Norms, Normative Propositions, and Legal Statements,” in G. Flistad
(ed.), Contemporary Philosophy: A New Survey. The Hague: M. Nijhoff, 3: 127–52.
Dancy, J. (2004). Ethics without Principles. Oxford: Oxford University Press.
Dummett, M. (1975). “Can Analytical Philosophy be Systematic, and Ought it to be?,”
reprinted in M. Dummett, Truth and Other Enigmas. London: Gerald Duckworth & Co.
Ltd., 1978, 437–58.
Dworkin, R. (1975). “Hard Cases,” reprinted in Dworkin, Taking Rights Seriously.
Cambridge, MA: Harvard University Press, 1977, 81–130.
Eklund, M. (2011). “What Are Thick Concepts?,” Canadian Journal of Philosophy 41:
25–50.
Elstein, D. and Hurka, T. (2009). “From Thick to Thin: Two Moral Reduction Plans,”
Canadian Journal of Philosophy, 39: 515–36.
Enoch, D. (2011). “Reason-Giving and the Law,” in L. Green and B. Leiter (eds.), Oxford
Studies in the Philosophy of Law. Oxford: Oxford University Press, 1: 1–38.
Finnis, J. (1980). Natural Law and Natural Rights. Oxford: Clarendon Press.
Foldy, M. (1997). The Trials of Oscar Wilde. New Haven: Yale University Press.
Fuller, L. (1964). The Morality of Law. New Haven: Yale University Press.
Gibbard, A. (2003). “Reasons Thin and Thick,” Journal of Philosophy 100: 288–304.
Hare, R. M. (1952). The Language of Morals. Oxford: Clarendon Press.
Hart, H. L. A. (1961/1994). The Concept of Law. Second edition. Oxford: Clarendon Press.
Heuer, U. (2012). “Thick Concepts and Internal Reasons,” in U. Heuer and G. Lang (eds.),
Luck, Value, and Commitment: Themes from the Ethics of Bernard Williams. Oxford:
Hom, C. (2010). “Pejoratives,” Philosophy Compass, 5: 164–85.
Kelsen, H. (1960/1967). The Pure Theory of Law. Second edition, 1960. Translated by Max
Knight. Berkeley and Los Angeles: University of California Press, 1967.
McDowell, J. (1981). “Non-Cognitivism and Rule-Following,” in S. Holtzman and
C. Leich (eds.), Wittgenstein: To Follow a Rule. London: Routledge & Kegan Paul,
141–62.
McKeever, S. and Ridge, M. (2006). Principled Ethics: Generalism as a Regulative Ideal.
Oxford: Oxford University Press.
Marmor, A. (2009). Social Conventions: From Language to Law. Princeton: Princeton
University Press.
Murphy, M. (2004). “Natural Law Theory,” in M. Golding and W. Edmundson (eds.),
The Blackwell Guide to the Philosophy of Law and Legal Theory. Oxford: Wiley-Blackwell,
15–28.
Postema, G. (1982). “Coordination and Convention at the Foundations of Law,” Journal of
Legal Studies, 11: 165–203.
Raz, J. (1974). “Kelsen’s Doctrine of the Basic Norm,” reprinted in J. Raz, The Authority of
Law. Oxford: Clarendon Press, 1979, 122–45.
Raz, J. (1975/1990). Practical Reason and Norms. Second edition. Princeton: Princeton
University Press.
Raz, J. (1977). “Legal Validity,” reprinted in J. Raz, The Authority of Law. Oxford:
Clarendon Press 1979, 146–59.
Raz, J. (1979). The Authority of Law. Oxford: Clarendon Press.
Raz, J. (1980). “Sources, Normativity, and Individuation,” postscript to Raz, The Concept
of a Legal System. Second edition. Oxford: Clarendon Press, 1970/1980, 209–38.
Raz, J. (1981). “The Purity of the Pure Theory,” Revue international de philosophie, 35:
441–59.
Scheffler, S. (1987). “Morality through Thick and Thin,” Philosophical Review, 96: 411–34.
Schroeder, M. (2009). “A Matter of Principle,” Nous, 43: 568–80.
Stalnaker, R. (1978). “Assertion,” reprinted in Stalnaker, Context and Content. Oxford:
Oxford University Press, 1999, 78–95.
Toh, K. (2005). “Hart’s Expressivism and his Benthamite Project,” Legal Theory, 11:
75–123.
Toh, K. (MS). “Legal Philosophy à la carte.”
Väyrynen, P. (2011). “Thick Concepts and Variability,” Philosopher’s Imprint, 11, <http://
quod.lib.umich.edu/cgi/t/text/text-idx?c=phimp;view=toc;idno=3521354.0011.001>.
Väyrynen, P. (2012). “Thick Concepts: Where’s Evaluation?,” in R. Shafer-Landau (ed.),
Oxford Studies in Metaethics. Oxford: Oxford University Press, 7: 235–70.
Walton, K. (1997). “Spelunking, Simulation, and Slime: On Being Moved by Fiction,”
in M. Hjort and S. Laver (eds.), Emotion and the Arts. Oxford: Oxford University Press,
37–49.
Williams, B. (1985). Ethics and the Limits of Philosophy. Cambridge, MA: Harvard
University Press.
Williams, B. (1995). “Truth in Ethics,” reprinted in Brad Hooker (ed.), Truth in Ethics.
Oxford: Blackwell Publishing, 19–34.
Williamson, T. (2003). “Understanding and Inference,” Proceedings of the Aristotelian
Society, suppl. 77: 249–93.
Yablo, S. (2001). “Go Figure: A Path through Fictionalism,” reprinted in Yablo, Things.
Oxford: Oxford University Press, 2010, 177–99.
12
Making Old Questions New: Legality,
Legal System, and State
Keith Culver and Michael Giudice
From the first steps of what is now regarded as the analytical tradition of legal
theory, from Hobbes to Hart and beyond, analytical legal theorists have long taken
the law of the sovereign state as the primary example of legality.1 Their theories of
the nature of law have been firmly rooted in what Joseph Raz has described as the
“assumption of the importance of municipal law,” which “reflects our, or at least
my, intuitive perception that municipal legal systems are sufficiently important and
sufficiently different from most other normative systems to deserve being studied
for their own sake.”2 On this widely shared view, state legality is sufficiently
prominent and familiar that its correct explanation also serves as a criterion of
adequacy for any general theory of law.
The power of this view rests on much more than the intuitive perception that
modern municipal legal systems are important and distinctive. Post-Westphalian
sovereign states and their operation of systems of laws have in fact played a
dominant role in shaping forms of social life—the organization of political com-
munities, markets, educational practices, and so on. No theoretical study of social
life can ignore the role of the law-state without opening itself to serious objections
as to its completeness. Yet at the same time it is clear that the role of the law-state in
social life can be overstated, or simply given more than due attention at the expense
of an adequately comprehensive and nuanced account of contemporary social life.
In recent years analytical legal theory has been charged with just this sort of
overstatement of the importance of the law-state to characterization of social life.
Legal pluralists such as William Twining and Brian Tamanaha, for example, who
are deeply knowledgeable about analytical jurisprudence, have repeatedly pointed
1 By way of admitting that this is identification of a dominant trend to which there has been some
opposition, it is worth noting Hans Kelsen as an exception. See Kelsen (1966). Also, a note on
terminology: for the most part we will use “legality” instead of “law,” to emphasize the variety of types
of social situation which display elements of legal order or lawfulness. “Law,” we believe, is too strongly
associated with the idea of state law, and so for this reason is undesirable. On the elements of legality
and legal order, see Culver and Giudice (2010: ch. 4).
2 Raz (2009a: 105).
280 Keith Culver and Michael Giudice
out that state legality is simply one variety of legality, and has only ever been one
variety of legality amongst others.3 Twining and Tamanaha both maintain that the
existence of other varieties of legality demonstrates the limited scope of analytical
jurisprudence, and further, shows why its aim to provide a general theory of law is
likely misguided. Where analytical legal theory seeks to explain the essence or
nature of law as an ultimately unitary or monolithic phenomenon intrinsically
linked to the career of the law-state, no such phenomenon exists in law’s actual
habitat. Instead observers find only diversity of phenomena and diverse legal
concepts serving diverse purposes.4 This may be called the legal pluralist challenge
to state-centered theories of legality.5
To date the legal pluralist challenge has elicited little interest and little reaction
from analytical and state-centered legal theorists.6 It might even be wondered
whether there really is any challenge to be met. The debate between state-centered
legal theory and legal pluralism might not really be a debate at all, and might instead
be better read as an awkward comparison of two different kinds of theories
addressing two very different subject-matters. State-centered legal theory, it
might be said, is analytical in method and focused on the most general possible
explanation of state legal systems, even at the occasional cost of declaring some
phenomena to be outliers. Legal pluralism, by contrast, can be said to use analytical
tools of the same sort, yet in its sensitivity to diversity of legal phenomena, legal
pluralism is more content to tolerate conceptual-explanatory tensions while captur-
ing descriptively the widest possible range of diverse manifestations of legality at all
levels of social organization. We may have found a new Janus-faced figure in
jurisprudence: to assert that legal theory captures legality is to say that analytical
theorists look abstractly in one direction, asking how general a theory of munici-
pally originating law can be, while legal pluralists look pointedly in another, asking
what municipally originating legal theory might be overlooking.
Could this really be the end of the matter? Possibly so, but even if this is the end,
such a conclusion can only be asserted once the legal pluralist challenge has been
3 See Twining (2000 and 2009) and Tamanaha (2001 and 2008).
4 Tamanaha writes: “The long history of failed attempts at articulating an essentialist concept of law
should be taken as instructive—there is something wrong with the ways in which the question of what
law is has been posed and answered. The source of the intractable difficulty lies in the fact that law is a
thoroughly cultural construct. What law is and what law does cannot be captured in any single concept,
or by any single definition. Law is whatever we attach the label law to, and we have attached it to a
variety of multifaceted, multifunctional phenomena: natural law, international law, state law, religious
law, and customary law on the general level, and an almost infinite variety on the specific level, from lex
mercatoria to the state law of Massachusetts and the law of the Barotse, from the law of Nazi Germany
to the Nuremberg Trials, to the Universal Declaration of Human Rights and the International Court
of Justice. Despite the shared label ‘law’, these are diverse phenomena, not variations of a single
phenomenon, and each one of these does many different things and/or is used to do many things.
There is no law is . . . ; there are these kinds of law and those kinds of law; there are these phenomena
called law and those phenomena called law; there are these manifestations of law and those manifest-
ations of law. No wonder, then, that the multitude of concepts of law circulating in the literature have
failed to capture the essence of law—it has no essence . . .” (Tamanaha (2001: 193)).
5 For other types of challenges to essentialist theorizing about law, see Priel (2007) and Leiter
(2003).
6 For a recent exception, see Waluchow (2010).
Making Old Questions New 281
taken seriously. Our investigation will show that the Janus-faced image is quite
misleading: there is in fact much to gain from close comparison of legal pluralism
with state-centered analytical legal theory. The comparison shows more points of
common orientation than might be initially supposed, and more importantly for
our purposes as defenders of a variety of analytical approach, the comparison shows
that the presumptive centrality of state legality to analytical jurisprudence ought to
be reconsidered, and can be reconsidered without damage to the enduring insights
of analytical legal theory.
Here is our plan. First, we shall identify three characteristic features of the core
commitments of state-centered legal theory and contrast those commitments with
the legal pluralist challenge to them. The second part of our chapter examines the
ways in which the legal pluralist challenge—regardless of its independent merits as a
theory of law—shows that state-centered legal theory is incomplete on its own
terms. We intend to show in particular that state-centered legal theory lacks a
satisfactory solution to the problem of the continuity of legal systems because it
lacks an adequate explanation of the relation between law and state. Third, we will
then take some initial steps toward specification of the characteristics of an adequate
account of the relation between law and state. Here legal pluralist insights are most
relevant, and show most clearly the limitations of state-centered legal theory. In the
fourth part of our discussion we shall identify areas of remaining disagreement
between state-centered legal theory and legal pluralism, arguing that the challenge
of legal pluralism can be taken seriously even without accepting all of its insights. In
particular, we shall argue that even if legal pluralists are ultimately correct in their
contention that law does not have an essence, that conclusion requires more
support than is provided by the legal pluralist challenge. In the fifth and final
part of our discussion we shall attempt to make good on the promise identified in
the title of our chapter, showing how some old jurisprudential questions can be
made new for the next set of challenges we believe analytical jurisprudence must
face.
1. State-centered legal theory and the legal pluralist challenge
State-centered theorizing about law embodies three main features. The first is so
familiar that it risks being regarded as beneath analytical notice: it is the view that
the legality of the law-state is a systemic legality. That is, there are no free-floating
state laws: all laws in the law-state are members of its system of laws. While there are
important variations amongst theorists regarded as members of the analytical camp,
all of Bentham, Austin, Kelsen, Hart, and Raz agree that legality as found in the
law-state is systemic. Raz’s expression of this view is particularly instructive. Raz
suggests that state legal systems can be identified in terms of three characteristic
claims they exhibit: comprehensiveness, supremacy, and openness.7 In attributing
7 Raz (2009a: 116–20).

claims to legal systems, Raz is not simply personifying what are in fact better
regarded as emergent properties of the complex collection of norms and institutions
giving rise to a legal system. Rather, on Raz’s view, the claims of legal systems are
found in the attitudes and practices of a class of law-applying officials and insti-
tutions. So the systemic claims of comprehensiveness, supremacy, and openness are
epiphenomena of the attitudes and practices of law-applying officials and insti-
tutions, detectable in their operation of institutions in application of laws. In
exhibiting a claim of comprehensiveness a legal system (or more properly, its officials
in action) claims authority to regulate all aspects of social life within the self-
declared boundaries of that system, whether by prohibition, requirement, or
permission. Legal systems differ in this way from other normative systems, such
as sports associations and universities, which have limited purposes and so limited
spheres of regulation within their broader social situation and legal system. A claim
to supremacy can be seen in a legal system’s self-understanding that its standards
provide the ultimate foundation or authority for all other normative standards
recognized by the system. In states with a written constitution, the hierarchy of
rules and procedures which embodies a legal system’s claim to supremacy is readily
apparent: constitutions typically claim to be the supreme law of the state, and
authorize and empower institutions and officials to create, vary, and extinguish
laws. A claim to supremacy can also be detected in the way a legal system views
conflicts between its legal norms and other types of normative standards. Where
legal norms conflict with religious norms or moral norms, for example, legal norms
hold priority unless otherwise stipulated by the legal system, whose authority over
such special stipulations preserves and demonstrates its claim to supremacy. Finally,
a legal system typically claims to be open: the norms of other systems—e.g. sports
leagues, trade unions, universities, etc.—can be supported by legal systems and
“adopted” or certified as permitted to operate in certain ways within a given legal
system. Examples of such certification include use of legislation to establish subse-
quently largely self-governing institutions such as universities or professional
bodies. Examples of adoption are most typically seen when norms of other systems
are applied by courts to resolve disputes arising within their own systems.8 Taken
together, the features of comprehensiveness, supremacy, and openness serve well to
explain the way in which, as Raz puts it, “ . . . we feel that legal systems not only
happen to be the most important institutionalized system governing human
society, but that that is part of their nature.”9
The state-centered view’s identification of the law-state with systemic legality
does not, as the “openness” claim demonstrates, preclude the state-centered view
from forming further views about non-state forms of legality which might exist
8 Raz (2009a: 120). Raz also views conflict of laws cases in terms of the relation of adoption, in
which domestic courts apply the norms of a foreign system to resolve a dispute, but without the foreign
norms thereby becoming part of the domestic system.
9 Raz (2009a: 116). Here we are only presenting Raz’s view of state law, and we ought to note that
we do not think it is true that all state legal systems exhibit claims of comprehensiveness, supremacy,
and openness. Federal systems, for example, pose a particular problem. See Culver and Giudice (2010:
ch. 2).
within, between, or across states. Such views comprise the second main feature of
state-centered legal theory as it views all other forms of legality from the perspective
of the presumptively basic form and perspective of state legality. This view typically
results in either an adverse comparison identifying non-state legality as deficient in
one way or another, or revelation that what appears to be non-state legality is in fact
ultimately authorized by and so a creature of state legality. For example, both Hart
and Kelsen thought that international legality was a defective kind of state legality.
In Hart’s view, international legality lacked a system-constituting rule of recogni-
tion, and so amounted to a primitive or non-systemic form of legality. Kelsen also
thought that international legality was primitive, though mainly because it lacked
effective centralized institutions of enforcement. Yet this assessment of inter-
national legality did not rule out the possibility, via the openness characteristic of
the law-state’s systemic legality, that at least some of what is called international law
might be or become legal via support in the form of state practice or recognition.10
For example, there seems to be nothing in Hart’s account of a rule of recognition
which would rule out the possibility that a state legal system’s rule of recognition
might come to include religious, international, transnational, religious, Aboriginal,
or customary legality among its authoritative sources of law.11
The third feature of state-centered legal theory is its view, manifested in the
tendencies of state-centered legal theorists, that state legality is not just sufficiently
important to warrant study, but so central that study and explanation of inherently
systemic state legality amounts to study and explanation of the nature or essence of
law full stop. This view is readily apparent in the views of Austin, Hart, and Raz.
In The Province of Jurisprudence Determined, for example, Austin proceeds by
narrowing the class of true laws12 (laws “properly so called”) until he settles on
the theory that only commands issued by state sovereigns are laws. And where
Austin thought that he had found the “key to the science of jurisprudence” in the
idea of a command, Hart sought to show that the real “key” lies in the union of
primary and secondary rules, amongst which there is a single rule of recognition
which validates all the norms of the legal system and claims supremacy.13 Notice
that neither Austin nor Hart specified or qualified their claims as claims to have to
found the key to the science of jurisprudence of state legality; they appear content
for their views to be understood as scoping over all legality. It remains possible, of
course, to suppose on a charitable reading of Austin, Hart, and Raz that they have
simply chosen to focus exclusively on state legality without much emphasis of
the fact of that choice. These choices have nonetheless had the tendency of
10 Roger Cotterrell puts the assumption like this: “ . . . all law is created by the state or derives
authority and validity from its recognition by agencies of the state, and that its guarantee is its
enforceability by the state.” Cotterrell (2008: 1, 10).
11 See MacCormick (1999).
12 Sometimes with surprising results. Constitutional law, for example, was only really positive
morality. See Austin (2000: 253–61). On this score Austin’s view has been echoed more recently by
theorists wishing to deny legal status to negotiated or cooperative orders. See Roberts (2005).
13 Hart (1994: chs. 5 and 6).
overreaching and turning into claims about the nature or essence of law,14 or what
might be overly quick dismissal of those who aim to take non-state legal phenom-
ena seriously. It is precisely this combination of state-based overreaching and
under-sensitivity to additional prima facie legal phenomena that legal pluralists
contend has severely limited the scope of analytical jurisprudence.
It may seem from the preceding discussion that exactly those aspects of the state-
centered analytical view of legality which enable it to characterize systemic yet
absorptive state legality are at the same time weaknesses of the view in the event that
it can be shown that there really are non-state legal phenomena whose legality
cannot plausibly be attributed to state-sponsored certification, temporary adoption,
or similar processes. Legal pluralists assert just this challenge, contending that
state-centered analytical legal theory proceeds with two crucial—and ultimately
mistaken—assumptions in the construction of general theories of law: (i) that all
law is either contained in or traceable back to an authorizing law-state;15 and
(ii) that for the purposes of analysis, states are stable and their legal systems have
meaningful boundaries, which is to say that legal norms operating in a given legal
system are only those authorized by the given state legal system. As legal pluralists
attempt to demonstrate with observationally available evidence, both recent and
historical phenomena cast significant doubt on these assumptions.
Since the emergence of legal pluralism as a theoretical approach in the hands of
anthropologists studying social life in colonial and post-colonial states, it has been
widely adopted by socio-legal theorists to explain law within and across states in a
variety of contexts. Indeed, legal pluralism is used to explain the diversity and
interaction of all forms and levels of social order.16 Amongst legal pluralists,
Twining and Tamanaha are perhaps most relevant to investigation of the possibly
Janus-faced relation between legal pluralism and analytical jurisprudence, in sig-
nificant part because both begin from and are well aware of the explanatory merits
of analytical approaches. Two main features of their views are useful to understand-
ing the contrasting approaches of legal pluralism and analytical jurisprudence. Both
Twining and Tamanaha maintain (i) that a general jurisprudence, to be genuinely
general, must take as its subject matter legality wherever and whenever it exists, and
since state legality is only one kind of legality, it cannot be given any special
theoretical priority. In this way, both Twining and Tamanaha are pluralists about
the sources or types of legality. Twining and Tamanaha also both maintain (ii) that the
diversity of sources and types of legality makes it impossible or at least unproductive
to settle on an essence or nature of law: legality’s manifestations are simply too
diverse.17 In this way, both Twining and Tamanaha are pluralists about the concept
of law itself.
14 See Tamanaha (2001: 146) on “analytical imperialism.”

15 This assumption goes hand in hand with the assumption that states divide the normative order of
the world without remainder, and the view that state legal systems necessarily claim supremacy.
16 For a recent and useful overview of legal pluralist literature see Tamanaha (2008).
17 See Twining (2009: 117) and Tamanaha (2001).
As we said in our introduction, it is relatively easy at this point to see how the
conversation between state-centered legal theorists and legal pluralists might come
to an end and settle into an agreement to cohabit within the metaphor of a new
kind of Janus-faced jurisprudence. In this cohabitation, state-centered legal theor-
ists might insist that state legality is still very much prominent, yet still incompletely
understood so warranting the continued centrality of the law-state in legal theory
(though it might be more accurate to call itself a theory of state legality18). Legal
pluralists in this context might be said to simply exhibit a different preference,
maintaining that all forms of legality ought to be explored in a general jurispru-
dence. Each approach’s conclusions about the nature of legality would of course be
very different, but only because the two types of theories began with very different
aims and subject matters.
Our central concern here is not with the possibility of a genuinely general theory
of law, so we shall leave that question to one side as we consider more thoroughly an
avenue of investigation opened by consideration of the metaphor of a Janus-faced
discipline of jurisprudence. Here we shall investigate whether the legal pluralist
challenge to the analytical assumption of the fundamentally systemic nature of
legality raises an explanatory challenge for analytical theory, as it encounters the
question of continuity of legal systems over time. We take up this question in part
because legal pluralists have drawn attention to a tension in analytical approaches as
they attempt to reconcile the apparently closed nature of legal systems with
conditional openness to absorption or adoption of norms from without, and in
part because analytical attention to the phenomena driving this explanation and
tension take the analytical approach closer to legal pluralists’ preoccupations with
consideration, identification, and characterization of diverse forms of legality. The
results of the investigation are of immediate relevance to the fate of the Janus-faced
standoff between rival approaches to characterization of legality, and eventually, to
the possibility of a truly general jurisprudence.
2. The problem of continuity of legal systems
As we noted earlier, one of the key features of state-centered legal theory is the view
that state legality is always and everywhere systemic. Analytical theorists have
developed increasingly sophisticated explanatory devices to capture the systemic
nature of law, deploying the ideas of sovereignty, basic norm, rule of recognition,
and comprehensiveness, supremacy, and openness. These devices are well suited to
static legal systems or those made momentarily stationary for purposes of analysis,
enabling explanation of how it is that, at some given moment, the norms and laws
of a particular state form one legal system. The explanatory formulae are familiar:
norms and laws form one legal system if they are all issued as commands by the
sovereign, or authorized by a single basic norm, or validated by a single rule of
18 See Tamanaha (2001).

recognition, or consist of all those standards law-applying institutions are under a
duty to apply as standards of the comprehensive, supreme, and open legal system.
These explanations are not, however, complete explanations, to the extent that
they capture legal systems frozen in time for analysis. The second part of the
explanatory task arrives with the question of how to explain the continuity, or
dynamic nature of state legal systems. This is a multi-part question of how to
identify and explain the conditions under which a legal system comes into exist-
ence, persists, and ceases to exist as it is displaced by another, falls prey to natural
disaster, or another of the perils of social life. As Raz observed in 1998, the problem
of continuity remains among the unsolved problems of analytical jurisprudence,19
but not only that, there has been little interest in bridging the gap between
explanations of systematic legality in static and dynamic forms.20 As Raz suggests,
interest has simply “shifted elsewhere.”21 This lack of interest is somewhat remark-
able given widespread acknowledgment of the inability of available theories to
provide a solution, a series of failures worth reviewing as a source of salutary lessons
regarding the explanatory challenge posed by the problem of continuity.
Austin thought that the conditions of creation, persistence, and disintegration of
legal systems mirrored the conditions of creation, persistence, and disintegration of
independent political societies, of which Austin had states primarily in mind. Yet
such a view, Raz argues, cannot explain how it is that independent political societies
can come to recognize the jurisdiction or legal authority of outside institutions yet
still remain meaningfully independent, a possibility which might do well in
explaining the character of member states within the European Union.
Kelsen thought that the problem of continuity was to be solved by seeing that a
legal system’s identity and continuity rested with authorization by a historically first
constitution. The problem with Kelsen’s view, however, is that it does not provide
the means to explain how legal systems (such as Canada’s) can become independent
through legal authorization, such as constitutional devolution of power.22 Canada,
for example, reached its present stage of independence from the United Kingdom
by virtue of its Constitution Act, 1982, an act of the Canadian Parliament, which
established domestic constitutional amendment power and the Charter of Rights
and Freedoms. The Constitution Act, 1982, however, is also Schedule B to the
Canada Act, 1982, an act passed by the UK Parliament, which grants—or perhaps
recognizes—Canada’s independence. On Kelsen’s view, there are at least two
possible explanations. The Constitution Act, 1982 might be considered the histor-
ically first constitution of the Canadian legal system. Or it might be the case that
Canada does not yet have an independent legal system since its historically first
constitution still lies somewhere in the constitutional history of the United King-
dom. The trouble is that there seems to be no way to decide in Kelsen’s view
between these two very different possibilities.
19 Though here it might depend on who you ask. John Finnis argues that the problem is unsolvable
on the terms set by legal positivism, but a solution is available from natural law theory. See Finnis
(1973).
20 Raz (2009b: 58). 21 Raz (2009b: 58). 22 See Finnis (1973).
Hart’s views might seem to fare somewhat better than those of Kelsen, as Hart
developed one dimension of a solution to the problem of the continuity of law-
making rules by introduction of the idea of a rule of recognition, an analytical
device whose invention was motivated by observation of the defects of Austin’s
habit-based account of sovereignty.23 Surely the idea of a system-constituting rule
of recognition, rooted in admittedly dynamic official practices, at least points the
way toward explanation of continuity of legal systems? Raz provides reason to
doubt that explanation of the rule of recognition is in part also an explanation of the
continuity of legal systems. Hart’s explanation is disturbingly thin, Raz explains,
exactly where it ought to be thick. Raz asks:
What changes are consistent with the continued existence of the same rule, and what
changes compel the admission that a new rule has replaced the old one? It is easy to bring
examples for either situation, as well as examples of borderline cases. However, it is not the
existence of borderline cases, which are inevitable, that is disturbing. The disturbing fact is
that Hart’s theory provides no clue as to how to draw the conceptual distinction.24
In Raz’s view, the diagnosis of Hart’s oversight is simple:
If his theory fails to provide a complete solution to the problem of identity it is because he
overlooked not only part of the answer but also a whole question: that of the relation of law
and state.25
We will suggest later that Hart’s oversight may not actually be a deficiency of his
theory, but instead a decisive strength: his apparent omission of explanation of an
obviously close tie between legality and the state is in fact a strength enabling his view
to more closely approach full-blooded generality as a general jurisprudence. But let us
first follow Raz’s suggestion further, since his reasons for supposing legality and state
systems to be one and the same are relevant to our evaluation of whether analytical
legal theory and legal pluralism really are in a Janus-faced relation. If we are to solve the
problem of continuity, according to Raz, we must raise the question of the relation of
legality and state in a way which avoids the extremes of Austin’s and Kelsen’s views.
Whereas Austin attempted to reduce or define law entirely in terms of political facts of
sovereigns, habits of obedience, and coercion, Kelsen explained the state entirely in
legal terms, as a juristic entity, without reference to political facts of any kind. Raz
rejects both views, asserting instead that the solution lies somewhere between them:
A theory of law must be based, at least partly, on a theory of the state, and denying this has
been one of Kelsen’s gravest mistakes. A theory of state, however, is partly based on a theory
of law—the two are intimately connected.26
23 Recall that in Hart’s view, Austin’s account of sovereignty, which was based on the habitual
obedience of subjects to particular sovereigns, would signal the end of a legal system once a sovereign
died, since it would take time for habits of obedience to a new sovereign to form. See Hart (1994:
51–61).
24 Raz (2009a: 98).
25 Raz (2009a: 98).
26 Raz (2009a: 99). But of course Kelsen was never aiming at a social theory of law. If his theory is
interpreted in this way, it shouldn’t come as a surprise that it will look mistaken.
Raz has not, however, provided a theory of the state which would explain the
continuity of legal systems, even while observing, as we noted above, that
the problem remains unresolved—a situation which still stands, so far as we are
aware.27 Raz’s brief remarks in 1971 on a strategy for solving the problem are
nonetheless instructive: “A state is the political organization of a society, it is a
political system that is a subsystem of a more comprehensive social system.”28 Legal
systems, in particular, are themselves subsystems of the political systems of the
states of which they form a part. It follows, then, “that since the continuity of a
legal system is tied to the continuity of the political system, the former is affected by
the fate of the non-legal norms that happen to form part of the political system
concerned.”29
In section 3 we shall explore some of the assumptions in Raz’s suggestion. It will
be useful in advance of that discussion to attempt to explore further the meaning of
the suggestion, couched as it is in open-textured phrases asserting that there is an
“intimate relation” between law and state, in which the legal system as a “subsys-
tem” is “affected by the fate” of norms in the larger political system. Intimacy and
fate, in particular, demand further explanation, as they are little in evidence
elsewhere in Raz’s jurisprudence. These terms connote a kind of blending of
interests and purposes characteristic of interdependence: but interdependence of
what kind, with what points of connection and bearing on which less intimate
relations with other social and legal systems? It may be initially tempting to think of
the relation as quasi-isomorphic, a reflexive relation between two norm systems
which likely vary principally in terms of their claim to supreme authority: where a is
a norm of the state system with property g, there is a corresponding norm p of the
social system with property g; and where there is a norm p of the social system with
property g, there is a corresponding norm a of the legal system with property g. So
where there is a legal norm that swimming in the lake is permitted, there is a
corresponding social norm that swimming in the lake is permitted; and where there
is a social norm that swimming in the lake is permitted, there is a corresponding
legal norm that swimming in the lake is permitted. While perhaps prima facie
trivial, this reading of Raz’s view is actually revealing to the extent that it asserts a
relation of interdependence between social and legal systems which might serve as a
reminder in the event of finding a social norm without a corresponding legal norm
that one might need to look harder, or question whether the social norm really is a
social norm. Yet the relation between social and legal systems cannot be quite so
intimate, since as Raz observes in the passage immediately quoted above, the legal
system is a “subsystem of a more comprehensive social system,” meaning that the
more comprehensive social system is larger so there is no simple one-to-one relation
of social to legal norms. The intimacy involved is evidently rather more complex. It
may be helpful for at least heuristic, exploratory purposes to conceive of the inter-
systemic relation as a variety of supervenience: a non-symmetric supervenience in
27 But for a recent account on the relation between law and state, see MacCormick (1999: ch. 2).
28 Raz (2009a: 100). 29 Raz (2009a: 100).
which properties of the smaller system of authoritative, exclusionary legal norms

supervene on the system of non-authoritative, non-exclusionary norms of the
larger social system. So for a given state system with norms (a, b, c) exhibiting
respectively properties (g, h, i), those norms and their properties supervene on
corresponding norms of the broader social system (p, q, r) exhibiting respectively
properties (g, h, i). So a swimming in the lake b on Sundays c may be g legally
permitted only with h supervision and i prohibited after dusk, and this situation
corresponds to social norms (p, q, r) with properties (g, h, i) permitting supervised
Sunday swimming before dark. Yet the supervenience is non-symmetric, to the
extent that while the social system may exhibit both supervening associations of
legal norms (a, b, c) with corresponding social norms (p, q, r) with common
properties (g, h, i), the social system may exhibit further norms which lack
corresponding legal norms and properties. So the social system may exhibit, for
example, norms (l, m, n) and associated properties bearing no further relation to the
legal system. The supervenience relation between these non-symmetric systems
is nonetheless reflexive to the extent that changes to legal system norms and
properties supervene on social system norms and properties. If this is a plausible
rendering of Raz’s view of an “intimate relation” as in effect an assertion of non-
symmetric supervenience of legal systems on their associated social systems, what is
perhaps surprising at this juncture is what is left incomplete. An assertion of
supervenience is an assertion that for some set of property-bearing objects under
consideration, a certain pattern or variation is evident.30 An assertion of super-
venience is not at the same time an explanation of why the supervenience
occurs—such an explanation amounts to “superdupervenience.”31 Without push-
ing beyond plausibility the exploratory suggestion that Raz’s view is usefully viewed
as a kind of assertion of supervenience, it is nonetheless worth noting what
application of the distinction between supervenience and superdupervenience
helps us to see in the nature and reach of Raz’s view. Raz’s explanation following
assertion of an “intimate” relation between legal and social systems does not appear
to extend on toward superdupervenience, in explanation of the ontological priority
of elements of legal or supporting social systems. Nor do his views contain
assertions regarding dependence between elements of legal and social systems.
This is perhaps sensible to the extent that his view is left open to supervening
legal norms’ originating in either the legal system or the social system; but it
does leave a possible dependency relation unexplained, as the “intimate relation”
explored above does not appear to arrive with any further explanation of how the
legal system can be “affected by the fate” of non-legal norms of the social system.
In short, relations of dependence between social and legal systems seem to be
presumed rather than explained. We are offered a static snapshot of the intimate
relationship with little understanding of how it changes, and how, if intimacy
collapses, we are to detect that collapse.
30 McLaughlin and Bennett (2011: especially section 3.7).

31 Horgan (1993).
3. The relation between law and state
We have arrived at a puzzling position. Hart’s use of the rule of recognition to

account for continuity may be problematic to the extent that it does too little to
explain the relation of law and state, but Raz’s view sets us only a little ahead, and
raises other problems along the way. No analytical legal theorist, it seems, is willing
to deny that the continuity of legal systems depends on the continuity of surround-
ing and encompassing political and social conditions. That is, none want to deny
that the existence and persistence of law depends upon political and social facts. Any
theory of law which attempts to identify and explain the social nature of law must
explain how law sits upon social foundations. And surely if the social conditions of
existence of a political community fail, its legal system will fail as well. Yet as
preceding analysis showed, if these are plausible explanatory propositions, their
plausibility is not an output of Raz’s explanation of the relation between law and
state via explanation of the interaction between a state legal system and a state-wide
social and political system. As we saw, while Raz’s view states or maps the pattern of
shared properties of norms in a state legal system and a social system, he offers no
explanation of their interaction with one another, or their interaction with other
systems—explanations including, perhaps, how a given social system rejects as an
imposter the claim of a non-corresponding legal system to authorize a supervening
norm. What is instead evident is that what leads Raz to focus on state legal systems as
the object of continuity are his assumptions about the state. In particular, Raz relies
on the assumptions that the social and political systems which encompass legal
systems and constitute the state must be both (i) unified systems of norms and more
importantly (ii) bounded by state borders. In what follows we shall contend that
neither assumption is safe. States may internally embody multiple and possibly
conflicting political and social systems in various relations with one another, and
may tend towards fragmentation and instability rather than unity and stability. Even
states with meaningfully independent legal systems may depend to a significant
extent on the existence and stability of international political and social norms,
which themselves might not be systemic, unified, and stable.
Let us look within the state first. Is there a social system—a system of norms, in
Raz’s view—of a state which partly constitutes and is partly constituted by the
state’s legal system, standing in an “intimate relation” with the legal system and
affecting its fate? Here a number of perplexing questions immediately arise.
First, if there is a social system of norms, is it a system of norms in the same way
or sense that a legal system is a system of norms which, as we have noted, claims
through the attitudes and actions of norm-applying institutions to be comprehen-
sive, supreme, and open? And if it is a system of the same kind, what are its norm-
applying institutions determining the membership of the norms of the system?32
32 Notice that if it is a system of the same kind as a legal system, in every state we would have
competing claims of supremacy, as both the social system and the legal system would each be claiming
supremacy over each other.
These questions may on reflection seem wrongheaded: perhaps, since the relation
between law and state is intimate, we ought not to look for distinct norm-applying
institutions of the social system. These social institutions might just be the same
institutions as the norm-applying institutions of the legal system. So just as, for
example, courts apply the norms of the legal system, they are also social actors who
apply the norms of the social system. This suggestion has more than trivial
plausibility to the extent that actors within different institutions may be members
of both, as officials in courts applying legal norms are simultaneously members of a
broader system of social norms. Yet this suggestion carries an unacceptable explana-
tory cost, as it seems to leave us with no means of distinguishing the norms of the
social system from the norms of the legal system, since we can no longer distinguish
these by pointing to their distinct owners and operators. More importantly, this
isomorphic relation seems to be inconsistent with the implication of Raz’s claim
that a legal system is nested in a political system, in turn nested in a social system,
supporting a reading of Raz’s view as admitting some but likely not complete
overlap in the institutions of legal and broader political and social systems. While
Raz does not offer anything approaching a theory of social systems confirming this
reading, other related remarks are consistent with this reading, as, for example, he
suggests that Dicey’s constitutional conventions are good examples of non-legal
political norms whose continued existence and acceptance matters for the continu-
ity of the legal system.33 What is distinctive about constitutional conventions is that
they not enforced by courts,34 but are maintained by a state’s political institutions,
especially its executive, legislature(s), political parties, and other actors. Yet even
this observation does not deliver an explanation of the systemic character of the
social system whose operations affect the fate of the legal system—it remains to be
explained just what binds the social system as a system.
Second, we might also wonder whether the social system is really, or necessarily,
or entirely a system of norms. Raz’s discussion of constitutional conventions
certainly points in this direction. He also offers as an example of non-legal norms
that matter for the continuity of a legal system what he calls “norms of respect,”
which are norms held by society in general, or “some important sections in the
society,” “for at least some of the laws and some authorities.”35 Yet no explanation
of why respect must take the form of a norm or norms is given. It certainly seems
possible that attitudes of acceptance and recognition towards the legal norms or
legal system, which do not themselves exist as norms,36 legal or otherwise, could
explain how and when legal systems and their norms persist, such that if the
attitudes of acceptance and recognition come to an end, so does the legal system.
Notice, however, that such attitudes of acceptance and recognition, as a matter of
social fact about the existence and continuity of legal systems, are likely very diffuse,
33 Dicey (1952: ch. 14).

34 Courts will, however, often recognize the existence of conventions. See e.g. the Canadian
patriation reference case Reference re a Resolution to amend the Constitution [1981] 1 SCR 753.
35 Raz (2009a: 100).
36 Such attitudes might simply exist as a matter of habit, with no critical or normative features.
as legal systems typically need a very wide social base to exist. The social base likely
overlaps but more importantly probably extends beyond the base of legal officials
and norm-applying institutions of the legal system.37
Third, and most importantly, do these social norms or attitudes always point
towards and support the laws of the state legal system? In a stable and well-
functioning state, the answer is usually yes; but there are instances where even in
relatively stable states the social base of norms and attitudes may additionally or
conflictingly support non-state norms. Consider, for example, instances of market-
driven norms and industry standards, which depend upon their acceptance and
practice by market and industry institutions and businesses. Often these norms are
more stringent than state laws,38 and sometimes they are norms of a transnational
(and so again, non-state) nature.39 Consider as well states in which entire regions or
sections are effectively self-governing, despite the claims or pretensions of the
centralized government and legal system. Tamanaha has written extensively
about social life and law in the federated states of Micronesia,40 and Boaventura
de Sousa Santos has observed the relative isolation of favelas in Brazil.41 In
Micronesia indigenous communities operate mainly according to indigenous cus-
tomary laws, even when these conflict with state laws, and such a social situation is
widely recognized and accepted by both state officials and the indigenous commu-
nities themselves.42 Similarly, favelas are effectively governed by internal leaders,
not state officials (who rarely venture into such areas), and this relatively stable state
of affairs is recognized and accepted by both state officials and the relevant actors in
the favelas. In each instance, there are wider social norms which support both state
as well as non-state (and sometimes outlaw) norms. What Tamanaha’s and Santos’s
examples have in common is not just identification of the possibility, but in fact
evidence of the reality of failure of a state legal system’s claims to supremacy and
comprehensiveness.
Now let us turn to the state’s external legal-normative environment. It may be
easily seen just why the nature of state legal systems might not be adequately
explained in isolation from international law, or at least in isolation from an
account of international conditions. All principles of sovereignty in international
law define or regard sovereign statehood partially in relational terms. For instance,
consider the elements of statehood laid out in the Montevideo Convention of
1933, formalizing in a North and South American treaty various elements of
sovereignty gradually emerging in practice and state vs. state international litiga-
tion, such as the locus classicus Lotus case of the Permanent Court of International
37 For a sample of recent work on questions of popular acceptance and efficacy, see Alexander and
Schauer (2009); Adler (2006); and Postema (2008).

38 See e.g. Cashore (2002) and Cashore, Auld, and Newsom (2004).
39 The lex mercatoria is the best-known example.
40 See Tamanaha (2001: 145–6).
41 See Santos (2002: 99–162).
42 So we might say that such self-governance and conflicts are preserved by wider social norms, i.e.
norms which encompass both state and non-state norms.

Justice.43 Article 1 of the Convention provides that to be a state there must be a

territory, a permanent population, and an effective government capable of
governing the territory and population and entering into international relations.44
The benefits of sovereignty so conceived include the rights to territorial integrity
and freedom from intervention, which arrive with recognition by other states
(although such freedom is increasingly becoming subject to institutionally recog-
nized moral conditions, especially human rights conditions45). Under this concep-
tion of sovereignty, the social practices of surrounding states, especially their
recognition and cooperation, matter very much to the existence and operation of
state legal systems. There can be little doubt, for example, that the sovereignty of
Mexico, and so the existence of its legal system (while perhaps not the content of its
particular legal norms, though even here there is likely influence), depends a great
deal on the United States’ government’s recognition of Mexico’s government,
borders, and citizens, among other things. Any social fact account of a state legal
system must admit as much: if the practice of recognition of state sovereignty were
to end, state legal systems would come to an end as well.46
Taken together, these questions amount to a demand for explanation of the various
interactions which might occur amongst the elements of the independent and
internally unified yet corresponding legal and broader social systems mapped and
distinguished by Raz. In the absence of such an explanation, we are left with an
account of a momentary legal system, and no means by which to approach the
problem of continuity. Perhaps surprisingly, this result is consistent with Raz’s own
account of at least early iterations of his view: in The Concept of a Legal System he
restricts his account to investigation of momentary legal systems: “Apart from a few
remarks in this section, this chapter is concerned only with the problem of identity of
momentary systems.”47 More importantly, he explicitly disavows as a task of analyt-
ical jurisprudence the investigation of social phenomena needed to explain the
interaction between legal and social systems where the former is “affected by the
fate” of the latter, writing somewhat enigmatically of a dependency relation between
legal system and “social forms:”: “The identity of legal systems depends on the identity
of the social forms to which they belong.”48 Exploration of the relation between legal
system and broader social system or “social form” is outwith the responsibilities or
capacities of analytical jurisprudence, according to Raz, pressing on to claim that:
The criterion of identity of legal systems is therefore determined not only by jurisprudential
or legal considerations but by other considerations as well, considerations belonging to other
social sciences.
43 Available at <http://www.taiwandocuments.org/montevideo01.htm>. Accessed 2 May 2011. See
also SS Lotus (Fr. v. Turk.), 1927 PCIJ (ser. A) No. 10 (September 7). Available at <http://www.
worldcourts.com/pcij/eng/decisions/1927.09.07_lotus.htm>. Accessed 21 May 2012.
44 The Montevideo Convention is discussed in Buchanan (2004: 264). On the essentially relational
quality of sovereignty, see also Endicott (2010).

45 See e.g. Reisman (1990). For illuminating discussion see Raz (2010a) and Raz (2010b).
46 In fact, where the recognition of the sovereignty of particular states has been withheld, the legal as
well as social order of those states often collapses as well.

47 Raz (1980: 187). 48 Raz (1980: 189).
Not wishing to trespass on other fields, I shall confine myself to the problem of the
identity of momentary systems.49
While this deference to the methods and insights of other disciplines may seem
sensible on general grounds of methodological modesty, Raz’s omission of even an
attempt to borrow constructively from “other fields” leaves his view short of
precisely the evidence or “other considerations” needed to transform his account
of the legal system-social system relation from map to explanation. In light of this
apparent acceptance that resolution of the problem of continuity lies in the hands
of others outside analytical jurisprudence, it is quite striking that Raz claims in
1998 that the problem of continuity remains for analytical jurisprudence, and
remains unsolved because interest has “shifted elsewhere.”50 It is difficult to avoid
concluding that Raz aims to have it both ways: to claim simultaneously that the
explanation of the legal system-social system relation key to resolution of the
problem of continuity must be left to “other” social science, while maintaining
nonetheless that the problem of continuity is a jurisprudential problem. If this is
the case, it is no surprise that analytical legal theory has yet to develop or receive a
solution to the problem of continuity: resolution requires a kind of interdisciplinary
diplomacy which is acknowledged as necessary yet shunned as somehow requiring a
presumably wrongful or at least undesirable “trespass.”
What can we draw from the preceding observations? One conclusion stands out:
by seeing how the project of state-based legal theory needs to complete itself—by
investigating the relation between law and state as mediated and represented by the
legal system-social system relation, it becomes easier to see where the intersection,
rather than opposition, between state-based legal theory and legal pluralism lies. If
the social norms and conditions which are necessary to support law can be seen, in
some contexts, to be disparate and not a unified system, not restricted within the
borders of single states, and to support non-state legality as well as state legality,
the tight association between law and state begins to loosen. And this is precisely
the destination at which the legal pluralists arrive, albeit by a different route, one
which never shared the core commitments of analytical, state-based legal theory.
A further set of insights regarding analytical legal theory can be drawn from
observation of how the limitations of Raz’s views point toward convergence with
legal pluralism regarding the subject or subjects of legal theory. We presented in
section 2 above Raz’s contention that Hart had no solution to the problem of the
continuity of legal systems, because Hart did not ask the question to which the
solution would be the answer: the question of the relation between law and state.
Our subsequent analysis and extension of Raz’s view now enables us to reconsider
the plausibility of this criticism of Hart. In The Concept of Law and elsewhere, Hart
attempted to take seriously the social conditions of law’s existence, in ways
especially evident in his development of a system-constituting rule of recognition
whose existence and scope is to be determined by observing the social practices of
recognition of the various officials of a particular system. What is missing from
49 Raz (1980: 189). 50 Raz (2009b: 58).

Hart’s view is an account of how these social practices amount to a wider social
and political system of norms, a single sovereign state, in effect, which stands in
a necessary and one-to-one relation with the state legal system.51 Contra Raz,
however, this is plausibly regarded not as a weakness but as a strength of Hart’s
theory. Hart’s view escapes the questions and problems we have noted in Raz’s
view, to the extent that Hart’s conception of legality is not inherently state based
and thus compelled to view legality in terms of a state in correspondence with a
geographically bounded and unified social and political system of norms. Amongst
the views of analytical legal theorists, Hart’s position seems most likely to be able
to make sense of non-state forms of legality, such as international legality and
transnational legality.52
4. Remaining disagreement
Discussion in the last section might give rise to the impression that the difficulties
faced by Raz’s state-based view should incline analytical jurisprudes toward a
Hartian approach which in turn leads to a convergence with legal pluralism and
collapse of the possibility of a Janus-faced relation between pluralism and analytical
jurisprudence. This is not, however, the view we aim to convey or support. While
all should welcome joint acknowledgment between analytical and pluralist theorists
regarding the existence and importance of non-state legal phenomena, at least one
key issue signals that disagreement is likely to remain: the question of just how
inclusive the concept of legality ought to be. For legal pluralists, there is little reason
why non-state social norms ought not to be considered law. Often they are spoken
of, thought of, and function as law in creating and maintaining social order. For
state-centered legal theorists, however, this extends the concept of law far too
broadly and empties it of all meaning. The goal of analytical jurisprudence—or
at least one of its goals—is, after all, to explain how legality and laws are related to
but also different from normativity generally and a wider class of social norms.53 It
is beyond the scope of this chapter to resolve this disagreement here; yet it is safe to
say that it is not a disagreement that can be resolved by simply adducing more facts
or by appealing to any shared or ordinary understanding of law or shared concept of
law. Rather, resolution of this disagreement will likely require a return to a much
more extensive investigation of the theoretical aims of descriptive-explanatory
theories of law. We can, however, briefly return to Hart’s account of the difference
between wide and narrow concepts of law for some clues as to the ingredients of a
resolution of the question of the proper degree of inclusiveness of an explanation
of the concept of law. As Hart notes in explaining the “choice” between a wide
51 “Every state—by which is meant a form of political system and not a juristic person—has one
legal system that constitutes the law of that state, and every municipal legal system is the law of one
state.” Raz (2009a: 98–9).
52 Though not without some modification. See Culver and Giudice (2010).
53 See Waluchow (2010).
positivist concept of law, which includes within the category of law morally evil but
legally valid rules, and a narrow natural law concept of law, which excludes morally
evil rules from the category of law, he appeals to two success criteria for theories or
concepts of law:
If we are to make a reasoned choice between these concepts, it must be because one is
superior to the other in the way in which it will assist our theoretical inquiries, or advance
and clarify our moral deliberations, or both.54
We think it would be difficult to deny the gains for theoretical and moral inquiries
by including transnational legality within the concept of law. Note here that, in
relying on a reasoned choice, what matters most is what gets included within the
purview of legal theory, and not what theorists are willing to consider law. We tend
to side with those who would include a greater diversity of social norms within the
concept of law, but we also hope to have shown that the very project of state-based
legal theory, with its need for an account of the relation between law and state,
already includes, and necessarily, greater investigation of the diversity of social
norms within the subject matter of legal theory.
Notice, however, that to accept a greater diversity of types or sources of legality is
to accept only a part of the legal pluralist challenge raised by Tamanaha and
Twining, namely, that there is a diversity of types and sources of legality. Accepting
this does not amount to accepting pluralism about concepts of law. Consider, for
example, Tamanaha’s claim that the sheer diversity of varieties of legality shows
why law can have no essence: that is, that state law, religious law, aboriginal law,
transnational law, and other varieties of law are simply too different to warrant any
presumption that they share an essence. But how do we know these are all varieties
of legality to begin with? Tamanaha’s answer is this: they are all law because as a
matter of convention and practice we (or some group) have attached the label “law”
to them.55 The labeling test, however, plainly will not do. As Twining observes, it
leads to the following strange conclusion: since we speak of the “laws of cricket”
but only the “rules of association football,” we must conclude that cricket is
regulated by law but football is not.56 To this insight regarding the insufficiency
of an approach too-reliant on lexical use, we can add a further argument against
inference from the plurality of types or sources of legality to a plurality of concepts
of law. As Tamanaha and Twining admit, analytical jurisprudence has been
squarely focused on explanation of state legality, which in part means that few, if
any, have even attempted to find or construct a concept of law which meaningfully
illuminates diverse types or sources of legality. With so few or perhaps no substan-
tive attempts on offer, it is likely premature to pass negative judgment on the
prospects of a truly general or universal jurisprudence which provides a truly general
or universal concept of law, in favor of a usage-tracing approach which leads to
54 Hart (1994: 209).

55 Tamanaha (2001: 151, 166).
56 Twining (2009: 101, n. 69). See also Twining (2003) and Himma (2004). Yet even Twining is
explicitly reluctant to being characterized as holding an essentialist view about law.
highly counter-intuitive, indeed absurd conclusions, depriving us along the way of

meaningful characterization of the ways in which legal order is quite plainly
distinguished from social order and social practice.
5. Old questions made new

The title of this chapter promises renewal of old jurisprudential questions, a
renewal whose motivation, it should now be apparent, lies in the need to respond
to the legal pluralist challenge and its surfacing of social phenomena which cannot
be ignored by any jurisprudential approach claiming to take social facts seriously.
The force of the legal pluralist challenge is particularly evident in the context of the
continuity problem, once viewed by analytical jurisprudes as the problem of
explaining the existence and persistence of distinct legal systems, and in particular,
distinct state legal systems as they are challenged by such phenomena as revolution
and devolution.57 This conception of the problem of continuity is committed to
regarding the state legal system as a central unit of analysis. There are several
familiar features of the idea of state legal system which characterize this commit-
ment, surveyed above and expressible in digest form: the state legal system contains
all the legal norms of the system, claiming supreme and comprehensive authority
and organized hierarchically as superior to all other social norms in the broader
social system of which the legal system is a member, with legal norms recognizable
by their pedigree in particular hierarchically organized institutional sources or
official practices.58 Such a system typically makes its claims of comprehensive,
supreme authority with respect to a given geographic territory, and views compet-
ing normative claims as either members of entirely separate systems or as in effect
subordinate to and within the operational control of the system.
These features are certainly found in—at least the idea of—the modern sover-
eign state, in which the creation, application, and enforcement of law are central-
ized, and there is no appeal or reliance on extra-state institutions or rules for
validation or authorization. Conceiving the problem of continuity of law in the
particular terms of the continuity of state legal systems is, however, just one way of
conceiving the problem of the continuity of law. Given the historical prominence
of sovereign states, this particular choice has evident merit; but in light of our
accumulating argument, there are reasons to suppose that we might need to
reconceive the terms of the problem. Completion of the project of state-based
legal theory, which includes offering a solution to the problem of the continuity of
legal systems, requires explanation of the supporting and associated social and
political norms and attitudes. These, however, may or may not be systemic, and
57 See e.g. Finnis (1973).

58 There is of course a problem of circularity looming here, as the norms of a legal system appear to
determine who the officials and institutions are but the officials and institutions are responsible for
determining which norms are legal norms of the system. For an assessment of the problem of
circularity, and its associated problem of indeterminacy, see Culver and Giudice (2010: ch. 1).
may or may not be restricted by state boundaries.59 More importantly, such social
and political conditions may or may not converge in supporting the laws of the state
legal system. They may sometimes support non-state legal orders, which might be
contained within single states or converge to create trans-state or supra-state forms
of legality. The problem of the continuity of law might then be more aptly
understood not as a problem about how to draw clear borders around the edges
of state legal systems, but instead how to trace the various ways in which state and
non-state legality come into existence, and overlap and interact with each other.
Why might this be a better way of conceiving the problem of continuity? Theories
of law ought, amongst other achievements, to explain the social reality of law, and
the social reality of law might be untidy in ways not amenable to explanation by
state-based legal theory. Perhaps more importantly, we must recover and revisit
analytical legal theory’s initial motivation for attention to the problem of the
continuity of state legal systems, a motivation rooted in the perfectly sensible idea
that we ought to be able to track the borders of legality. This motivation has been
lost in analysis, stuck in problem subdivision as efforts to grapple with the problem
began with “momentary” legal systems and failed to advance to the fully-fledged
problem of continuity. Yet any theory of law purporting to be a social fact theory
must recognize that while laws may at least appear to be unchanging and so outside
of time, the social life of law is ever-dynamic and within time. There can be no fully
satisfying theory of law or account of the concept of law which does not face and
account for the dynamic relation within legality and between legality and surround-
ing social phenomena. The problem of continuity is accordingly not secondary, to
be resolved once legality is given a foundational characterization; rather, it is co-
fundamental, and can never be ignored. Analytical legal theorists owe thanks to
legal pluralists for spurring renewed attention to this fact, and owe themselves a
redirection of jurisprudential effort.
References
Adler, M. D. (2006). “Popular Constitutionalism and the Rule of Recognition: Whose
Practices Ground US Law?,” Northwestern University Law Review, 100: 719–805.
Alexander, L. and Schauer, F. (2009). “Rules of Recognition, Constitutional Controversies,
and the Dizzying Dependence of Law on Acceptance,” in M. D. Adler and K. Himma
(eds.), The Rule of Recognition and the US Constitution. New York: Oxford University
Press.
Austin, J. (2000). The Province of Jurisprudence Determined. New York: Prometheus.
Buchanan, A. (2004). Justice, Legitimacy, and Self-Determination. Oxford: Oxford
University Press.
Cashore, B. (2002). “Legitimacy and the Privatization of Environmental Governance: How
Non-State Market-Driven (NSMD) Governance Systems Gain Rule-Making Authority,”
Governance: An International Journal of Policy and Administration, 15: 503–29.
59 See e.g. Cotterrell (2008).

Cashore, B., Auld, G., and Newsom, D. (2004). Governing through Markets: Forest
Certification and the Emergence of Non-State Authority. New Haven: Yale University Press.
Cotterrell, R. (2008). “Transnational Communities and the Concept of Law,” Ratio Juris,
21: 1–18.
Culver, K. and Giudice, M. (2010). Legality’s Borders. New York: Oxford University Press.
Dicey, A. V. (1952). Introduction to the Study of the Law of the Constitution. Ninth edition.
London: Macmillan and Company.
Endicott, T. (2010). “The Logic of Freedom and Power,” in S. Besson and J. Tasioulas
(eds.), The Philosophy of International Law. Oxford: Oxford University Press.
Finnis, J. (1973). “Revolutions and Continuity of Law,” in A. W. B. Simpson (ed.), Oxford
Essays in Jurisprudence. Second series. Oxford: Clarendon Press.
Hart, H. L. A. (1994). The Concept of Law. Second edition. Oxford: Clarendon Press.
Himma, K. (2004). “Do Philosophy and Sociology Mix? A Non-Essentialist Socio-Legal
Positivist Analysis of the Concept of Law,” Oxford Journal of Legal Studies, 24: 717–38.
Horgan, T. (1993). “From Supervenience to Superdupervenience: Meeting the Demands of
a Material World,” Mind, 102: 555–86.
Kelsen, H. (1966). Principles of International Law. Second edition. Toronto: Holt, Rinehart
& Winston of Canada.
Leiter, B. (2003). “Beyond the Hart/Dworkin Debate: The Methodology Problem in
MacCormick, N. (1999). Questioning Sovereignty. Oxford: Oxford University Press.
McLaughlin, B. and Bennett, K. (2011). “Supervenience,” in E. N. Zalta (ed.), The Stanford
Encyclopedia of Philosophy (Winter 2011 Edition), <http://plato.stanford.edu/archives/
win2011/entries/supervenience/>.
Postema, G. (2008). “Conformity, Custom, and Congruence: Rethinking the Efficacy of
Law,” in M. H. Kramer, C. Grant, B. Colburn, and A. Hatzistavrou (eds.), The Legacy of
H. L. A. Hart: Legal, Political, and Moral Philosophy. Oxford: Oxford University Press.
Priel, D. (2007). “Jurisprudence and Necessity,” Canadian Journal of Law and
Jurisprudence, 20: 173–200.
Raz, J. (1980). The Concept of a Legal System. Second edition. Oxford: Clarendon Press.
Raz, J. (2009a). The Authority of Law. Second edition. Oxford: Oxford University Press.
Raz, J. (2009b). “Two Views of the Nature of the Theory of Law: A Partial Comparison,” in
Between Authority and Interpretation. Oxford: Oxford University Press.
Raz, J. (2010a). “Human Rights without Foundations,” in S. Besson and J. Tasioulas (eds.),
The Philosophy of International Law. Oxford: Oxford University Press.
Raz, J. (2010b). “Human Rights in the Emerging World Order,” Transnational Legal
Theory, 1: 31–47.
Reisman, M. (1990). “Sovereignty and Human Rights in Contemporary International
Law,” American Journal of International Law, 84: 866–76.
Roberts, S. (2005). “After Government? On Representing Law without the State,” Modern
Law Review, 68: 1–24.
Santos, B. (2002). Towards a New Common Legal Sense. London: Butterworths.
Tamanaha, B. (2001). A General Jurisprudence of Law and Society. Oxford: Oxford
University Press.
Tamanaha, B. (2008). “Understanding Legal Pluralism: Past to Present, Local to Global,”
Sydney Law Review, 30: 375–411.
Twining, W. (2000). Globalization and Legal Theory. London: Butterworths.
Twining, W. (2003). “A Post-Westphalian Conception of Law,” Law and Society Review,
37: 199–257.
Twining, W. (2009). General Jurisprudence. Cambridge: Cambridge University Press.
Waluchow, W. (2010). “Legality’s Frontier,” Transnational Legal Theory, 1: 575–85.
13
Legal Disagreements and the
Dual Nature of Law*
Andrea Dolcetti and Giovanni Battista Ratti
Disagreement does not necessarily have to be overcome. It may remain an

important and constitutive feature of our relations to others, and also be seen
as something that is merely to be expected in the light of the best explanations
we have of how such disagreement arises.
(B. Williams, Ethics and the Limits of Philosophy1)
1. Introduction
As a matter of fact, people disagree about the laws in force in their communities.
Sometimes, when people try to solve their disagreements, they may witness other
disagreements amongst lawyers, judges, and legislators about the requirements of
their legal system. If they are legal philosophers, they may disagree not only about
the law of their community, but also about the nature of law in general. In this
chapter we consider what legal disagreements can reveal about the nature of law.
Before sketching the structure of this work, two caveats are in order.
The first caveat concerns the expression “nature of law.” In our opinion, the use
of this expression should be avoided, if it is meant to entail the ontological claim
that law has an essence, conceived of as a stable (set of) necessary feature(s) common
to all legal systems.2 In this chapter, we use the expression “nature of law” simply to
* A first draft of this chapter was presented at The Nature of Law: Contemporary Perspectives
Conference, hosted by the Philosophy Department at McMaster University, Hamilton (ON), Canada,
11–15 May 2011. The authors would like to thank Professors P. Chiassoni, R. Guastini, B. Leiter,
S. Sciaraffa, F. Shecaira, K. Toh, and W. Waluchow for their comments and Ms Daniela Simone for
her editing suggestions. The authors are solely responsible for any errors.
1 Williams (2006: 133).
2 H. L. A. Hart considers questions such as “what is the nature (or the essence) of law?” to be a more
obscure formulation of the question “what is law?”, which can be understood as a request for a
definition of law. Hart (1994: 6). See also Raz (2009). On necessity, cf. Van Fraassen (1977).
302 Andrea Dolcetti and Giovanni Battista Ratti
refer to some features of law which have greater explanatory power in comparison to
other features of law.3
The second caveat concerns the term “law.” This term notoriously has a
plurality of meanings. There are well-known differences between: (a) a law, that
is a normative provision belonging to a certain legal system; (b) the law, that is
an institutionalized set of norms existing in a given space-time dimension (i.e. a
particular legal system); and (c) law, that is a type of normative institution which
one can recognize as occurring in a number of societies.4 We hold the view that it is
important for a legal theorist to keep the questions “what is law?” and “what is the
law?” separate in order to avoid generalizations about law on the basis of aspects
peculiar to a particular legal system.5 Nonetheless, the identification of remarkable
properties of a given legal system might provide a good counterexample to a
(supposedly) universal claim about the nature of law.
The hypothesis that this chapter intends to test is that the presence of specific
types of disagreements in most contemporary legal systems—those that we call
interpretative disagreements 6—is a characteristic feature of law. Once this propos-
ition is verified, we demonstrate that both disagreements on the interpretation of
legal sources, and an agreement on the identification of those sources, are character-
istic features of law. Furthermore, we argue that disagreements about the interpret-
ation of the sources of law are not simply compatible with, but indeed presuppose,
an agreement on the identification of those sources.
In order to verify our initial hypothesis we discuss one of the arguments that
Ronald Dworkin has developed to challenge H. L. A. Hart’s legal positivism
(section 2). We begin by noticing that Dworkin’s “argument from disagreement”
is problematic, since it contains two fundamental ambiguities concerning the
expressions “grounds of law” and “theoretical disagreements” (section 3). In light
of a refined taxonomy of legal disagreements (provided in section 4) we consider
how the ambiguities in Dworkin’s reconstruction of legal disagreements may affect
his critique of Hart’s positivism (section 5). Finally, we consider the relevance of the
notion of legal disagreements for general jurisprudence. On the basis of the
linguistic character of legal norms, we maintain that most contemporary legal
systems are characterized, on the one hand, by an interpretative (meaning-content)
disagreement, and on the other hand, by an institutional (source-based) agreement.
We conclude by showing that these characteristic features of modern law are related
in such a way that it makes sense to talk of a “dual nature” of law (section 6).
3 See Ross (1961), and Schauer (2013). See also Endicott (2013).
4 Finnis (1987: 368): “Positivist and natural law theories in jurisprudence are not, and do not even
look like, theories about the law of any particular community (in the sense of offering to identify
propositions of law which are true for that legal system), or about the criteria for identifying the law
which are used by the lawyers and judges of any particular community. They look like theories about
what law—a(ny) legal system—‘necessarily is’ (at least in its paradigmatic instantiations, its central
cases)”.
5 For a discussion, see Twining (2009).
6 These types of legal disagreements are discussed in section 4.2.
Legal Disagreements and the Dual Nature of Law 303
2. Dworkin’s argument from disagreement
In his critique of contemporary general jurisprudence and, in particular, of Hart’s

legal positivism, Dworkin focuses attention on the importance of the phenomenon
of disagreements for legal practice. According to Dworkin, a sound theory of law
must be able to provide an explanation of disagreements concerning issues of law.7
Dworkin maintains that legal disagreements—especially disagreements between
judges—cannot be properly explained by a theory, like Hartian positivism, that
defends a conception of law based on agreement.8 Hart’s theory of law as the union
of primary and secondary rules can indeed be reconstructed as relying upon an
agreement between legal officials on the identification of the valid legal rules which
constitute their legal system (i.e. a shared consensus, at least amongst officials,
about the rules of recognition of that system).9
The identification of a supposed lack of explanatory power when dealing with
legal disagreements is the second stage of a broader critique which Dworkin has
addressed to Hart’s legal positivism over the course of two decades. The first stage,
as is well known, aimed to attack legal positivism by emphasizing the fact that
judges often feel legally bound to use moral principles in deciding legal cases.10 The
second stage, which emerges from the pages of Law’s Empire, relies upon an
argument developed in light of two dichotomies: (a) between grounds of law
and propositions of law; and (b) between empirical and theoretical disagreements
about law.11
(a) According to Dworkin, the object of disagreements about the law is the truth
or falsity of a given proposition of law. Propositions of law are defined as: “all the
various statements and claims people make about what the law allows or prohibits
or entitles them to have.”12 Dworkin maintains that lawyers, judges, and ordinary
people generally assume that some propositions of law, at least, can be true or false.
Propositions of law can be true or false in virtue of “other, more familiar kinds of
propositions” which “furnish” what Dworkin calls the grounds of law.13
7 According to Dworkin, lawsuits raise three different kinds of issues: issues of fact, issues of law,
and issues of political morality. The first concerns the actual historical events in the controversy,
whereas the last concerns moral rights and wrongs, i.e. moral disagreements emerging in courts and
revealing what people think the law should be. Issues of law, instead, reveal what people think the law
is. Dworkin maintains that lawyers and judges seem to disagree very often about the law governing a
case. On this point, see: Dworkin (1986: 3). This book is meant to be, in the words of its author, a
book “about theoretical disagreement in law”. See Dworkin (1986: 11).
8 Dworkin (1986: 34–5).
9 For a discussion of the thesis that rules of recognition are to be understood as conventional rules,
see Dickson (2007).

10 Judges make use of moral principles, at least, in deciding hard cases. See Dworkin (1977).
11 As Scott Shapiro puts it: “whereas the first critique seeks to exploit the alleged fact that judges often
take the grounds of law to be moral in nature, the second critique tries to capitalize on the alleged fact that
judges often disagree with one another about what the grounds of law are.” See Shapiro (2007: 41).
12 Dworkin (1986: 4).
13 For this reason, one might say that propositions of law are parasitic on propositions that “furnish”
the grounds of law. In the example offered by Dworkin, the proposition of law “No one may drive over
(b) Dworkin distinguishes two ways in which it is possible to disagree about the
truth of a proposition of law. First, people might agree about the grounds of law,
but disagree about whether those grounds are in fact satisfied in a particular case
(empirical disagreements about law). Second, people might disagree about the
grounds of law, i.e. about which other kinds of propositions, when true, make a
particular proposition of law true (theoretical disagreements about the law). In the
former case—the case of empirical disagreements about law—people agree about
when the truth (or falsity) of other (more familiar) propositions would make a
particular proposition of law true (or false).14 In the latter case—the case of
theoretical disagreements about the law—people would agree about what the
statute books and past judicial decisions have to say about a case. Yet, they might
still disagree about what the law actually is, since they might disagree about whether
statute books and past judicial decisions exhaust the pertinent grounds of law.
It seems to us that both these dichotomies are problematic. The expression
“grounds of law” is ambiguous since it may denote, in one sense, the (possible)
sources of law—such as constitutions, statutes, judicial decisions, international
treaties, or customs—whereas, in another sense, it may denote the meaning of
these sources. A study of most contemporary legal systems demonstrates that
although it is possible to disagree about what the sources of law are, people
might also disagree just on their meaning. Therefore, we suggest distinguishing
two different concepts of “grounds of law,” to which two different concepts of
“theoretical disagreements” correspond.
It is worth noting that this distinction, overlooked in Dworkin’s argument, has
not been fully considered in recent analyses of the so-called “Hart-Dworkin
debate.”15 Brian Leiter, for instance, maintains that “the key theoretical disagree-
ments for Dworkin concern the meaning of the acknowledged sources of law,” but
in the same article he also affirms that “a theoretical disagreement is a disagreement
about the criteria of legal validity, that is, about the content of what Hart calls the
Rule of Recognition”16—that is about what sources count as valid law. Scott
Shapiro holds the view that theoretical disagreements occur when legal participants
all agree about the sources of law but “dispute their legal significance,” although on
the previous page he maintains that this kind of disagreement “involves conflicting
claims about what the grounds of law are.”17 This apparent confusion may stem
from the fact that both kinds of disagreements, in some sense, can be said to be
55 miles an hour in California” might be true if, for instance, it is true that “a majority of California’s
legislators has raised their hands when a text of that effect lay on their desks”.
14 For example, two judges might agree that (it is true that) “the speed limit is 55 miles an hour in
California” if (it is true that) “the official California statute book contains a law to that effect”.
Nonetheless, the two judges might still disagree about whether the book does in fact contain such a law.
see Dworkin (1986: 5): “People can disagree about what words are in the statute books in the same way
they disagree about any other matter of fact”.
15 Leiter (2003). For a reconstruction of the arguments proposed in response to Dworkin’s critique
of Hart’s positivism, see Ratti (2009).

16 Leiter (2009: 1217 and 1222).
17 Shapiro (2007: 36–7).
theoretical: one may hold a certain theory of the sources of law, just as one may
hold a certain theory of interpretation. Nonetheless, in order to understand
Dworkin’s argument against Hart’s theory of law it is necessary to keep disagree-
ments about the identification of legal sources and disagreements about the
interpretation of legal sources clearly separate.
3. Identifying and interpreting legal sources
The ongoing debate that has stemmed from the anti-positivist argument that
Dworkin employs in Law’s Empire is characterized by a certain lack of clarity
regarding the notion of “grounds of law” and, consequently, the notion of “theor-
etical disagreement.” This is not the only reason why we are not in favor of the
use of this expression. In our opinion, Dworkin’s use of the same adjective—
“theoretical”—to describe both cases is rather unfortunate, for at least three
reasons. First, it obscures the fact that different types of disagreements might be
at play. Second, it suggests the (misleading) idea that such disagreements are com-
pletely theorized amongst the participants: on the contrary, as we will see later on, this is
not necessarily the case. Lastly, and more importantly, the term “theory” evokes the idea
of a scientific explanation of a certain phenomenon; whereas in fact, the phenomenon
of legal disagreements results from a divergence in evaluative doctrines (i.e. normative
theories) about sources and their meanings. Consequently, we suggest using the
terms “doctrine” and “doctrinal,” where Dworkin uses “theory” and “theoretical.”
As we suggested in the previous section, the participants in the discussion of the
Hart–Dworkin debate seem to conflate two different meanings of “grounds of law,”
which may denote, in a first sense, what is usually called “sources of law,” and, in a
second sense, may refer to the meaning of those sources. We suggest, therefore, keeping
the disagreements lawyers might have about what the sources of law are separate from
disagreements lawyers might have about the interpretation of the sources of law.
Both types of disagreements—about the identification of legal sources and about
the interpretation of those sources—are conceptually separate and their possible
relations can be described in terms of the following matrix:
(i) agreement on possible sources, but disagreement on their interpretation;
(ii) disagreement on possible sources, but agreement on their interpretation;
(iii) agreement both on possible sources and their interpretation;
(iv) disagreement both on possible sources and their interpretation.18
Although these four scenarios are logically possible, the reality of contemporary
legal systems leads us to think that one would probably be reluctant to affirm that a
18 The second case mirrors, from a logical point of view, the analysis of conditionals which is
provided in Quine (1960: } 46). Agreements on interpretation take place on the basis of the hypothesis
that a certain source is valid. This can also be discussed by reference to a counterfactual scenario, as in
the case of the interpretation of sources which are surely invalid (such as, for example, the application of
the rules of Roman law within the contemporary Italian legal system).
legal system exists in the case of persistent disagreement about the possible legal
sources of that system and their interpretation.19 It seems plausible to conceive of a
legal system whose existence—at least for a short period of time—is consistent with
the contingent overlapping of different authoritative sources recognized as legal
sources. Yet, most contemporary legal systems seem to be characterized by an
agreement on the (legal) sources of the system coupled with persistent disagreement
on their interpretation.20
It must also be observed that disagreements on the legal relevance of an
authoritative source often appear to depend on disagreements about the correct
interpretation of legal provisions dealing with legal sources—in Hartian terms we
might say the interpretation of rules of change.21 This observation suggests a way of
explaining disagreements about the sources of law.22 Yet, in this chapter we intend
to focus on the relationship between disagreements about the sources of law and
disagreements about their meaning (i.e. the result of the process of interpretation).
Before doing so, a clearer classification of legal disagreements is needed.
4. A taxonomy of legal disagreements
As a result of the previous analysis of the Dworkinian notions of “grounds of law”

and “theoretical disagreements,” and for the purposes of this chapter, we would like
to offer a refined taxonomy of legal disagreements, by distinguishing the following
four types of disagreements:23
(i) Source-recognition disagreements;
(ii) Interpretative disagreements;
(iii) Decision-making disagreements;
(iv) Axiological disagreements.
4.1. Source-recognition disagreements

“Source-recognition disagreements” may be defined as those situations where
diverging conceptions of the valid legal sources in a certain legal system are held
by different participants.24 Such disagreements are at the heart of jurisprudential
19 The opposite scenario is a legal system characterized by stable agreement on both sources and
their interpretation—a situation that few societies have ever known, and certainly not for long.
20 For discussion of this point, see section 6.
21 Hart (1994: 95–6). See also Bobbio (1994).
22 This explanation implies a certain theory of legal sources: i.e. what counts as a legal source
depends, mainly, on what is provided by authoritative texts, whose interpretation may well be
controversial. For further discussion, see Jori (2010).
23 This taxonomy is obviously not intended to cover all the possible disagreements that might take
place within the legal domain.

24 In what follows, we will focus on these situations from the perspective of officials (mainly judges)
as Hart and Dworkin do. However, it must be observed that divergent views about the valid sources
may also materialize amongst the participants at large in a legal system (i.e. the addressees of rules of
conduct).
thinking: the rationally indeterminable debate between legal positivist and natural
law theories can be reconstructed in a way that provides a clear instance of this type
of disagreement. At a lower level of abstraction, lawyers sometimes disagree about
whether particular kinds of free-floating standards (such as the principle that “no
man may profit from his own wrong” used by the majority in Riggs v. Palmer) are
indeed part of the law, whereas disagreement about pedigreed sources, enacted by a
competent normative authority, are less frequent (at least in non-pathological legal
systems).25
4.2. Interpretative disagreements

“Interpretative disagreements” may be defined as those situations where different
and diverging doctrinal conceptions about the meaning of legal sources in a certain
legal system are held by different participants.26 “Interpretative disagreements”—
we submit—are at the heart of a lawyer’s craft: lawyers are generally expected, at
least in the legal systems we are acquainted with, to be good creators of viable legal
arguments which might oppose the arguments proposed by other parties in the
process. This characteristic is clearly triggered by two aspects of most contemporary
legal systems:
(1) the presence of a multiplicity of principles and values upon which the system
is considered to be founded, which may constitute the basis for different and
incompatible ascriptions of meaning to legal sources;27
(2) the presence of pedigreed rules or customary practices of legal interpretation
which, by admitting a non-ordered set of canons of interpretation, bring
about multiple and conflicting results (i.e. meanings).28
4.3. Decision-making disagreements

If the descriptive thesis that contemporary legal systems admit different canons of
interpretation of legal sources, which are not ordered by the legal authorities or by
jurists’ practice, is true, then another possible type of legal disagreements can be
generated by those interpretative canons: decision-making disagreements. Such
disagreements usually arise when different judges solve the same legal problem in
a different way: this phenomenon is particularly visible in the case of not unani-
mous decisions, especially when majority and dissenting opinions are explicitly
expressed.29 Consequently, we define “decision-making disagreements” as those
25 The locus classicus is Hart (1994: 117–23).

26 There is a sub-distinction (which we will elaborate on in section 6) regarding disagreements
about the doctrine of interpretation chosen, and the particular ascriptions of meaning to a certain legal
source which can be triggered by the same doctrine of interpretation or by different doctrines of
interpretation.
27 See Schauer (1991: 73ff.). 28 Kelsen (2000). 29 Kelsen (2000: xiv).
situations where two or more legal actors select diverging solutions for an individual
case amongst different (equally justifiable and possibly contrasting) legal norms.30
4.4. Axiological disagreements

Different value judgments might be found to underlie source-recognition disagree-
ments, as well as interpretative and decision-making disagreements. “Axiological
disagreements” may be defined as those situations where diverging conceptions
regarding (the existence or the content of) values are held by different partici-
pants.31 It is important to stress that axiological disagreements have effects, as it
were, on the degree to which a solution might be found “locally.” Jurists—amongst
them, typically judges—identify normative solutions to legal issues by means of a
“local-centered method.”32 This is to say that they try to solve the issue which they
are faced with by resorting to those resources that offer a direct solution. When
such rules offer a solution which is regarded as grossly unjust, jurists generally tend
to resort to deeper normative or axiological levels of the legal system, to retrieve the
resources which might allow for a rereading of the rule in a fairer sense, or for a
defeasance of the rule in favor of another legal standard, considered to be more
adequate to the case at hand.33
5. Legal disagreements and Hart’s legal positivism

The purpose of this section is to consider whether Dworkin’s argument from
disagreement—as a critique of Hart’s legal positivism—is right on target. Brian
Leiter has maintained that Dworkin’s critique of legal positivism has been substan-
tially met. Although Leiter still believes that everybody interested in jurisprudence
should know of this critique, he argues that such a critique should, nevertheless, be
resigned to the past and is, therefore, beyond debate. Scott Shapiro holds the view
that Leiter is substantially correct when commenting on the “first act” of the Hart–
Dworkin debate—which focuses on the theses Dworkin propounds in “The Model
of Rules I”—but he is wrong for omitting consideration of the second act of the
saga, lurking in “The Model of Rules II” and taking full shape in Law’s Empire.34
Both Leiter and Shapiro analyze and discuss Dworkin’s argument from disagree-
ment, but they do not clearly state the distinction that we have drawn between
source-recognition disagreements and interpretative disagreements. Our suggestion
30 Guastini (2006: 70).

31 More precisely, people disagree about values when they express a judgment on either how a
(future) action should be guided or how a (past) action should be qualified. We maintain the view that
disagreements concerning these two kinds of value judgment affect the sparking of legal disagreements,
but they also affect the solution of such disagreements. For a discussion of this separation, with special
reference to Rawls and Habermas, see Besson (2005: 24–39).
32 See Alchourrón and Bulygin (1971: Ch. 1), and Ratti (2008a: Part II).
33 See Schauer (1998) and (2012).
34 For a critical review of this thesis, see Kramer (2008).
is that a reading of Dworkin’s critique in light of this distinction would make it

possible to defend Hart’s theory of law, without dismissing a salient feature of
contemporary legal systems that Dworkin has correctly identified (i.e. the phenom-
enon of interpretative disagreements).
One of Leiter’s central contentions is that Dworkin emphasizes a phenomenon
that is completely marginal to law: disagreements materialize only in a very small
percentage of cases.35 If most of the time people agree on the solution of a case,
what is the fruitfulness of explaining a certain object by pointing to some features
that only arise on unimportant occasions? For this reason, Leiter affirms that
positivism is a sound theory, since it explains agreement throughout the system.
Nevertheless, contemporary legal systems seem, indeed, to be pervaded by conflict-
ing principles that stimulate interpretative disagreements.36 Dworkin does not
deny that many solutions are by and large agreed on at a local level. More
importantly, however, Dworkin affirms that any solution (independent of its
degree of confirmation) is subject to the threat of “justificatory ascent;” that is to
say, any decision may, in future, be reviewed in the light of another interpretation,
triggered by a different contextualization of a normative provision. Any interpret-
ative decision is only pro tempore and in any case depends upon evaluative consider-
ations. No doubt, inconsistency between principles and subsequent disagreements
seem to be one of the key features of most contemporary legal systems.37
The contention that a legal system considered at a particular moment in time
could be regarded as being founded upon a general agreement on its different
components is not problem free, as the agreement reached, in fact, might often be
only an “apparent” one.38 However, if we look at the legal system in a dynamic
fashion, it seems quite misleading to affirm that it contains only a few disagree-
ments on rather marginal points. The same normative provisions are usually
interpreted over and over, without ever reaching a point of final confirmation:
one can easily appreciate that jurists will constantly disagree on their meaning given
their possible reinterpretation over time.
What are the main causes of such diachronic disagreements? As Guastini has
recently pointed out, linguistic ambiguity is not the main cause of interpretative
disagreements.39 Disagreements due to linguistic ambiguity make us face (statistically
35 Leiter (2009: 1227).

36 Contemporary legal systems are usually based on constitutions that incorporate supposed natural
rights. The first author to emphatically stress the constant conflict between such rights was Jeremy
Bentham: see Hart (1983b: 149–50).
37 Dworkin (2006: 52–3).
38 As Giovanni Tarello correctly put it, agreement in modern legal systems is often about normative
texts, not meanings. See Tarello (1980: 367). In some sense, there is an overall constitutional and
legislative agreement to agree to leave things for the judges to settle. In other words, there is a
theoretical agreement which conceals interpretative disagreements. A major fact of our contemporary
legal systems seems to be that our supreme and constitutional courts roughly agree on what counts as a
valid legal source but they might be split into opposing sides, which hold different and opposite views
about virtually everything else. It does not matter how many cases reach these superior courts: the
relevant fact is that the foundations of our legal systems, save for an abstract criterion to determine what
counts as a legal text, seem to be completely up for grabs.
39 Guastini (2006: 78–80).
rare) questions which have the following logical structure: “Does normative text
T express the norm N1 or the norm N2?” However, the majority of interpretative
disagreements stem from other (much more frequent) questions about the mean-
ing of a normative text: e.g. “Does T express only N1, or N2 as well?,” “Does T,
which surely expresses N1, also imply N2?,” “No doubt, T expresses N1. But, is
N1 defeasible?” According to Guastini, the meaning of a normative text can be
broken down into two parts: an uncontroversial kernel and a controversial halo.
This is not, however, equivalent to Hart’s core/penumbra dichotomy, for Hart was
thinking of controversies about subsumptive sentences (“Is this skateboard a
vehicle?”), whereas Guastini is pointing to controversies about genuine interpret-
ative sentences (“Does section 1 of the Sherman Act invalidate all the acts which
restrain trade or just some of them?”). In other words, Hart was preoccupied by the
doubtful qualification of individual cases under generic concept-words, while
Guastini deals with the ascription of meaning to sentences. Hart’s problem is a
real world-oriented one (“Does this concept x apply to instance y?,” whereas
Guastini’s is a text-oriented matter (“What norm does the legal provision
S express?”). This phenomenon of controversy about the halo of meaning of a
text seems to be central in modern law, since different sets of values, which can
supposedly be deemed to underlie the process of rule formulation, may be used to
devise questions about the scope of the meaning of a text (in this sense, they may
be said to be capable of extending the halo of controversy).40
It is now time to consider how Hart’s theory might accommodate the claim that
officials and citizens often disagree about the law. As is well-known, Hart maintains
that the model with most explanatory power is the idea of a legal system as the unity
of primary (duty-imposing) and secondary (power-conferring) rules. Secondary
rules are remedies for three defects—uncertainty, staticity, and inefficiency—
which characterize a simple form of social structure consisting only of primary
rules of obligation. In relation to the first defect, Hart notices that “ . . . if doubts
arise as to what the rules are or as to the precise scope of some given rule, there will
be no procedure for settling this doubt, either by reference to an authoritative text
or to an official whose declarations on this point are authoritative.”41 A “rule of
recognition” is the kind of secondary rule that needs to be introduced in the legal
system to provide a remedy for the system’s uncertainty. It is worth noting that, in
its simplest form, this rule might be just a list of rules (e.g. the written reproduction
of unwritten rules).42 The point stressed by Hart is that the criteria must acknow-
ledge some source as authoritative, in order to provide “conclusive identification of
the primary rules of obligation.”43 Yet, in modern legal systems the rules of
recognition—since there are various sources of law—are more complex, and they
usually refer to some general characteristic possessed by the primary rules of the
system, “having been enacted by a specific body, or their long customary practice,
40 See Ratti (2008b). 41 Hart (1994: 92).

42 Hart (1994: 94). According to Hart the existence of a rule of recognition is manifest in the
general practice of private persons and officials.
43 Hart (1994: 95).
or their relation to judicial decisions.” When there is more than one identifying
criterion, such criteria are usually ranked to prevent conflict.44 Possible conflicts
between different criteria of identification might mirror the type of disagreement
that Dworkin has called “theoretical disagreements” and that we have identified as
“source-recognition disagreements.”45
According to Hart, there are two minimum conditions for the existence of a legal
system:
(a) valid rules of behavior must be generally obeyed (valid according to the
ultimate criteria of validity in that system);
(b) its rules of recognition (setting the criteria of legal validity), rules of change
and rules of adjudication must be effectively accepted by officials.46
One might wonder whether it is possible for the officials not to accept the rule of
recognition and the other secondary rules of their legal system. Of course this might
be the case, but Hart considers this to be a symptom of a pathological legal
system.47 Interestingly, Hart also considers how rule-skepticism might affect the
rule(s) of recognition of a legal system, since the open-texture of all terms belonging
to any natural language might well affect the rules of recognition of the system.48
Here, it is important to distinguish between uncertainty in relation to an ordinary
validity question and uncertainty in relation to the ultimate criteria of legal
validity.49 In any case, by discussing the vagueness of legal language Hart does
not reject the idea that disagreements—in this case, what we have called decision-
making disagreements—can persist within a legal system.
If one wants to restate Dworkin’s critique in light of the taxonomy we have
suggested in section 4, one should say that Hart’s theory of law cannot provide a
sound account of source-recognition disagreements—i.e. disagreements about
the rule(s) of recognition—and/or interpretative disagreements. In our opinion,
Dworkin’s considerations are only relevant for a general theory of law in relation to
the second type of disagreement we considered in our taxonomy. Yet, we maintain
that Hart’s theory does not rule out the possibility of explaining this important legal
44 Hart (1994: 95).

45 On the relation between rules of recognition and sources of law, see Hart (1994: 101), and
related note on page 294. See also Hart (1994: 214), where Hart talks about a rule of recognition as one
“specifying ‘sources’ of law and providing general criteria for the identification of [ . . . ] rules.”
46 Hart (1994: 116).
47 Hart (1994: 123): “a substandard, abnormal case containing within in the threat that the legal
system would dissolve.”

48 Hart (1994: 147–54).
49 The open-texture of language “may affect the rule of recognition specifying the ultimate criteria
used in the identification of the law as much as a particular statute.” See Hart (1994: 123). On the
uncertainty of the ultimate criteria of validity, see Hart (1994: 148): “The distinction between
the uncertainty of a particular rule, and the uncertainty of the criterion used in identifying it as a
rule of the system, is not itself, in all cases, a clear one. But it is clearest where the rules are statutory
enactments with an authoritative text. The words of a statute and what it requires in a particular case
may be perfectly plain; yet there may be doubts as to whether the legislature has power to legislate in
this way. Sometimes the resolution of these doubts requires only the interpretation of another rule of
law which conferred the legislative power, and the validity of this may not be in doubt.”
phenomenon.50 It is true that Hart does not develop a full theory of interpretation,
although he addresses the issue of decision-making disagreements when he
discusses the open-texture nature of legal language. Furthermore, Hart’s theory of
law (as the union of primary and secondary rules) provides a convincing account
of source-recognition (dis)agreements, by explaining how, in order to exist, a legal
system requires a certain degree of agreement (at least amongst officials) on legal
sources.51
6. Agreement, disagreement, and the “dual nature” of law

This section is devoted to an analysis of the relevance of legal disagreements for
general jurisprudence. Our discussion of Dworkin’s and Hart’s accounts of legal
disagreements has led us to affirm that both the agreement amongst legal officials
on the sources of law and their disagreement on the interpretation of those sources
are characteristic features of a (contemporary) legal system. We have also argued
that agreement on legal sources is logically prior to any disagreement on their
interpretation and is a necessary—although not sufficient—condition for the
existence of a legal system.52 Next, we provide some further qualifications to the
basic notions of “agreement on sources” and “interpretative disagreement,” which
are central to our argument.
50 See, paradigmatically, Hart (1983a: 106): “The clear cases are those in which there is general
agreement that they fall within the scope of a rule, and it is tempting to ascribe such agreements simply
to the fact that there are necessarily such agreements in the use of the shared conventions of language.
But this would be an oversimplification because it does not allow for the special conventions of the legal
use of words, which may diverge from their common use, or for the way in which the meaning of words
may be clearly controlled by reference to purpose of a statutory enactment which itself may be either
explicitly stated or generally agreed. A full exploration of these questions is the subject-matter of the
study of the interpretation of statute.” We read this passage as stating that the convergence on a
particular decision may be the result of several interpretative and meta-interpretative factors. Hart’s
statement seems also to admit that there are problems of interpretation and meta-interpretation that he
does not deal with, which may complicate the neat picture of the dichotomy between clear and hard
cases. Amongst the several factors which may complicate the picture, there is, for instance, defeasibility,
which intervenes in apparently easy cases. For discussion of this factor, see Alchourrón (1996).
51 Two clarifications are in order here. First, any given legal system might have—and usually has—
more than one rule of recognition: these rules identify multiple legal sources of the system. Second,
most “issues of law” (using Dworkin’s terminology) are decided by reference to intermediate “grounds
of law” in the chain of validity of the system. Dworkin’s argument, therefore, needs to challenge Hart’s
theory at the level of the rules of recognition—that for Hart are ultimate and supreme.
52 Dworkin himself seems to hold a similar view. See Dworkin (1986: 90–1): “Law cannot flourish
as an interpretive enterprise in any community unless there is enough initial agreement about what
practices are legal practices so that lawyers argue about the best interpretation of roughly the same data.
That is a practical requirement of any interpretive enterprise: it would be pointless for two critics to
argue over the best interpretation of a poem if one has in mind the text of ‘Sailing to Bizantium’ and the
other the text of ‘Mathilda Who Told Lies’. I do not mean that all lawyers everywhere and always must
agree on exactly which practices should count as practices of law, but only that the lawyers of any
culture where the interpretive attitude succeeds must largely agree at any one time. We all enter the
history of an interpretive practice at a particular point; the necessary preinterpretive agreement is in that
way contingent and local.”
6.1. Agreement on legal sources

We ought to refine our rough conception of agreement (and conversely, disagree-
ment), by introducing some further distinctions, related to the kind of propos-
itional attitude which is required in order to explain the existence and the
functioning of legal systems. This, in turn, will allow us to determine the proper
place for disagreement (and conversely, agreement) in a successful jurisprudential
explanation of law. In other words, in this subsection we undertake to distinguish
between the conceptual issue of the identification of the objects to be interpreted
(for our present purposes: legal sources) and the social-empirical phenomena which
connect to, or accompany, such an identification.
We maintained that some kind of agreement on legal sources is necessary in
order for a legal system to exist. Strictly understood, this view might be called into
question if seen through the lens of the conceptual/empirical dichotomy we have
just mentioned. It seems possible that the identification of valid legal sources occurs
due to a mere overlapping of different conceptions of the validity of legal sources.
If a certain official has a certain conception Æ of legal sources, another official
holds conception , and another one holds conception ª, and the three conceptions
intersect in the case of the subset of sources ø, we can say that ø is the set of
extensionally identified legal sources which correspond to the overlapping
of extensions spelled out in the formulation of the three different conceptions of
legal sources.53 This, of course, does not appear to require any particular agreement:
on the contrary, it may appear to be the outcome of mere chance. So, the minimal
requirement to have a legal system seems to be the common identification, by
officials, of a certain set of extensionally delimited legal sources. The empirical
phenomenon whereby different officials identify the same set of present legal sources
we call “convergence.”54
Mere convergence, understood as the not necessarily intended intersection of
some conceptions of legal sources held by different officials, is at one extreme of the
continuum of empirical factors which can lead to the identification of the same
“objects” (as it were, the sentential basis) to be interpreted. At the other extreme—
which is more unlikely than mere convergence—it is a successful plan made by
officials in order to construct a legal system, by founding it on a common concep-
53 According to the standard version of set theory, the formal representation of such a situation is:
ø = Æ \ \ ª = {x | (x ∈ Æ) & (x ∈ ) & (x ∈ ª)}. On this issue, see Ratti (2012: chs. 1 and 4).
54 Note that the identification of certain “objects” to be interpreted may concern non-present or
non-actual objects: for instance, legal sources which are no doubt invalid (e.g. the legal sources of the
imaginary state of “Erehwon” or of the Roman Empire at the time of Trajan). We can sensibly have the
same criteria of identification of these sources, but since they are not current sources, they are not (or
no longer) liable, by definition, to be subject to empirical convergence (but only, as it were, to
“hypothetical convergence”). Observe that there is a difference between the “historical case” and the
“imaginary case,” since only the former is liable to empirical confirmation regarding the occurred
validity of the identified sources. However, when this judgment is uncertain because of our ignorance
of the past, both cases are, indeed, quite similar. The identification of “valid” legal sources is only
hypothetical: we reason as if they were valid, even though we know that they are not, or we do not
know whether they are. Observe, moreover, that both sets of legal sources are currently void.
tion of legal sources. Real legal systems seem to be somewhere in the middle of this
continuum:55 officials, by and large, share common views on what valid legal
sources are, but there are controversial cases which materialize from time to time.
In this sense, although agreement on legal sources does not seem to be a necessary
feature of every legal system, it does seem to be a pervasive feature of existing legal
systems. If we consider all possible “legal worlds,” the cases of a mere convergence
or a completely successful plan seem to be very distant from the legal world that our
reality instantiates.
In any case, we must stress the fact that the identification of the “object” to be
interpreted is conceptually prior to any interpretative disagreement, and this usually
takes the social form of agreement.56 Accordingly, in the case of a legal system, the
identification of an interpretable object (e.g. a statutory provision) is necessarily
prior to its interpretation (i.e. a legal norm).57
6.2. Disagreement on the interpretation of legal sources

Once the object to be interpreted is identified, interpretative agreement or disagree-
ment enters the picture. Here, the thesis that there is a pervasive agreement on the
interpretation of legal sources—regarded as a general thesis about the nature of law,
i.e. as a thesis regarding all (contemporary) legal systems—is clearly false. We are
acquainted with many legal systems which are pervaded by a great many disagree-
ments on the interpretation of legal sources. Moreover, such disagreements seem to
be triggered by the very institutional design of the canons of interpretation. Since
many canons are admitted, and no hierarchy is usually imposed on those canons,
solutions that might be derived from legal sources are manifold and none of them
can be regarded as the “true” one.58 The “interpretative disagreements game” is a
feature which appears to be physiological in those legal systems which do not rank
legal interpretative canons.59 As a consequence, no general thesis about massive
interpretative agreement, or about the preferential character of a certain interpret-
ative technique, can be held to be a thesis about the “nature” of law. At best, it may
be regarded as a simple parochial thesis about the “nature” of the law (i.e. of a
particular municipal legal system). Now, if a parochial thesis about massive inter-
pretative agreement in a certain municipal system is true, then one may be tempted
to assert that the general thesis about massive interpretative disagreements within
legal systems is false. But such a hasty impression might be mistaken. The
identification of a certain set of constant solutions may be due to social factors
55 This point is correctly stressed in Hart (1994: 117–18).

56 Although, as we have just mentioned, it may also take the form of mere convergence or of a
successful plan.
57 For a more detailed presentation of this topic, the synchronic and diachronic dimensions of the
legal system must be distinguished. For the sake of simplicity, however, we will not elaborate on this
point here.
58 Guastini (2004: 39).
59 On the notion of “interpretative game” in the legal domain, see Chiassoni (2000: 86–92).
which have nothing to do with interpretation stricto sensu. Several social factors—
such as the way in which lawyers stand to gain from the legal process, the costs of
justice, and so on—may trigger or hamper legal claims made to courts of justice.60
But the presence of such social factors cannot constitute the basis for a theory of
legal interpretation, much less a basis for the theorizing of interpretative agreement.
What these social factors trigger in some systems is decision-making convergence:
that is, the constant identification of certain solutions to solve certain legal cases.
But such convergence should not be equated with interpretative agreement. More
precisely, we may envisage four situations which might occur regarding the inter-
pretation of legal sources and the empirical phenomena of massive convergence
which can accompany them.
(a) The identification of the same solution derived from the same doctrine of
interpretation;
(b) The identification of the same solution derived from different doctrines of
interpretation;
(c) The identification of different solutions deriving from the same doctrine of
interpretation;
(d ) The identification of different solutions deriving from different doctrines of
interpretation.
In light of this classification, to affirm that a legal system is pervaded by agreement
on interpretation is conceptually ambiguous, since situations (a), (b), and (c) may
be regarded as cases of agreement, although the agreement is on partially different
objects. However, in the literature, the main indication of interpretative agreement
can be identified in decision-making agreement: but this is also ambiguous, for
cases (a) and (b) may be subsumed under this broad heading.
One of the main questions to be addressed is what is it that renders different
doctrines of interpretation different. Doctrines may differ by virtue of the canons
they admit, the ordering they impose on those canons, or both. As an empirical
matter, it must be noted that judges frequently agree on the canons to be admitted
within their doctrines, even though they disagree on their ordering with the same
frequency.61 Having said that, we must iterate the twofold tenet that, in contem-
porary legal systems:
(1) Canons of interpretation are usually not ranked by the lawgiver, so that
different solutions may be admitted which derive from the acceptance of the
same non-ordered doctrine of legal interpretation (i.e. from what we can call
“subscription of the lawgiver’s doctrine by the interpreters”);
60 Luque Sánchez (2011: 247).

61 Sometimes, judges disagree on the interpretation of legal provisions which identify the canons
imposed by the lawgiver: this is the main case of disagreement about the canons to be admitted.
(2) Different fundamental values, which legal systems are based on, are also not
abstractly ranked, so that different solutions may be admitted which derive
from the acceptance of the same doctrine of fundamental values as the basis
for legal interpretation.
From a purely methodologically stance, if our systems are correctly characterized
by a set of legitimately usable canons, not ordered by the lawgiver, relevant
agreement may only be at the level of the ordering (or lack of ordering) of these
canons: every canon is available, and many interpretative solutions are then
allowed at the lawgiver’s discourse level. Only at the meta-level of the interpreter’s
doctrine, is this equality of canons altered: and this alteration is the precise object of
disagreement (of course, disagreement may also revolve around the very viability
of making such an alteration). If one solution is frequently (or always) picked up by
officials for a set of individual cases, we can speak of a decision-making agreement.
Note, however, that this by no means implies an agreement at the abstract level of
the doctrine of legal interpretation: for instance, judge A may consider that the
systematic canon dominates, whereas judge B may see the teleological canon as
dominating; and it may well be the case that in many circumstances the results
of the two canons converge. Note, moreover, that the ranking of canons must
logically respond to extra-legal considerations, unless the law itself imposes a rank
(which is rarely the case). These considerations are mainly triggered by axiological
attitudes and require a full explanation of the ideological or idiosyncratic reasons
which motivate the officials. We also need to emphasize that the choice of a non-
ordered set of interpretative canons is strategically connected to the possibility of
picking up a canon or a certain aggregate of canons according to the demands of
axiological preferences in an individual case. If this is correct, the justification of
the selection of a certain decision in a particular case is mainly the extra-legal order
that judges impose on a certain set of canons. So, judges must be taken as
disagreeing on the ordering of canons. This is, so to speak, a meta-interpretative
disagreement, for it refers to the means by which we can alter the structure of rules
on interpretation. It must be observed, however, that such disagreements are
nothing but a sub-species of axiological disagreements, since they are situations in
which different value judgments on the preference criteria amongst canons are held.
As a matter of fact, what generally happens in contemporary constitutional
legal systems—we submit—is that jurists agree on the solution to be adopted in
less demanding cases (i.e. those cases that are further removed from issues of
the ordering of fundamental values and require less extreme orderings of canons),
in many cases for reasons of mere conformity, even though they hold different
doctrines of (meta-)interpretation (e.g. originalism vs. evolutionism). Meta-
interpretative disagreement is hidden here beneath decision-making convergence.
In more demanding cases, however, different meta-interpretative doctrines gener-
ally surface bringing about different interpretative solutions.
A resulting qualification must be made at the theoretical level. Jurists—especially
judges—talk as if there is an interpretative solution which, legally, is the best. But
this idea should generally be rejected as legal dogma.62 From a legal theoretical
point of view, if we recognize that there is no general lawgiver’s ordering of
competing doctrines of interpretation and fundamental values in most contempor-
ary legal systems, we must also recognize that there is no legally best interpretative
solution. As a consequence, when judges discuss different “best” interpretative
doctrines, these discussions must be taken, inter alia, as expressing their own
meta-interpretative disagreements about the “best” ordering of such canons and
values.
For the reasons we have just stated, the sentential basis of contemporary legal
systems is always liable, for methodological reasons, to competing interpretations.
This manifestly triggers massive interpretative disagreement at the highest levels of
the judiciary. The fact that such disagreement is not very visible at lower levels does
not necessarily mean that it does not exist.
6.3. Legal disagreements and general jurisprudence

What conclusions are to be drawn from these observations at the level of general
jurisprudence? General jurisprudence investigates common aspects (if any) present
(by and large) in legal systems. Contemporary constitutional legal systems are
largely characterized by the “interpretative game” of allowing several canons of
interpretation, which are not univocally ordered. In particular, some canons are
allowed, such as the strictly literal and the intentional canon, the a contrario
argument and the analogical argument, which necessarily bring about opposing
normative solutions.63 Moreover, interpretations based upon a rule’s background
(e.g. legislative history) or motivating principles are liable to produce necessarily
opposing results depending on the exact layer of axiological resources at which the
lawyer decides to stop his or her interpretative activity.
If this is an empirically accurate description of contemporary legal systems, one
should infer that what general jurisprudence needs in order to account for the
phenomenon of disagreements, is the combination of a theory of the (extensional)
identification of the set of valid sources of law,64 and a theory of the legal
interpretation of those sources (both supplemented by an explanation of the
meta-interpretative ideological preferences held by the officials). But once we
agree that common identification is a factual occurrence which takes place in our
legal systems, what is needed is a general theory of interpretation (and meta-
interpretation) that is able to make sense of the interaction between sources and
interpretation, which is created by the interpreters’ doctrines.
In this regard, it is fundamental to distinguish between legal norms as prescriptive
statements contained in the sources of law and legal norms as the meanings of those
62 Kelsen (2000: xiv): “The view, however, that the verbal expression of a legal norm has only one,
‘true’ meaning which can be discovered by correct interpretation is a fiction, adopted to maintain the
illusion of legal security, to make the law-seeking public believe that there is only one possible answer to
the question of law in a concrete case. Unfortunately, the contrary is true.”
63 Kelsen (1967: 352).
64 For a classical general overview on sets identification, see Russell (1920: 12–13).
statements. The first are the objects of legal interpretation (interpretanda), whereas
the second are the products of legal interpretation (interpretata). We have argued
that when one considers a legal system as a set of prescriptive statements it is
characterized by an institutional agreement on its sources; whereas, when con-
sidered as a set of meanings ascribed to those statements usually there is no
agreement on that ascription of meaning. The convergence which is found at
some levels of certain legal systems may be the product of the overlapping of
solutions derived by different interpretative or meta-interpretative doctrines.
A possible objection to this line of argument might be that we overemphasize
what lawgivers’ rules provide, and therefore we might be overlooking what jurists
really do. By constructing rules of interpretation as “open,” disjunctive, obligations,
we might be seen as referring to the norm-content of legal systems, but not
acknowledging that jurists frequently converge on a certain solution. In a nutshell:
jurists could choose amongst many different solutions, but—as a matter of fact—
they frequently (or even always) select the same solution,65 so that the possibility of
decision-making disagreement is irrelevant to the theoretical explanation of legal,
and especially judicial, practice.
In our view, this line of argument is misleading in that from convergence on a
certain set of particular solutions it infers that such solutions are univocally justified
by means of the same interpretative and meta-interpretative standards. But that is a
proposition which must be empirically proved, as our analysis suggests. Moreover,
overlooking the fact that, legally, different courses of interpretation are available
(mostly, in a diachronic perspective) obscures the fact that the common solution
is merely a pro tempore “conventionally right” solution. In any case, this line of
argument has two theoretical drawbacks.
The first arises from the assumption that those decisions which apply different
canons which trigger a different solution are mistakes. But such canons, with the
solutions they trigger, are available to the “dissenting jurist” or to the “future jurist:”
and usually jurists do not take issue with the choice of the canons made by other
jurists, which shows that they do not necessarily regard other possible arguments
(or at least meaning-ascriptions) as “mistakes.” So, some specification is in order in
relation to which decisions are correct and which are mistaken.66
Second, and more importantly, the empirical fact of convergence may be
accounted for in several ways, depending on the theoretical reconstruction of the
sociological, interpretative and meta-interpretative factors we have at our disposal.
To derive interpretative agreement from decision-making agreement is, in this
sense, not completely justified, since competing (and possibly) better explanations
are available.67
65 This is suggested by Leiter in his pyramid metaphor. See Leiter (2009: 1226ff.).
66 We hold the view that Kelsen’s theory of the frame of meanings may help in dealing with this
issue. See Kelsen (1967: 350–1).
67 Here we have an instance of the famous “Duhem–Quine” thesis about the underdetermination
of theory by evidence. On this thesis, see Leiter (2007: 145–6). Oddly enough, such a thesis, which is
relevant in Leiter’s proposal to naturalize jurisprudence, in this case runs counter to the features
identified by Leiter as characterizing legal systems.
7. Conclusion
In this chapter we have argued that Dworkin’s critique of Hart’s legal positivism, as
well as the recent debate originating from this critique, fails to clearly distinguish
different types of disagreements which can arise in the legal domain. We have also
maintained that Hart’s theory of law does not lack the tools to provide an account
of legal disagreements. Although his theory focused on the phenomenon of insti-
tutional agreement on sources (via the concept of a rule of recognition), Hart also
dealt with the explanation of what we have called decision-making disagreements.
Furthermore, his theory does not seem to preclude the possibility of pervasive
interpretative disagreements in law. We have stressed how the thesis that consensus
lies at the foundations of law (i.e. agreement about the sources of law) does not
imply that law is characterized by an overall agreement: in this respect, we think
that Dworkin’s reconstruction of Hart’s theory of law is flawed.
In section 6, we proposed a reconstruction of legal disagreements according to
which, by disentangling the different relations which may exists amongst interpret-
ative and axiological choices, one may regard contemporary legal systems as
characterized by an actual institutional agreement on legal sources coupled with a
potential interpretative disagreement (backed by a pervasive meta-interpretative
disagreement). This has allowed us to show how the fact that there is a convergence
of certain judicial solutions in certain legal systems is compatible with interpretative
and (meta-interpretative) disagreements. It is true that at an inferior court level such
disagreements are not generally liable to surface; however, the shared identification
of legal solutions at this level is not necessarily due to interpretative (and meta-
interpretative) agreement, because different doctrines of interpretation and (meta-
interpretation) may (and often actually do) trigger the same legal outcome. On the
contrary, at higher justificatory levels—in cases decided at a superior court level—
the absence, within the system, of an institutionalized hierarchy of canons of
interpretation and of fundamental values, makes interpretative disagreements per-
vasive. In light of the relationship between institutional agreement and interpret-
ative disagreement, one might indeed speak of a “dual nature” of law. This
relationship refers to the physiological existence of a legal system—where the
firm basis of legal sources is open to differing interpretations—whereas a persistent
state of disagreement about both legal sources and their meanings would pertain to
the pathology of a legal system, as Hart clearly stated more than fifty years ago.
References
Alchourrón, C. E. (1996). “On Law and Logic,” Ratio Juris, 9: 331–48.

Alchourrón, C. E. and Bulygin, E. (1971). Normative Systems. Vienna and New York:
Springer.
Besson, S. (2005). The Morality of Conflict: Reasonable Disagreement and the Law. Oxford:
Hart Publishing.
Bobbio, N. (1994). “Norme Secondarie,” in N. Bobbio, Contributi ad un Dizionario
Giuridico. Turin: Giappichelli.
Chiassoni, P. (2000). “Interpretive Games: Statutory Construction through Gricean Eyes,”
in P. Comanducci and R. Guastini (eds.), Analisi e Diritto 1999: Ricerche di
giurisprudenza analitica. Turin: Giappichelli.
Dickson, J. (2007). “Is the Rule of Recognition really a Conventional Rule?,” Oxford Journal
of Legal Studies, 27: 373–402.
Dworkin, R. (1977). “The Model of Rules I,” in R. Dworkin, Taking Rights Seriously.
London: Duckworth.
Dworkin, R. (1986). Law’s Empire. London: Fontana Press.
Dworkin, R. (2006). Justice in Robes. Cambridge, MA: The Belknap Press of Harvard
University Press.
Endicott, T. (2013). “The Generality of Law,” in A. Dolcetti, L. Duarte d’Almeida, and
J. Edwards (eds.), Reading HLA Hart’s The Concept of Law. Oxford: Hart Publishing.
Finnis, J. (1987). “On Reason and Authority in Law’s Empire,” Law and Philosophy,
6: 357–80.
Guastini, R. (2004). L’Interpretazione dei Documenti Normativi. Milan: Giuffrè.
Guastini, R. (2006). Lezioni di Teoria del Diritto e dello Stato. Turin: Giappichelli.
Hart, H. L. A. (1983a). “Problems of the Philosophy of Law,” in H. L. A. Hart, Essays in
Jurisprudence and Philosophy. Oxford: Oxford University Press.
Hart, H. L. A. (1983b). “1776–1976: Law in the Perspective of Philosophy,” in
H. L. A. Hart, Essays in Jurisprudence and Philosophy. Oxford: Oxford University Press.
Jori, M. (2010). Del Diritto Inesistente. Pisa: ETS.
Kelsen, H. (1967). Pure Theory of Law. Berkeley and Los Angeles: University of California
Press.
Kelsen, H. (2000). “Preface: On Interpretation,” in H. Kelsen, The Law of the United
Nations: A Critical Analysis of its Fundamental Problems. New Jersey: The Lawbook
Exchange Ltd, 2000.
Kramer, M. (2008). “Review of Arthur Ripstein (ed.), Ronald Dworkin,” Notre Dame
Philosophical Review, 1.
Leiter, B. (2003). “Beyond the Hart–Dworkin Debate: The Methodology Problem in
Leiter, B. (2007). “Why Quine is not a Postmodernist,” in B. Leiter, Naturalizing
Jurisprudence. Oxford: Oxford University Press.
Leiter, B. (2009). “Explaining Theoretichal Disagreement,” University of Chicago Law
Review, 76: 1215–50.
Luque Sánchez, P. (2011). “Dos problemas de una respuesta positivista al desafío de los
desacuerdos teóricos sobre el derecho,” in P. Comanducci and R. Guastini (eds.), Analisi e
Diritto 2010. Madrid: Marcial Pons.
Quine, W. V. (1960). Word and Object. Cambridge, MA: MIT Press.
Ratti, G. B. (2008a). Sistema Giuridico e Sistemazione del Diritto. Turin: Giappichelli.
Ratti, G. B. (2008b). “The Consequences of Defeasibility,” in P. Comanducci and R. Guastini
(eds.), Analisi e Diritto 2007: Ricerche di giurisprudenza analitica. Turin: Giappichelli.
Ratti, G. B. (2009). “Los desacuerdos jurídicos en la jurisprudence anglosajona,” in
P. Comanducci and R. Guastini (eds.), Analisi e Diritto 2008. Madrid: Marcial Pons.
Ratti, G. B. (2012). Diritto, Indeterminatezza, Indecidibilità. Madrid: Marcial Pons.
Raz, J. (2009). “On the Nature of Law,” in J. Raz, Between Authority and Interpretation.
Oxford: Oxford University Press.
Ross, A. (1961). “Review to The Concept of Law,” Yale Law Journal, 71: 1185–90.
Russell, B. (1920). Introduction to Mathematical Philosophy. London: Allen & Unwin.
Schauer, F. (1991). Playing by the Rules: A Philosophical Examination of Rule-Based Decision-
Making in Law and in Life. Oxford: Clarendon Press.
Schauer, F. (1998). “On the Supposed Defeasibility of Legal Rules,” in M. D. A. Freeman
(ed.), Current Legal Problems 51: Legal Theory at the End of the Millennium. Oxford:
Schauer, F. (2012). “Is Defeasibility an Essential Property of Law?,” in J. Ferrer Beltrán and
G. B. Ratti (eds.), The Logic of Legal Requirements: Essays on Defeasibility. Oxford: Oxford
University Press.
Schauer, F. (2013). “Necessity, Importance, and the Nature of Law,” in J. Ferrer Beltrán,
J. J. Moreso, and D. M. Papayannis (eds.), Neutrality and Theory of Law. Dordrecht:
Springer.
Shapiro, F. (2007). “The Hart–Dworkin Debate: A Short Guide for the Perplexed,” in
A. Ripstein (ed.), Ronald Dworkin. Cambridge: Cambridge University Press.
Tarello, G. (1980). L’interpretazione della legge. Milan: Giuffrè.
Twining, W. (2009). General Jurisprudence: Understanding Law from a Global Perspective.
Van Fraassen, B. (1977). “The Only Necessity is Verbal Necessity,” Journal of Philosophy,
74: 71–85.
Williams, B. (2006). Ethics and the Limits of Philosophy. London: Routledge.
14
Is There One Right Answer to the
Question of the Nature of Law?
Dan Priel *
Sir James Stephen is not the only writer whose attempts to analyze legal ideas
have been confused by striving for a useless quintessence of all systems, instead
of an accurate anatomy of one.
(Holmes 1897: 475).
In Europe, charters of liberty have been granted by power. America has set the
example . . . of charters of power granted by liberty.
( James Madison, quoted in Bailyn 1992: 55).
1. Introduction
For some time during the late 1970s and early 1980s Ronald Dworkin’s claim that
there are uniquely right answers to virtually all legal questions was the subject of
heated debate. But as often happens with such debates, it eventually ran out of
steam, even though neither side succeeded in convincing the other. It may therefore
seem a bit odd to return to this topic now. If I do so, it is because I wish to examine
it from an angle that I think it has not yet been looked at, and which I think is
relevant to some debates that are very much alive these days. More specifically,
I will argue that there is something about the right answer thesis that fits a certain
view prevalent in American political culture about politics and (consequently)
about the relationship between law and politics, and that this view is quite different
from the British mainstream view about law and politics. The more general and
more important goal of this chapter is to challenge the underlying idea of the search
for the nature of law, at least so long as it is understood as a conceptual inquiry that
purports to tell something about law outside the different political and cultural
environments in which different legal systems operate. Somewhat surprisingly,
* I thank participants at the Nature of Law Conference at McMaster University for their comments,
and especially to my commentator at the conference, Kevin Walton, and to Stefan Sciaraffa, for his
detailed written comments.
Is There One Right Answer to the Nature of Law? 323
many of the very same legal theorists who have argued vigorously against the view
that there is a single right answer to legal questions accepted (or actually, assumed)
this position at the meta-level inquiry on the nature of law. I think this is
assumption is mistaken, and my chapter is an attempt to explain why.
2. Changing places
Returning to his native Britain after a year in the United States, Hart recorded his
impressions for the BBC.1 These were then published in the BBC’s magazine The
Listener. This brief piece, “A View of America,” reflects Hart’s astonishment with
what he saw there. Coming from a country that was still slowly recovering from two
wars that had left it practically bankrupt, Hart was clearly enthralled about the tall
buildings, the wide roads, the big cars, and the bountiful universities. But apart
from that, the essay also contains some pertinent comments about American
politics and how different they were from what he was used to at home. One
thing that particularly impressed Hart was
the passion inspiring so many whom you meet for the moral issue. Argument soon breaks
through to what is believed, apparently, to be at the root of every problem—a moral
problem. And, more widespread than I could have believed, was the conviction that just as
there lurks at the bottom of almost every problem a moral question, so there must somewhere
be an answer, an answer perhaps for the sage—and he may be in the university—to provide.
(Hart 1958: 89)2
This is also a fairly neat summary of Dworkin’s right answer thesis. This description
would not be so surprising to those who know something about the debates that
would come to dominate legal philosophy in the following decades had it not been
for the fact that Hart wrote this some five years before Ronald Dworkin published
his first article. And yet Hart was not talking here not about American law but on
American politics. This, as I will try to show, is not a coincidence.
Unlike Hart, who only made a few visits to the United States, Dworkin lived and
taught in Britain for many years. But he always remained, I think, a relative outsider
to the legal and political culture in Britain. He has been, for example, a frequent
commentator on American constitutional cases in the New York Review of Books,
but to the best of my knowledge has never written a similar essay on a British case.
He has seen enough, however, to note the differences in the political cultures of the
two countries, writing once that “public debate in the United States is dominated,
to a degree British commentators find surprising, by discussion of what rights
1 Though what I write about may be more accurately described as English, throughout the chapter
I mostly talk about Britain and British law. Those troubled by this term should read “England” and
“English law” instead wherever these terms appear. Though less loaded, “American law” is also not free
from difficulties. Again, I stuck with this phrase aware of its imprecision. I believe that in neither case
does my choice of word get in the way of my argument.
2 Surprisingly, even though Nicola Lacey discusses this little essay at some length in her biography
of Hart (2006: 193–6), she does not mention this remarkable passage.
324 Dan Priel
people have” (Dworkin 1985: 31–2). Dworkin’s most detailed discussion of British
law and politics is found in a relatively unknown pamphlet (or is it a book?) entitled
A Bill of Rights for Britain. Summarized in a sentence, in this essay Dworkin
recommends that Britain’s law and politics be turned into something more or
less like their American counterparts. He writes there that in order for people to
“participate in government . . . constantly through informed and free debate . . .
Parliament must be constrained in certain ways in order that democracy be genuine
rather than a sham” (Dworkin 1990: 33, emphasis in original).3 Such a change, he
suggested, might make “[l]aw and lawyers . . . begin to play a different, more
valuable role in [British] society than they now even aim to have.” It would
encourage courts to “think more in terms of principle and less in terms of narrow
precedent.” Such a change would help foster “a legal profession that could be the
conscience, not just the servant, of government and industry” (Dworkin 1990: 23).
In short, Dworkin recognized that there are fundamental differences in the political
culture of Britain and the United States, and that those had fundamental implica-
tions for the way lawyers perceived their role in society.
3. Two ways to the rule of law

I want to suggest that these reflections on the political differences between the
United States and Britain that both Hart and Dworkin noticed are relevant to legal
philosophy. At this stage I will phrase my argument as weakly as I can: since law
interacts with politics, the fact that different political communities have different
conceptions of politics is likely to have an impact on their conception of law as well.
I will examine this argument in the context of one aspect of the rule of law, namely
that of the requirement that the content of the law not be affected by the identity of
the person charged with expressing it. This sense is sometimes captured by the
slogan “the rule of law and not of men,” and is associated with one sense of
objectivity, that of non-perspectival constancy or invariance. Put in everyday terms,
it is the demand that if different people are asked to state what the law says on a
particular question, they will (typically) offer the same answer, even if their political
views are different. (This, of course, is not the only sense of the rule of law.)
When legal philosophers address the question they usually do so by trying to
connect debates about objectivity in law to the subtle and sophisticated discussions
of objectivity among philosophers where the topic has been front and center of
virtually every field of philosophy. These jurisprudential contributions provided
important correctives to those who rejected the possibility of objective knowledge
as naive or even reactionary. However, in focusing on the question of objectivity in
these senses they shifted attention away from the issues that lawyers are mostly
interested in. For lawyers do not usually wonder about whether Truth exists, or the
ontological building blocks of Reality, but they are intensely interested, practically
3 These days Dworkin (2011: 398–9) has somewhat softened his support for judicial review.
interested, in potential problems that could undermine objectivity in the sense

I associated with the rule of law. The worry is that people’s personal (possibly
unconscious) prejudices, their political opinions, or cultural background, will affect
their determination of the law thereby undermining the possibility of invariance
among different people.
The unique problem of objectivity in law can thus be summarized as follows: (1)
law should be (in the specific sense explained above) objective; (2) law (in modern
society) primarily is concerned, directly or indirectly, with the relationship between
individual and the state; (3) questions of this sort are often politically controversial;
(4) legal provisions are often written in vague language that can be interpreted in
different ways, and therefore (5) law’s objectivity is in danger. More informally, the
worry is that, especially in the modern welfare state, there is tension between the
aims for objectivity in the law and many of the issues it is made to handle.
Setting up the problem at this level of abstraction is fairly easy. What makes
finding a solution to it difficult is that the issue is not simply one of finding the
optimal design of legal institutions to minimize a clearly set problem. Part of the
difficulty arises from the fact that the understanding of the problem itself changes in
different political environments, and therefore that there are several possible ways of
understanding what would count as a successful solution to it. One way of
understanding my argument below is that we can identify a difference in the
mainstream legal academic positions in Britain and the United States not simply
about the means of achieving an agreed optimal level of separation between law and
politics, but on what kind of relationship between law and politics is desirable. To
complete my argument about the error in the way debates about the nature of law
are currently conducted I will need to show, first, that the relationship between law
and politics is relevant to debates about the nature of law; and second, that different
answers to this matter—and as a result different understandings of the nature of
law—are going to be given in different political cultures.
My argument below can be summarized as follows. I will first argue that the
British response to the problem identified above is to try and keep law away from
political questions, whereas the American response has seen various attempts for
allowing courts to address political questions without entering the political fray.
I will then argue that we can see exactly these different responses in the respective
works of Hart and Dworkin. The third step is to show that these competing views
mark competing understandings of what law is, which at least in part are the result
of competing understandings of politics.
3.1. The British response

The traditional, and still typical, response in Britain to the potential challenge that
politics poses to the rule of law is to insist on a categorical separation of law and
politics. “To call a court ‘political’ is merely to deny it the character of a court of
law” (Oakeshott 1975: 412). This distinction is maintained through the more or
less conscious adoption of various institutional mechanisms aimed at minimizing
the potential infiltration of politics into the law, and no less importantly, of law into
326 Dan Priel
politics. Some of these means for achieving that goal are rather familiar, others
perhaps less so. I mention only a few of them here:
(1) Parliamentary supremacy. In the course of a discussion of the differences
between the United States and Britain Hart wrote how the “defenders of the British
government’s policy had come to think not merely that Parliament had unlimited
legislative powers as a matter of English constitutional law but that it was a general[,]
necessary and indeed self-evident truth that government by law could not be
limited by law” (Hart 1976: 551). Hart’s point was that though the matter was
presented as a conceptual truth it was merely a reflection of a particular political
choice, one that could be easily refuted by pointing, as Hart himself did in his
critique of Austin, to the United States as a counterexample (Hart 1994: 72–4).4
The fact is, however, that though there were some who sought to present the matter
in this conceptual fashion, the opposition to legal limitations on Parliamentary
power was always thoroughly normative. To its proponents Parliamentary sover-
eignty was a good idea.5 One of the good things about Parliamentary sovereignty
was that it created a clearer separation between legal and political matters, and made
it less likely that courts would be embroiled in the sort of political questions that
could compromise their objectivity. This remains a fundamental pillar of British law.
(2) Judicial deference to government action. Another doctrine used to keep law
and politics apart has been a very deferential approach to executive action. The
general attitude of British courts towards a judicial mindset has been described by a
prominent legal historian as one of “total deference to the executive” (Stevens 1978:
320).6 A broad doctrine of justiciability, an almost uninterrupted discretion to
public authorities (under what British public lawyers call “Wednesbury unreason-
ableness”),7 and narrow scope for tort liability of public authorities, have all served
to keep executive action relatively free from legal oversight. Of course, one can find
examples of all that in other places as well, but in comparison to courts in other
Western democracies, and more specifically in comparison with American courts,
in much of the twentieth century British courts were remarkably timid (Stevens
2005: 31–46). British courts did not, until the last few years, develop a detailed
protected rights jurisprudence that set limits to legislative and executive action, as
these were deemed inimical to Parliamentary sovereignty.
(3) A non-political process of judicial appointments. Judicial appointments in
the UK are made on the basis with relatively little influence by politicians, and are
based on an assessment of the legal ability of the candidates. This, of course,
contrasts sharply with the fashion in which judges in the US are typically
4 Many others have made this point against Austin before Hart, e.g. Bryce (1901: 89–94); Sidgwick
(1897: 25–8).
5 A summary of the overwhelmingly political arguments against such limitations is found in
Goldsworthy (2001: 233–4).

6 Also Stevens (1978: 388–9): “parliamentary sovereignty implied that there should be no serious
judicial questioning not only of acts of Parliament, but of decisions of the Civil Service.”
7 So called after Associated Provincial Picture Houses v. Wednesbury Corporation, [1947] 1 KB 223
(CA).
appointed, either in popular elections (in most states) or by a highly politicized

nomination process (in the federal courts).
(4) Focus on doctrine in legal education and legal scholarship. What does it mean
to be a good lawyer? In the best British law schools this still means, to a degree that
would astound American law professors, mastering a vast amount of legal rules by
reading a breathtaking number of cases, while giving relatively little consideration
to questions of social policy. In a kind of feedback loop mechanism doctrinal legal
education is both a product of and a mechanism for maintaining the separation of
law from politics: today’s students are lawyers and judges of the next generation,
and they continue to litigate, interpret, and develop the law in the same fashion.
This is not merely a matter of legal education: doctrinal, “black-letter” scholar-
ship—ranging from the brief case note to the comprehensive treatise—is a genre
still accorded great respect among British academics. One of the marks of this kind
of scholarship is the view that the solution of legal problems is still primarily to be
derived from legal materials. As a result even what counts as “theoretical” work in
Britain is geared much more towards conceptual analysis or the internal taxonomy
of legal categories than to the examination of the law from an interdisciplinary
perspective and it often lacks the explicitly normative orientation which is more
common in American legal scholarship. This kind of scholarship fosters the view
that law and politics are separate domains: social scientific data or political theory
belong to politics, “internal” answers derived from the analysis of legal materials are
the domain of lawyers.
3.2. The American response

The American scene is different. It was already in the first half of the nineteenth
century that Tocqueville famously wrote “[t]here is hardly a political question in
the United States which does not sooner or later turn into a judicial one” (Tocque-
ville 2003: 315 [}I.2.8]).8 And this enchantment with law as a central location for
political battle has greatly intensified since.
One source of all these differences is the United States Constitution, understood
primarily as a legal document that empowers courts to declare certain legislation
unconstitutional and that has no equivalent in Britain. Another difference was
pointed out by Hart in the very same essay from which I quoted above:
For [the English], surely, liberty is this: that there is a circle round each man, inside which he
can do as he please, and it is no concern of others; this is the liberty the Englishman has
inside his house and garden and behind its hedges. I think that this as an ideal makes little
appeal to an enormous number of Americans; I believe you can find what the American
means by liberty by looking at the Constitution of an American State. In the State of
Massachusetts the Constitution provides that any member of the public may introduce a
8 It is true that some have questioned Tocqueville’s claim (Graber 2004; cf. Schauer 2005), but
though it is true that not all political questions become judicial questions in the United States, it is true
that American courts have been involved in many politically contested questions, especially in
comparison to British courts.
328 Dan Priel
measure into the Legislature and argue for it before committees. And it seems to me that this
is what an American means by liberty; the right to take part in what he would call “the
decision-making process”. (Hart 1958: 89)9
Simply put, politics is a more participatory affair in the US than in Britain, and
going to court is one way of participating in politics in the United States. To be
sure, there have been many critics, both popular and academic, of the excessive
legalism in American life, complaining about too many laws, too many lawyers, and
too much faith in the redeeming powers of law. But these persistent complaints
need to be repeatedly made exactly because of the staying centrality of law in
American political life. They need to be mentioned against a background of
statements such as “[l]itigation is an important political activity: courts exercise
political authority, modify substantive laws, and allocate resources. . . . Litigation
declares and changes fundamental cultural values” (Abel 1987: 454–5).10 (It is hard
to imagine such a statement being made in the British context.)
The result of all this is a very different understanding of the relationship between
law and politics in American law from the way this relationship is understood in
Britain. This makes the fears of law becoming indistinguishable from politics more
pressing; at the same time it implies that many of the institutional solutions for
maintaining the rule of law that were adopted in British law are simply not available
in the United States, and it requires the adoption of different institutional solutions
to match the very different problems that arise when courts actively engage in
politics. Once again I can only mention briefly a few of them.
(1) Giving up. Perhaps the most radical response to the difficulty is to give up
on the attempt to maintain the separation between law and politics. The nomin-
ation proceedings of Supreme Court justices are now televised and nominees are
assessed in quite openly political terms. In many states judges are elected and can be
removed from office by the public. There is a thriving industry of scholarship
produced by political scientists assessing court decisions in terms of their political
orientation, and in response to the challenge that the law is political some scholars
have even openly celebrated the contribution “political courts” make to the
American polity (e.g. Peretti 1999). The situation in Britain could hardly be
more different. As one of the few political scientists to take an interest in the
workings of the courts put it, “[i]t is possible for students of British government to
pursue their studies in blissful ignorance of the contents of even the existence of the
law reports” (Drewry 1985: 373; Drewry 2009). Compared with literally hundreds
9 Once again Hart captured here Dworkin’s view years before he himself articulated it. Dworkin
(2011: 365) contrasts one view of liberty according to which liberty consists in the idea that “people
must be permitted to play a role in their own coercive government: that government must in some
sense or another be self-government” with another according to which “people must be free of coercive
government over some substantial range of their decisions and activities.” Dworkin, of course, defends
the former.
10 Or consider the following words from Bogus (2001: 41): “American common law is both more
democratic and more dynamic than its British counterpart. The people play a larger role in American
than in British common law, and the common law plays a larger role in the American than in the
British system of governance.”
of books and articles by political scientists examining the extent to which courts are
influenced by politics, there is an astounding dearth of studies of this sort in Britain.
This lack of scholarly interest can only be explained by the assumption that law and
politics are largely separate.
(2) Theories of interpretation. It is not often noted, but there is no British equi-
valent to American lawyers’ obsession with theories of interpretation. This work, at
both the statutory and constitutional level, is, at bottom, an effort to find a way to
be able to identify the content of law in a way that both acknowledges the political
content of the law, but tries to keep law apart from it. This is evident in approaches
such as originalism and textualism, which are presented as the only antidotes to the
dangers of politicization of the law. But the same is true of other approaches too.
Dworkin, for example (whose approach I discuss below), addresses this difficulty by
acknowledging the political character of the law, but argues that the right inter-
pretive methodology can avoid the dangers associated with political courts. These
debates are required because of, and can be understood as institutional solutions to,
the role of American courts in politics.
(3) Economic analysis of law (and “theoretical” scholarship more generally). Economic
analysis of law has proven a remarkable success in the United States, but it has not
had much impact in Britain. It is tempting to explain this disparity by appeal to the
American obsession with theory and contrast it with English anti-intellectualism,
but there is, I think, a deeper reason for this difference. In the British context
economic analysis is perceived as an alien influence that could potentially lead to
undermining the integrity of the law (its “self-understanding”). It is thus not merely
pointless, it is potentially dangerous. I suspect that the success of economic analysis
of law has much to do with the need to maintain objectivity in the face of law
deeply infused with politics. Since doctrinal analysis is no longer perceived as
sufficiently robust to prevent the politicization of the courts—this is the realist
legacy that has never had much impact in Britain—economic analysis is (or was)
conceived as a non-political alternative. At least in its initial stages, it was promoted
for its ability to provide precision and objectivity that traditional methods of legal
reasoning lacked.
* * *
I could list many other ways in which the fundamental differences in understanding
the relationship between law and politics manifest themselves in these two legal
systems. Political activism in the United States is quite often legal activism; there is
much less of that in Britain. (Of course, this may only reflect political activists’
prediction that such a course of action would be a waste of time and money; but
that is exactly the point.) In addition one may mention the attitude towards judicial
review, the different attitude towards deciding politically controversial matters
(abortion, slavery, health care), and consequently, the very different place of law
in public discourse (US Supreme Court decisions are commonly front page news,
but those of its British counterpart much more rarely). If I were to summarize the
difference between the two positions in a sentence I would put it this way: the study
of the American Constitution is, by and large, a study of American constitutional
330 Dan Priel
law; the study of the British constitution is, to a much greater degree, the study of
British politics.
No doubt the sketch I offered above is very schematic, and on each side there is
now growing appreciation of the virtues of the opposing view. In Britain even
before the adoption of the Human Rights Act there were those who sought to find a
basis for judicial review of legislation in the principles of the common law. In the
United States, after years in which judicial review was perceived as the greatest gift
American politics has bestowed on the world, there are now those who argue that
the US will be better off without it. But in both cases these are still minority views
and they are battling very entrenched ideas. It is also possible that British law may
be criticized by showing that all the institutional mechanisms mentioned above that
are aimed to keep law outside politics fail to do so, and that in fact British law is no
less political than American law. And yet at least at the level that legal philosophers
seem to be interested in, that of officials’ attitudes from the internal point of view,
the “self-understanding” of British law(yers) is fundamentally different from that of
American law(yers).
4. Back to the jurisprudential fray

I suspect the response of many legal philosophers to what I have said so far will be
that it is all very interesting, perhaps even true, but that this is not jurisprudence,
and has little to do with its concerns. For the general view among them is that “in
spite of many variations in different cultures and in different times, [law] has taken
the same general form and structure” (Hart 1994: 240). Most legal philosophers
accept this statement without argument and have consequently left all the matters
I discussed in the last section to others: this is history, or politics, or perhaps
constitutional theory or comparative law; this is not jurisprudence properly so
called.
What I will try to show is that these matters are relevant to jurisprudence, and
indeed have in fact been at the background of some well-known jurisprudential
debates. I begin by examining the respective views of Hart and Dworkin and show
that they reflect very different views on the relationship between law and politics,
and ones that are in accordance with the dominant view on the matter in Britain
and the United States, respectively.
4.1. The different ways of understanding the connection

between law and politics
I began by distinguishing between Hart and Dworkin’s perceptions of the other’s
native country and I suggested that their respective theories of law may have been
influenced by their perception of the law in their own native country. This
suggestion is not usually well received among legal philosophers. Michael Moore,
for instance, considered it the sort of mistake one occasionally hears from students
who are just embarking on their jurisprudential studies (Moore 2000: 80–1; for a
similar attitude see Endicott 1998). Nonetheless, I will try to demonstrate that
there may be deep and important truth in this view.
As the quotes above demonstrate, Hart has clearly recognized the differences
between American and British political cultures, and has also recognized the very
different role of politics in American and British law. None of this, apparently,
made him question the existence of a single nature to law. His brief references on
the interaction between law and politics, made in the context of discussing
American law (Hart 1983: 124–5) did not suggest to him that they had anything
to do with what law is. And yet, aspects of his general account reflect the British
approach to this relationship which he imperceptibly generalized to claims about
the nature of law in general. One place where this is particularly evident is Hart’s
discussion of adjudication. It is, as he later admitted (Hart 1994: 259) a fairly
cursory discussion (which is in itself revealing and unlikely to happen to an
American only a few years after Brown v. Board of Education). What is found
there is striking, for when Hart has to explain the sources of indeterminacy in the
law, he focuses almost exclusively on the inherent vagueness of natural language
(Hart 1994: 127–9): It is clear that a car is a kind of vehicle, but it is less clear
whether a bicycle is. In other words, the main source of disagreement is not political
divergence at the foundation of the legal system (or in his terminology, different
people having different rules of recognition due to their different political back-
grounds), but rather a local, linguistic disagreement at the periphery of law. It is
only when law runs out and the judge is required to fill a void when there is no
more law to follow that she will have to consult her moral and political views.
Thus, the British approach to the problem of objectivity mentioned above—that
of keeping law and politics separate—becomes a central tenet in Hart’s account of
the nature of law. Matters are very different with Dworkin. For him the constant
contact between law and politics is perhaps the central background problem that
calls for solution, the major threat to the possibility of objective law. It is for this
reason that he begins his critique of Hart by pointing out that when disagreements
occur in the law their source is usually political, not linguistic. Dworkin therefore
starts his analysis at a point that those who wish to maintain the separation of law
from politics would already find unacceptable: “Judicial decisions,” he says, “are
political decisions” (Dworkin 1978: 88). Thus, the separation between law and
politics, in the way it is understood in British law—two (largely) non-overlapping
domains—is off the table from the start. Instead, Dworkin seeks to solve the
problem of objectivity by relying on a distinction between what I will designate
as “politics” and “Politics.” (The terms are mine.) The former is what one sees on
TV: spin, personal attacks, simplistic slogans, talking points instead of argument; it
is also the kind of things that give politics a bad name: logrolling, filibustering, the
adoption of unprincipled position for the sake of personal promotion. For politics
Dworkin has nothing but contempt. Debates preceding “legislative decision[s]
about some great issue of principle” are “rarely” of “high quality.” Rather, such
“[o]rdinary politics generally aims . . . at a political compromise that gives all power-
332 Dan Priel
ful groups enough of what they want to prevent their disaffection, and reasoned
argument elaborating moral principles is rarely part of or even congenial to such
compromises” (Dworkin 1996: 344–5; also Dworkin 1985: 146). By contrast,
Politics is commitment to principle through the exploration of the moral require-
ments involved in living in a community. He says, for instance, that in a commu-
nity governed by “the model of principle” members of the community “accept that
they are governed by common principles, not just by rules hammered out in
political compromise.” And he adds: “Politics has a different character for such
people. It is a theater of debate about which principles the community should adopt
as a system . . . ” (Dworkin 1986: 211).11
What is the role of law in this? Dworkin’s answer is that law, properly under-
stood and practiced, is true Politics. The first step in the argument is that “our
government shall be republican rather than despotic” (Dworkin 1996: 111, also
345, for a similar emphasis on a “republican” view of law), and the recognition that
what I called “politics” “destroys civic republicanism” (Dworkin 2000: 234, also
369, where Dworkin links the “degraded and negative” political discourse in the
US to the claim that Americans cannot claim “with a straight face, to be governing
[them]selves”). The second step is that law is the main means for promoting this
republican ideal:
We have an institution that calls some issues from the battleground of power politics to the
forum of principle. It holds out the promise that the deepest, most fundamental conflicts
between individuals and society will once, someplace, finally, become questions of justice.
I do not call that religion or prophecy. I call it law. (Dworkin 1985: 71)12
Law and lawyers thus have a dual role within the American polity: one is to elevate
politics to Politics by providing a model of what political discourse should look like;
the other is to limit the domain of politics by transferring certain questions from
politics to law (and thus to Politics) if politics cannot be transformed into Politics.
The first strategy is in view in Law’s Empire when Dworkin argues that political
decision making should be based on the same principles and reasoning that guide
legal decision making (Dworkin 1986: 184–6; Dworkin 2006).13 Elsewhere
Dworkin explains that political institutions could rise to the level of Politics only
under what he calls “partnership democracy,” which, as its name indicates, is
distinctive for its participatory view of democracy—one that works to encourage
the kind of political debate he sees in judicial proceedings.
However, when the political community fails to create this form of democracy
Dworkin is clear in favoring the second strategy: “individual citizens can . . . exercise
11 For comparison consider Crick (2000: 32), a book that has a good claim to be the classic view of
politics in Britain: “ ‘Politics’ . . . simply summarizes an activity whose history is a mixture of accident
and deliberate achievement. . . . It is not as such motivated by principle. . . . Political principles are,
whatever they are, principles held within politics.” See also Crick (2000: 47).
12 Dworkin (1996: 345) offers a (highly romantic) explanation of how law achieves this. Friedman
(2004: 1290–1) makes precisely the same point.

13 I criticize the position of Dworkin (2006) in Priel (2007a).
the moral responsibilities of citizenship better when final decisions involving

constitutional values are removed from ordinary politics and assigned to courts,
whose decisions are meant to turn on principle, not on the weight of numbers or
the balance of political influence” (Dworkin 1996: 344; also Dworkin 1978: 85).
The judge’s role in this debate is not that of a follower (of rules set elsewhere), but
primarily as that of a moral leader. For this reason we even find in Dworkin, an
exhortation for (as Hart put it) “the sage . . . [who] may be in the university” to
provide the right answer to political questions.14
Accepting that, the question still remains how we can know—even in a partner-
ship democracy—that such debates would give us the objectively right answer to
the questions of Politics. The answer, quite simply, is that the right answer is the
best answer to come out of the debate. This crucial point about Dworkin’s position is
central to his thinking: the right answer is not “out there,” it is simply what emerges
as the “winner” in a principled debate. It is always a tentative winner, because it is
always open to challenge, but an objectively right answer means, for him, nothing
more than the best answer so far found, the one that we have been convinced cannot
be successfully challenged, in proper political (i.e. Political) discourse.
What is often missed in jurisprudential discussions of Dworkin’s work is the
extent to which his account reflects a view that is fairly familiar within American
legal discourse. To see this we must step outside the narrow bounds of analytic
jurisprudence, although because of space limits I can provide here only one detailed
example.15 In an essay first published in 1981 (now reprinted in Fiss 2001) Owen
Fiss has sought to address the challenge to objectivity posed by certain then-
emerging critical scholars. Proponents of this “new nihilism” contended, as Fiss
put it, that “[a]ll law is masked power” (Fiss 2001: 151). Fiss’s response was
somewhat different from Dworkin’s: unlike the latter who denied the intelligibility
of external criticism Fiss acknowledged it and even relied on it to distinguish
between internal legal criticism and external ethical (or religious) criticism. But
what looks at first like a significant difference turns out to be quite small, for as Fiss
put it, “[e]ndogenous change is always preferred . . . [and therefore t]he external
critic struggles to work within the law . . . ” (Fiss 2001: 159; also 160–2, for an
elaboration on the ways in which appeal to morality is required to maintain the
objectivity of law).
One need not accept Fiss’s romanticized view of American law to recognize in his
words a particular vision of addressing the challenge to the objectivity of law, one
14 Recall: “The courts are the capitals of law’s empire, and judges are its princes, but not its seers and
prophets. It falls to philosophers, if they are willing, to work out law’s ambitions for itself, the purer
form of law within and beyond the law we have” (Dworkin 1986: 407). To the same effect, Dworkin
(2011: 109).
15 Another famous example that is superficially very different but bears some deep unmistakable
resemblance to Dworkin’s ideas is Wechsler (1958), contrasting “the ad hoc in politics” with the
“judicial process . . . that . . . must be genuinely principled” (1958: 15), and who, interestingly, con-
trasted his approach to that taken by the House of Lords (1958: 17). There are also looser links
between Dworkin’s idea and much other work that has sought to reconcile the republican ideal of
self-government of “We the people” of the Constitution with the higher law, the “unalienable Rights,”
mentioned in the Declaration of Independence, and the role of law in answering this question.
334 Dan Priel
that seeks to find it through engagement in moral and political debate within an
“interpretive community” (Fiss 2001: 170). The judges’ authority derives from the
fact that they “are insulated from the political process” (i.e. what I called politics),
but that is so in order for them “to engage in a special kind of dialogue over the
meaning of [public] values” (i.e. what I called Politics). This judicial engagement
with public values is “an essential part of the process through which a morality evolves
and retains its public character” (Fiss 2001: 199). For this reason “courts should not
be viewed in isolation but as a coordinate source of governmental power, and as an
integral part of the larger political system” (Fiss 2001: 54, also 34–5).16
The view reflected here, then, is not simply that morality contains right answers
to moral questions, and since morality is part of the law, law contains right
answers too. (Note also that it has nothing to do with whether morality is a
condition of validity.) At bottom Dworkin’s view holds that law can transform
political debate and in this way avoid the problem of objectivity: following his
prescription is what it means to find the right answer to moral questions. If one
accepts this view the audacious right answer thesis becomes trivial, you might even
say tautological. But in the Hartian perspective it is almost incomprehensible.
Because Hart does not recognize the distinction between two kinds of politics, on
his view when politics is introduced into the law, political disagreement is simply
“reproduced” in the law. It is hard to see how that could aid the search for legally
correct right answers or guarantee the objectivity of law. This view is evident when
Hart criticizes those who think that “there is no central element of actual law to be
seen in the core of central meaning which rules have, that there is nothing in the
nature of a legal rule inconsistent with all questions being open to reconsideration
in the light of social policy” (1983: 72, emphasis in original; Shapiro 2011: 327–9,
argues against Dworkin on the basis of a similar attitude). For Hart to recognize the
existence of law is to recognize a domain in which politics, even Politics (“social
policy”), cannot touch.17
4.2. Legal validity?

What we have seen so far is that the relative separation of law and politics in English
law replicates itself in Hart’s account of the nature of law, whereas the deep ties
between law and politics in American law are at the heart of Dworkin’s theory of
law. That this fact has not received much attention is not, I think, a coincidence.
Unlike the connections between law and morality, the interaction between law and
16 Another similar suggestion is found in Barber and Fleming (2007: 156): the right approach to
constitutional interpretation, they say, “can only mean an interpretation of the Constitution that tries
to redeem its expressed claim to be an instrument of justice, the general welfare, and the other goods
listed in the Preamble.” Such examples could be easily multiplied.
17 The two jurisprudential views thus reflect two opposed views on the place of law in practical
reasoning. Hart’s positivism sees law as concerned with guidance, and clear guidance requires replacing
moral reasons. Dworkin’s view sees law as concerned with participation and ultimately in self-
government, and hence sees the point of law in engaging people in moral reasons. The issue is explored
more fully in Priel (2012b).
politics has been marginalized in analytic jurisprudence. Perhaps this reflects the
dominant view that morality is constant and therefore the relationship between it
and law is (fairly) constant too. (This is true also of moral anti-realist legal
philosophers such as Kelsen or Alf Ross.18) As such the law/morality borderline
seems to match an inquiry into the nature of law (as currently understood) much
more than the shifting boundary between law and politics. But in ignoring the law/
politics divide legal philosophers may have missed the most important challenge to
their search for a single nature to law. The purpose of this section is to show in what
way the differences identified above between British and American law about the
place of politics, and the corresponding differences between Hart and Dworkin,
reflect a difference that does not merely reflect different variations on a single nature
of law, but must be understood as competing answers to the question “What is law?”
To the extent that the question of the relationship between law and politics is
addressed in jurisprudential literature it is typically located in the rule of recognition.
According to a popular view (found already in Hart 1994: 71, 106, 247) whereas the
Americans have adopted a constrained sovereign, the British have not. Translated to
the language of contemporary jurisprudence this means that moral considerations
make up part of the American rule of recognition but not part of the British one.
There are various problems with this claim, not least that it assumes that what is
at stake between competing jurisprudential theories are different perceptions of
legal validity.19 In the present context, this view fails to capture the sense in which
the British constitution is fundamentally a political, not a legal, entity (it is difficult
to talk about a “document” in this context). It is a constitutional structure in which
“[e]verything that happens is constitutional. And if nothing happened that would
be constitutional also” (Griffith 1979: 19).20 It is from this that law is understood
as simply the voice (the “command,” if you don’t mind the Austinism) of whoever
happens to be in power. Sovereignty on this view is the power to make laws. It is a
power that the otherwise unlimited sovereign can choose to create legal limits on its
law-making power. This is, effectively, the way Hart interprets the US Constitution
(Hart 1994: 68–9), and this is how today the Human Rights Act is reconciled with
the doctrine of Parliamentary sovereignty. By contrast, in the United States
sovereignty is ultimately understood to vest in the people (Wood 2002: 159–62),
and the Constitution, created by the people, is thought to award some limited
powers, including limited law-making powers, to the different branches of
government. As such the law is understood fundamentally not as the manifestation
of political authority, but as what sets its limits (cf. Bailyn 1992: 175–81, 185–8,
201–3). In the British picture judicial engagement in politics is ultimately an
18 Though popular (especially among philosophers), the view that morality is largely unchanging is
not universally accepted. My view is that Dworkin, despite his endorsement of moral objectivity,
rejects this position. This by itself is a significant aspect of his work, that is very relevant to understand-
ing his jurisprudence, but not one I can consider here.
19 I challenge this view in Priel (2011a).
20 Griffith’s terms are stark, but the idea is an old one. Bagehot expressed a similar idea when he said
(1963: 221) that “The ultimate authority in the English Constitution is a newly-elected House of
Commons.” On the decline of the idea of fundamental law in Britain see Gough (1961: 168–207).
336 Dan Priel
illegitimate attempt to usurp political power, to undermine the sovereign; in the
American one, judicial engagement in politics is an (almost) inevitable outcome of
the need to make sure government does not exceed its given powers.
This is the philosophical divide that makes intelligible so many of the differences
between law and politics between the two countries. These two views emanate from
two very different underlying political theories about what makes the use of
political power legitimate, and they shape the most basic features of what law is
in these two countries. Yet in the present picture all this is treated as irrelevant to
understanding the nature of law, quite possibly exactly because these differences are
the result of competing political theories. But once we see this, much of what I said
above that was inexplicable in the “validity-first” picture that is dominant in
contemporary jurisprudence begins to make sense. It is not simply that the United
States and Britain have different tests of validity. It is that they have different
political traditions, which result in a different understanding of what law is, which
in turn explains numerous practical differences between their legal systems.
5. The implications for the search for the nature of law
John Gardner once described Dworkin’s work as that of a “theoretically ambitious

lawyer” (Gardner 2004: 173), suggesting that for all that is valuable or interesting in
it, it is not really legal philosophy. I suspect my argument would be classified in the
same way: indeed, exactly because my discussion reveals such differences between
American and British law it might be judged irrelevant to the question of the nature
of law. I suspect what I have said so far might be similarly dismissed as insufficiently
philosophical. There were just too many facts, too much history, perhaps even too
much law, to count as proper legal philosophy. These are exactly the kind of
contingent matters that legal philosophers consider to be the domain of others.
Rather than a reason for criticizing legal philosophers, one might say my discussion
so far shows that they have been correct to ignore all these matters.
In proper philosophical style, then, let me present a syllogism:
(1) There is no right answer to what constitutes “true” politics, or no single
account of the legitimacy of the state; there is, if you wish, no non-political
“nature” to politics to be discovered through conceptual analysis.
(2) Different states have different forms (“conceptions”) of politics that reflect
their different values, traditions, and history.
(3) There are connections between law and politics.
(4) Different understandings of politics result in different roles, understandings,
forms (“conceptions”) of law in different states.
(5) The different conceptions of law go to the heart of what on any plausible
view of the matter counts as the nature of law.
Hence: (6) The search for the (single, unique) nature of law is misguided.
I did not argue for (1), but I assume (and hope) it would not be controversial.
Denying it is hardly going to help the search for the nature of law, for that would
only mean that for the sake of an account of the latter, legal philosophers (unless
they deny (3)) would have to develop a side-specialty of identifying the true
nature of politics. In any case, my demonstration of (2) above can be seen as
validation of (1) as well. (3) too is, I think, uncontroversial. It is the remaining
premises, (4) and especially (5), that are likely to evoke most resistance, but my
argument in the last section has shown, I hope, the very different understanding
of politics in the two countries, and how they result in a profoundly different
understanding of what law is. To reiterate just one point noticed by Hart, the
dominant idea of freedom in Britain saw it as the opposite of law; in the United
States, freedom was a product of law. This is a fundamental divide that explains
not just the fundamentally different understandings of what counts as law, but
also many of the lower-level differences between law in these two countries. The
alternative advocated here to the prevailing jurisprudential orthodoxy claims that
these facts must figure in any attempt at depicting the “nature” of American and
British law. I argue that any inquiry that purports to account for law’s “self-
understanding” but cannot account for these differences, fails in the task it sets for
itself.
Those who wish to reject my argument might contend that at best what it
shows is the need for making discussions on the nature of law more abstract.
Now, of course, one may define the term “nature of law” any way one wishes
and it is definitely possible to define it to mean those necessary features we
happen to find in all legal systems. Even then I think my argument above
shows that one is likely to find much less than most legal philosophers seem to
think. More importantly, even if the nature of law simply means those things
that are necessarily true of all legal systems, one would then wonder why
this question is of any interest: Is philosophy in the business of collecting
necessary truths about things in the world? The search for the nature of law was
presumably considered worth engaging in because it was believed that by
engaging in it one could learn something valuable about a certain social
practice; it was believed that the search for the nature of law was somehow
illuminating of the law. The solution suggested here avoids the challenge posed
by giving the term “nature of law” a purely technical meaning and by stripping
this kind of inquiry of value. This solution to my challenge, then, looks more
like an admission of defeat.
If the search for the nature of law is to be meaningful, and if it is to bear some
resemblance for what most people mean when they talk about the “nature of X” or
“essence of X,” namely an inquiry that gives us the most significant aspects of a
certain phenomenon, such an inquiry should be understood in quite different
terms. I propose, informally, that a fact belongs to the nature of a thing if it
explains many other facts about it. Slightly more formally, a fact F is part of the
nature of social practice S if (and only if?) it figures in an explanation of many
features of S. Even the more formal definition is still imprecise, but for my purposes
here it will do. I propose that the way the relationship between American and
338 Dan Priel
British law and politics is part of each legal system’s nature, as defined here, for one
thing we have seen above is that the relationship between each legal system and
politics was part of the explanation of very many aspects of each; and though I have
said nothing about it, they are also relevant for explaining some of the differences
between the content of the laws of those legal systems.
All this does not amount to a logical proof, but it does amount to a demonstra-
tion that two legal systems have a very different understanding of what law is as a
result of their different political traditions and their corresponding differences in
understanding the relationship between law and politics. There is, however, a
logical challenge I can direct at defenders of the nature of law project: the issues
raised in this chapter can be deemed irrelevant only if American law and British law
belong to the same kind, otherwise the differences between them are simply irrele-
vant to the question of each legal system’s respective nature. In other words, there is
a different way of understanding my argument, i.e. as the claim that American law
and British law are different kinds, in which case the set of facts I described about
each law may be necessary for understanding its nature but irrelevant to under-
standing the nature of the other. Put differently, my point is that when proponents
of the prevailing understanding of nature of law rely on examples from both British
and American law in support of their views on the nature of law, they assume
without argument that the two are species of the same kind. However, from a
logical point of view there is warrant in doing so and there is no telling whether they
are right on this matter. Since there is no fact of the matter on this question, the
assumption that they are—an assumption that is crucial for the entire enterprise of
the search for the nature of law—is unwarranted. Furthermore, this assumption can
only make sense if one implicitly assumes some further substantive views about what
law is. The whole enterprise is thus in some sense circular as it can only get off the
ground by presupposing the truth of certain premises that are presented at the end
of the inquiry as findings about the nature of law. (This problem is complicated
many times over when other legal systems, contemporary, historical, some may
even want to add hypothetical, are added to the story.)
To amplify this last point imagine you had been asked to offer an account of the
nature of American law or the nature of British law, a sort of account of the central
features of the legal system, one that includes the self-understanding of lawyers in
each of these legal systems as to the enterprise they are engaged in. We could
conclude that British and American law belonged to the same kind only at the end of
our inquiry if we found out that the accounts of each that we had previously offered
turned out to have a roughly identical core and that their differences were only
confined to those aspects that we had not considered as part of either legal system’s
core. But if we attempted to undertake such an inquiry and found out that the
accounts were conflicting at their core that would imply that these two legal systems
do not belong to the same kind. My claim in this chapter is that when we do just
that, it turns out that the two legal systems indeed conflict at their core. Of course,
I may be wrong about that, but until something like this is tried—and I know of no
legal philosopher writing on the nature of law who tried to do that—the entire
project rests on wholly question-begging assumptions.
Perhaps, however, we may understand the search for the “nature of law” more
weakly, i.e. as not actually concerned with finding the nature of law. Instead this
term might simply designate an attempt to look at law from some greater distance,
without any commitment to classificatory accuracy. Is the difference between my
views and those of legal philosophers looking for the nature of law nothing more
than that they prefer the long shot view whereas I am more interested in the close
up? I do not deny that standing at some “distance” from a phenomenon can reveal
what a closer look may miss (Jackson and Pettit 2004). Even understood in this
more relaxed way, however, there is a serious problem with it, for if this is the
reason for the search for the nature of law, then the right distance from the object of
inquiry is not to be determined by the object of inquiry. Rather, it must be chosen
by legal philosophers according to criteria that are external to the object of inquiry. As
on this view there is no correct degree of “zoom,” the only appropriate way of
choosing between various possible distances must be determined according to what
we gain from the explanation. If that is the case, my view is that in their search for the
nature of law legal philosophers have been standing too far from the law, and that
some of the debates that preoccupied legal philosophers in the last few years under
the banner of the nature of law were deeply unilluminating of their object of
inquiry. Don’t take my word for it: on this interpretation of the search for the
nature of law a good test for the optimal “distance” from an object of inquiry is the
degree of interest others find in it. The fact that both legal academics and moral and
political philosophers (let alone practitioners or lay people) have not found much of
interest in these debates, is more than suggestive of the fact that the distance from
the object chosen by legal philosophers has been the wrong one.
Opting for the distant explanation can also lead to the opposite problem, viz. the
tendency to assume the local and familiar is general and universal. I have already
suggested what may be a problem in the work of both Hart and Dworkin: an
assumption that one can extrapolate from the legal system one is familiar with to
universal claims about the nature of law. Let me illustrate this point with another
example. A few years ago there was an internet debate on the question why Anglo-
American legal scholars take relatively little interest in the work of Hans Kelsen and
in deontic logic, both topics on which legal philosophers from continental Europe
write much more. The debate attracted many prominent legal philosophers, who
offered various explanations.21 I read this exchange and was struck by the fact that,
with the exception of one fleeting comment (by an Italian, made with regard to
Kelsen), no one mentioned what I thought was the obvious explanation. Kelsen’s
legal thought, despite his many years in the United States, remained firmly rooted
in a particular conception of law that is closer to what one finds in civil law systems.
Once it is recognized that for all its abstraction Kelsen’s approach fits some legal
systems better than others, the puzzle is solved. The same is true of deontic logic.
Deontic logic would be considered important for analyzing legal argument in a legal
21 See <http://prawfsblawg.blogs.com/prawfsblawg/2007/10/why-no-deontic-.html>; <http://prawfs
blawg.blogs.com/prawfsblawg/2007/10/why-no-kelsen.html>; <http://leiterlegalphilosophy.typepad.
com/leiter/2007/10/why-dont-amer-1.html> (all visited, April 8, 2011).
340 Dan Priel
system that puts a premium on deductive argumentation. This is exactly what one
finds. A prominent German legal scholar has recently written that even today “[t]he
typical German judgment, like its French counterpart, strives after the ideal of
deductive reasoning” (Zimmermann 2005: 27, also 38; similarly Markesinis 1986:
366). By contrast, legal philosophers from common law jurisdictions, told from
Day One of law school that “[t]he life of the law has not been logic; it has been
experience” (Holmes 1881: 1),22 are much less likely to think that deontic logic is
going to capture anything important about the law.
I point out these vestigial localisms in supposedly general theories of the nature
of law not in order to castigate particular legal theories for their insufficient
generality, but to point out that legal philosophers’ tendency to stand far away
from legal practice, their relative lack of interest in the nitty-gritty details of the
organization of particular legal systems, all adopted in the name of the search for the
nature of law, may actually result in too quick generalizations from few familiar
cases and thus undermine the inquiry for the nature of law.
Perhaps noticing the difficulties with talking about the nature of law in general,
several legal philosophers have turned in recent years to talking about “our” concept
of law (e.g. Raz 2009: 94–5). Unfortunately, this seemingly more modest claim
suffers for the same problem of demarcation: Quite simply, there is no logical way
of demarcating “our” concept of law without circularity unless we have a non-
question-begging way of demarcating who “we” are, and none of those who
suggested this possibility has even hinted towards a way of addressing this question.
This is not a fanciful complaint. In challenging the search for the nature of law
I relied not on anthropological studies of small pre-industrial communities living in
the thick of the Amazon rainforests, not on the law of an ancient and now extinct
political community, not on the law in contemporary Russia,23 not even on the
distinction between common law and civil law that some have suggested involved
some fundamentally different ways of thinking (Legrand 1997). I looked at the
two legal systems from which the most prominent analytic legal philosophers
hail, two legal systems that share a long history. If “our” does not include both of
them, I do not see what this word is supposed to cover.
At this point the challenger may reply that I am wrong to say that we have no way
of knowing that American and British law are not the same kind: after all, people treat
them as belonging to the same kind, which is why, for example, they consider it a
valuable exercise to compare them. This response, however, is both unconvincing
and self-defeating. It is unconvincing, because it is asserted rather than shown. The
fact that people use the same word to describe both is not sufficient to tell us that the
word is used to refer to the same kind. (We are, after all, repeatedly reminded that
legal philosophy is not lexigraphy.) This usage is perfectly acceptable for everyday
22 Several American philosophers (e.g. Dewey 1924; Cohen 1916) have expressed similar views on
law. Many other statements expressing the same idea from English and American judges and scholars
are found in Waddams (2003: 1–2).
23 For example Ascherson (2004): “when [Putin] says ‘law’ he means what we would call ‘order’.”
The issue is discussed further in Kahn (2006) and Kurkchiyan (2003).

purposes, but it does not require commitment to the view that the things described
by the same word share certain necessary features. In fact, my guess is that if you
asked some British lawyers what they thought of American law, many of them would
reply “it has the appearance of law, but it really is politics masquerading as law,”24 or,
more politely, that American law is a very different beast from British law. This
answer is also self-defeating because if we take people’s attitudes seriously then we
should pay at least as much attention to their profound thoughts and self-understand-
ing of their own legal system as to their superficial comparisons with others. Put
differently, one cannot invoke people’s attitudes at one point and ignore them
completely at another. And it has been exactly my claim that those attitudes are
profoundly different in Britain and the United States.
6. Some further objections answered

I foresee four additional objections to the position taken here. The first is that the
argument must be wrong because the existence of a debate on the question of the
nature of law shows that there is something that the disputants are talking about.
Andrei Marmor (2001: 6) once wrote that “controversy over the content of a rule of
recognition does not prove that there is no such a rule. Even when there are several
ways of understanding a rule (or anything else for that matter), there must be
something there that people can understand differently and argue about.” Though
the context of this remark was different, Marmor’s parenthetical remark shows he
considers his point to be general. Transferred to our context the argument would be
that the fact that people disagree about the nature of law shows that such a thing
exists. Unfortunately, this is a bad argument. Controversy may be the result of
shared mistaken belief over the existence of something that in fact does not exist. No
one (I hope) would infer the existence of God merely from the existence of debates
about the nature of God. In our context, debates about the nature of law could exist
so long as participants share the mistaken belief that such a thing exists. The
purpose of this chapter is exactly to challenge this belief.
A second objection is the opposite of one of the objections considered above, i.e.
that I have been tendentious in my choice of two very different legal systems for
discussion. There is American exceptionalism (including American legal exception-
alism) on one side and British exceptionalism (in not having a written constitution)
on the other. If I had chosen different countries, I would not have gotten such
different outcomes. There are, however, several problems with this objection. One
is, of course, that in some respects the American and the British legal systems are
fairly close: they are both legal systems of Western developed countries with a long
shared history. And as already mentioned, these happen to be the countries from
which the most prominent analytic legal philosophers come from, and it is very rare
indeed to hear them thinking of either (or both) of these legal systems as somehow
24 I have not conducted a survey but Stevens (2009: 651) and Birks (1996: 98–9) come close to
saying just that.

342 Dan Priel
an outlier in the search for the nature of law (especially when the whole point of the
search for the nature of law is to find those features from which there are no
outliers).25 The most important point, however, is that the moment one starts
looking closely, virtually every country is “exceptional” in one way or another.
French law and German law are the paradigmatic examples of civil law jurisdic-
tions, and yet there are important fundamental differences between them; German
law has served as the model of much Chinese law, but there are fundamental
differences between German and Chinese law as well. This does not entail (nor did
I argue) that each legal system has its own “nature,” or even that nothing
illuminating can be said about law in general. It says, however, that in order to
identify without circularity what belongs to the object of inquiry one must identify
a certain role that law plays within a certain social or political organization. Because
of the variety of such organizations, it cannot be assumed in advance that the thing
called “law” in all of them will have the same nature. This is only something we
could find at the end of our inquiry.
This point leads to the third challenge, and that is that my argument in fact
shows that law has a nature, namely that the law is concerned with objectivity. The
critic may even go on to quote me saying that there is “tension between the aims for
objectivity in the law and its subject-matter.” Can I make this statement without
assuming law has some nature? This is an important point for clarifying the scope of
my argument. First, as already said, I do not deny that there are certain observations
true of both American and British law (and quite possibly of other legal systems).
There may even be observations true of legal systems more generally. But to observe
at legal systems and notice some things they have in common is not philosophy,
and it is best done with recognized fact-gathering techniques that, to put it gently,
are not the mark of contemporary legal philosophy. Second, there may be some
“philosophical” questions (whatever that means exactly) that arise with regard to
many, perhaps even all legal systems. But these “puzzles” (as I called them in Priel
2007b: 193–5), need not, singly or together, amount to anything that could
plausibly be called the nature of law (unless one uses this term in the purely
technical sense mentioned above), nor do these puzzles become less puzzling if
they are found to be true of only some legal systems. Most importantly, there is no
reason to think in advance that even if the same puzzle arises in all legal systems,
that the answer to it will be the same in all legal systems. To pick a favorite puzzle, it
is perfectly possible that we will have one explanation for the normativity of law in a
contemporary modern welfare-state democracy and quite another one in a medieval
feudal society. We have seen a different understanding of the relationship between
law and politics between British and American law; the differences may be even
greater in a place where an absolute monarch declares “the state is me.” Since law is
not something that falls on societies from the sky with a note “take me or leave me”
25 Atiyah and Summers (1987: 256, 411–20) is a rare (and perhaps for this reason neglected)
exception.
attached, it can be (and has been) molded according to the particular political
culture in which it exists. No one doubts the truth of this claim as far as the content
of law is concerned; what I do here is extend it to the deeper and more general
elements of different legal systems.
The last objection I can foresee is that what I have argued so far is a kind
of a skeptical claim. As such it may be an amusing intellectual game, but like other
skeptical claims we should really ignore it and go on with our daily business,
including philosophers’ daily business, which has always been the identification
of the nature of things. However, if my claim is “skeptical” in any sense, it is
skeptical only with regard to a particular, and in historical terms recent, under-
standing of jurisprudence. In the English-speaking world it was not before John
Austin that jurisprudence was understood in this way (and I would argue that even
he is a questionable case). Before him there is simply no one that I know of,
definitely not Bentham or Hobbes, who was concerned with the question of the
nature of law in the manner this term is understood these days. It was only around
fifty or so years ago that Anglophone legal philosophy turned in earnest to this
question (in German-language jurisprudence things may have begun a bit earlier);
this happened at a particular point in history, when other branches of Anglophone
philosophy were also interested in questions of a similar kind. And yet in other areas
these questions were fairly quickly abandoned: already in 1979 Brian Barry, who
knew a thing or two about the time and place in which this sort of inquiry
flourished, wrote mockingly about questions of this sort, “I cannot remember
when I last read a discussion about the criteria for a good cactus or an extra-fancy
apple” (Barry 1979: 632).26 It is, I think, no coincidence that the revival in political
philosophy, pronounced dead in 1960s, took place at the time that interest in such
“conceptual” questions declined, and by now, in many branches of philosophy the
search for the “nature” of things has come under attack and has been largely
abandoned.27
26 But unlike other areas of philosophy this sort of question seems to be alive and well in
jurisprudence (Dickson 2009). It is notable that it is mostly legal positivists who seem concerned
with these sorts of questions, whereas natural lawyers like Finnis and Dworkin dismiss them as
unimportant. This should have alerted legal positivists to the fact that these natural lawyers are not
in the business of giving an analysis of the nature of law, but rather of giving a political (non-neutral)
account of law. Many of the apparent tensions Dickson finds in the work of Finnis (Dickson 2009:
170–3), disappear if this is realized. In the case of Dworkin, the matter is discussed in more detail in
Priel (2011a).
27 See e.g. in metaphysics: Van Fraassen (2002): chapter 1 entitled “Against Analytic Metaphysics;”
in epistemology Stich (1990: 19–21): section entitled “The Irrelevance of Analytic Epistemology;”
Kitcher (1992); in ethics Darwall et al. (1992: 123): describing how “[m]oral philosophers shed the
obsessions of analytic metaethics;” Timmons (1999: 15–30): describing the move from analytic to
post-analytic metaethics); in philosophy of mind Fodor (2004); and generally Harman (1999). They
could all be wrong, of course, but to suggest that abandoning the search for the nature of law is
somehow anti-philosophical displays lack of awareness on what some of the world’s leading philoso-
phers think is (and is not) philosophy.
344 Dan Priel
7. From the nature of law to human nature?
Does this mean that it makes no sense to talk about the “nature” of law? If we mean
by this term what most legal philosophers have given it in the last fifty years, then
I believe the inevitable answer is “yes.”28 The only way to avoid the problems
raised here is to engage in normative inquiry. Whatever its faults may be such an
account would not suffer from the problem of circularity identified above. Such
an account can be more sociological or historical—seeking to understand the
development of law in a certain time or place, to answer what law was for those
living then, or it can be more directly normative, seeking to explain where law fits
within a political theory, what law must be in order to be legitimate. On certain
assumptions the latter account may be universal in the sense that it applies in all
places, not in the sense that it exists in all places. (Claiming universality for such
an account presupposes that there is one way law may be legitimate, and this
assumption may be false.)
Perhaps, however, there is a completely different way of thinking about the way
to get to a universal account of the nature of law. Perhaps the nature of law is not a
matter to be discovered by a priori reflection on law, but rather by looking more at
humans and their nature. This idea may sound new, but it is in fact one as ancient as
jurisprudence, and it is only the lack of interest of most legal philosophers in the
history of their subject (prior to 1961) that may have obscured this point. It is part
and parcel of the natural law tradition, and it is also central to the work of those
often considered the first legal positivists, Thomas Hobbes and Jeremy Bentham.
Since this organizing idea has been rejected, probably with the work of John Austin,
jurisprudence has been losing its way (Priel unpublished).29
A contributing factor for the lack of interest in human nature may have been the
skepticism towards the idea in the humanities and social sciences around the time
Hart published The Concept of Law. More recently, however, the idea of human
nature has undergone something of a revival. Building on work in psychology and
anthropology that identified numerous traits and habits found in all human
societies (Pinker 2002: 435–9, passim), there have been some attempts to extend
these ideas to law (Kar 2006; Guttentag 2009; Mikhail 2011). Without passing
judgments on any of the particular ideas defended in these works, I will say that
I am sympathetic to the view that scientific findings could contribute to questions
that have traditionally been thought to belong to philosophy (Priel 2011b; Priel
28 For the sake of completeness I should add that my argument here is not the only one I have
advanced against this sort of inquiry. Others appear in Priel (2007b, 2008). Against the “case study”
approach of the present chapter, the approach taken in these two essays in more abstract.
29 In Hart (1994: 193–200) we still see the remains of this approach in the brief discussion of the
minimum content of natural law. It contains some, broadly Humean, remarks about human nature,
but even these are mostly negative (i.e. an attempt to challenge what he considered to be mistaken
views about human nature espoused by natural law theorists) and their relevance for understanding the
content laws are likely to have, but not the basis for Hart’s positive account of law. Even this minimal
concern for human nature is gone in more recent work by legal positivists, who have told us that an
account of the nature of law must be able to explain what law is even for non-human societies.
2012a). It is worth bearing in mind, however, that at this stage the research this
work is based on is still controversial, and its relevance to jurisprudence raises
additional difficulties. I cannot discuss these issues here, but I wish to highlight two
points about how different this approach would be from the currently dominant
view in jurisprudence. First, these inquiries rely heavily on factual discoveries of
social and cognitive scientists, whose methods are very different from those used by
contemporary legal philosophers. Second, the substantive results of this approach
are also likely to look quite different from those of contemporary jurisprudence.
The “direction” of their argument is from human nature to the nature of law, rather
than from observations about law to the nature of law. Fully cognizant of the fact
that there are considerable differences between legal systems, this sort of approach is
unlikely to try to distil the few features that make something into a legal system, but
rather try and identify what facts about human nature entail (if anything at all)
about the shape legal institutions are likely to take. Though the aspiration is to
identify some such universal tendencies among humans, it does not require the
discovery of necessary truths.
What such inquiries might reveal about law is not a topic I can address in a few
pages. The only purpose of the discussion has been to show where I think legal
philosophers should be looking if they seek more universalist findings about law in
general. But ending on this note may seem to undermine the thrust of everything
I have said before. That, however, is not a challenge to me, but to reality. It reflects
one of the most difficult questions in the explanation of all of human affairs, namely
how to reconcile the fact of universal human nature with the fact of immensely
different cultures. Swings in popular and academic opinion on the matter—the
sixties and seventies were the “relativistic” decades, and now universalism is on the
up30—reflect the extent to which this complex question may be influenced by
shifting fashions, but it seems clear that there is quite a bit of both in all human
societies, and that a complete account of social institutions would have to take
account of both.
Legal philosophers may have thought that they could contribute to the universal
end of the inquiry by taking on questions that seemed to be independent of
scientific inquiry and even human nature: the “nature” of law is not a question
that seems amenable to scientific inquiry (Priel 2012a), but the difficulties with this
sort of inquiry, some of which have been shown here, make this alternative
unattractive. The universalist alternative proposed here avoids these difficulties,
but poses a different challenge to legal philosophers. Put bluntly, it seems to leave
little room for philosophical contribution now that the research on human nature
has largely been taken up by scientists. This means legal philosophers will have
to rethink what they are doing. One possibility is for them to serve as a kind of
30 Consider the following titles: The Language Instinct: How the Mind Creates Language (Pinker
1994): “The Moral Instinct” (Pinker 2008); The Art Instinct: Beauty, Pleasure, and Human Evolution
(Dutton 2010); The Faith Instinct: How Religion Evolved Why It Endures (Wade 2009); “The Property
‘Instinct’ ” (Stake 2004); “Is There a Law Instinct?” (Guttentag 2009); The Compassionate Instinct: The
Science of Human Goodness (Keltner et al. 2010).
346 Dan Priel
go-between trying to connect the general, scientific inquiry on human nature with
the diverse, normative examination of diverse social institutions. Beyond that, legal
theorists will have to turn to political theory and explain the place of law within it.
Thus, in the end, we reach a conclusion similar to the one I have made in the
beginning of this section: jurisprudents are most like to make a real contribution by
adopting a more consciously normative approach to legal theory, one that relies on
scientific findings on human nature for the sake of a better theory of the proper role
of law. To do that, however, legal philosophers must be willing to open up their
discipline both to political theory and to scientific work on human nature.
8. Conclusion
This has been a story with a twist. Much of the work in jurisprudence of the last fifty
years has been concerned with the discovery of objective truth about the “nature” of
law. What I have argued here is that to find the answer to this question one needs to
understand the way the idea of objectivity in law is understood, and how different
political traditions have led to different characterizations of law in response to the
problem of objectivity. If one accepts that there is no one way of understanding the
relationship between law and politics and that the way this relationship is understood
touches upon and affects the most basic aspects of what law is, then it follows that
there is also no single answer to the question of the nature of law.
Why has this conclusion eluded some very sophisticated thinkers? One reason
must be that the debate about the nature of law has been framed as a debate about
the relationships between law and morality. But the focus on the relationships,
necessary or otherwise, between law and morality has obscured from view questions
about the relationships between law and politics, questions which arguably are far
more important for understanding law in the modern regulatory state. Understand-
ing the law–politics relationship calls for more attention to law’s daily workings—
the way judges understand their position in relation to other branches of govern-
ment, the way judges are appointed, the way they write their opinions, the way law
is taught, as well as the way political elites perceive the right way of solving social
problems. It requires a broader view of what law does: It is no coincidence, I think,
that when legal philosophers write about substantive law they focus on common
law areas, those that fit the law–morality divide far better than areas such as
securities regulation or administrative law (which highlight the law–politics divide).
One of the unfortunate implications of the focus on the “nature” of law within
jurisprudence is that those working in jurisprudence have come to see all this as
irrelevant detail, not sufficiently interesting or important for serious philosophical
reflection. The result has been a growing gulf, almost a conceptual separation,
between law and jurisprudence. If nothing else, I hope this chapter has shown that
this has not helped make jurisprudential debates more philosophically sophisticated
or better focused, and it has not helped to get us closer to finding the general features
of law.
References
Abel, R. L. (1987). “The Real Tort Crisis: Too Few Claims,” Ohio State Law Journal, 48:
443–67.
Ascherson, N. (2004). “Law v. Order,” London Review of Books, 20 May, 22–4.
Atiyah, P. S. and Summers, R. S. (1987). Form and Substance in Anglo-American Law:
A Comparative Study of Legal Reasoning, Legal Theory, and Legal Institutions. Oxford:
Clarendon Press.
Bagehot, W. (1963). The English Constitution. Edited by Richard Crossman. Second
edition. London: Fontana (second edition first published in 1872).
Bailyn, B. (1992). The Ideological Origins of the American Revolution. Enlarged edition.
Cambridge, MA: Harvard University Press.
Barber, S. A. and Fleming, J. E. (2007). Constitutional Interpretation: The Basic Questions.
New York: Oxford University Press.
Barry, B. (1979). “And Who Is My Neighbor?,” Yale Law Journal, 88: 629–58.
Birks, P. (1996). “Equity in Modern Law: An Exercise in Taxonomy,” University of Western
Australia Law Review, 26: 1–99.
Bogus, C. T. (2001). Why Lawsuits Are Good for America: Disciplined Democracy, Big
Business and the Common Law. New York: New York University Press.
Bryce, J. (1901). Studies in the History of Jurisprudence, vol. 2. Oxford: Clarendon Press.
Cohen, M. R. (1916). “The Place of Logic in the Law,” Harvard Law Review, 29: 622–39.
Crick, B. (2000). In Defence of Politics. Fifth edition. London: Continuum.
Darwall, S., et al. (1992). “Toward a fin-de-siècle Ethics: Some Trends,” Philosophical
Review, 101: 115–89.
Dewey, J. (1924). “Logical Method and Law,” Philosophical Review, 33: 560–72.
Dickson, J. (2009). “Is Bad Law Still Law? Is Bad Law Really Law?,” in M. Del Mar and
Z. Bankowski (eds.), Law as an Institutional Normative Order. Burlington, VT: Ashgate,
161–83.
Drewry, G. (1985). “The GCHQ Case: A Failure of Government Communications,”
Parliamentary Affairs, 38: 371–86.
Drewry, G. (2009). “A Political Scientist’s Perspective,” in Louis Blom-Cooper et al. (eds.),
The Judicial House of Lords 1876–2009. Oxford: Oxford University Press, 439–53.
Dutton, D. (2010). The Art Instinct: Beauty, Pleasure, and Human Evolution. Paperback
edition. New York: Bloomsbury.
Dworkin, R. (1978). Taking Rights Seriously. Revised edition. Cambridge, MA: Harvard
University Press.
Dworkin, R. (1985). A Matter of Principle. Cambridge, MA: Harvard University Press.
Dworkin, R. (1990). A Bill of Rights for Britain. London: Chatto and Windus.
Dworkin, R. (1996). Freedom’s Law: The Moral Reading of the American Constitution.
Cambridge, MA: Harvard University Press.
Dworkin, R. (2000). Sovereign Virtue: The Theory and Practice of Equality. Cambridge, MA:
Dworkin, R. (2006). Is Democracy Possible Here? Principles for a New Political Debate.
Princeton: Princeton University Press.
Dworkin, R. (2011). Justice for Hedgehogs. Cambridge, MA: Harvard University Press.
348 Dan Priel
Endicott, T. A. O. (1998). Book review, Law Quarterly Review 114: 511–15.
Fiss, O. (2001). The Law as It Could Be. New York: New York University Press.
Fodor, J. (2004). “Water’s Water Everywhere,” London Review of Books, 21 October:
17–19.
Friedman, B. (2004). “The Importance of Being Positive: The Nature and Function of
Judicial Review,” University of Cincinnati Law Review, 72: 1257–303.
Gardner, J. (2004). “The Legality of Law,” Ratio Juris, 17: 168–81.
Goldsworthy, J. (2001). The Sovereignty of Parliament: History and Philosophy. Oxford:
Gough J. W. (1961). Fundamental Law in English Constitutional History. New edition.
Oxford: Clarendon Press.
Graber, M. (2004). “Resolving Political Questions into Judicial Questions: Tocqueville’s
Thesis Revisited,” Constitutional Commentary, 21: 485–545.
Griffith, J. A. G. (1979). “The Political Constitution,” Modern Law Review, 42: 1–21.
Guttentag, M. D. (2009). “Is There a Law Instinct?,” Washington University Law Review,
87: 269–328.
Harman, G. (1999). “Doubts about Conceptual Analysis,” in Reasoning, Meaning, and
Mind. New York: Oxford University Press, 138–43.
Hart, H. L. A. (1983). Essays in Jurisprudence and Philosophy. Oxford: Clarendon Press.
Hart, H. L. A. (1958). “A View of America,” The Listener, 59: 89–90.
Hart, H. L. A. (1976). “Bentham and the USA,” Journal of Law and Economics, 19: 547–67.
Hart, H. L. A. (1994). The Concept of Law. Second edition. Oxford: Clarendon Press.
Holmes, O. W. (1881). The Common Law. Boston: Little, Brown.
Holmes, O. W. (1897). “The Path of the Law,” Harvard Law Review, 10: 457–78.
Jackson, F. and Pettit, P. (2004). “In Defence of Explanatory Ecumenism,” in Mind,
Morality, and Explanation: Selected Collaborations. Oxford: Clarendon Press, 163–85.
Kahn, J. (2006). “The Search for Rule of Law in Russia,” Georgetown Journal of International
Law, 37: 353–409.
Kar, R. B. (2006). “The Deep Structure of Law and Morality,” Texas Law Review, 84:
877–942.
Keltner, D., et al. (eds.) (2010). The Compassionate Instinct: The Science of Human Goodness.
New York: W. W. Norton.
Kitcher, P. (1992). “The Naturalists Return,” Philosophical Review, 101: 53–114.
Kurkchiyan, M. (2003). “The Illegitimacy of Law in Post-Soviet Societies,” in D. J. Galligan
and Marina Kurkchiyan (eds.), Law and Informal Practices: The Post-Communist
Experience. Oxford: Oxford University Press, 25–46.
Lacey, N. (2006). A Life of H. L. A. Hart: The Nightmare and the Noble Dream. Oxford:
Legrand, P. (1997). “Against a European Civil Code,” Modern Law Review, 60: 44–63.
Markesinis, B. S. (1986). “Conceptualism, Pragmatism and Courage: A Common Lawyer
Looks at Some Judgments of the German Federal Court,” American Journal of
Comparative Law, 34: 349–67.
Marmor, M. (2001). Positive Law and Objective Values. Oxford: Oxford University Press.
Mikhail, J. (2011). Elements of Moral Cognition: Rawls’ Linguistic Analogy and the Cognitive
Science of Moral and Legal Judgment. New York: Cambridge University Press, 2011.
Moore, M. S. (2000). Educating Oneself in Public: Critical Essays in Jurisprudence. Oxford:
Oakeshott, M. (1975). “The Vocabulary of a Modern European State (Concluded),”
Political Studies, 23: 409–14.
Peretti, T. J. (1999). In Defense of a Political Court. Princeton: Princeton University Press.
Pinker, S. (1994). The Language Instinct: How the Mind Creates Language. New York:
HarperCollins.
Pinker, S. (2002). The Blank Slate: The Modern Denial of Human Nature. London: Penguin.
Pinker, S. (2008). “The Moral Instinct,” The New York Times (Magazine), 13 January.
Priel, D. (2007a). “In Search of Argument,” Texas Law Review, 86: 141–64.
Priel, D. (2007b). “Jurisprudence and Necessity,” Canadian Journal of Law and
Jurisprudence, 20: 173–200.
Priel, D. (2008). “The Boundaries of Law and the Purpose of Legal Philosophy,” Law and
Philosophy, 27: 643–95.
Priel, D. (2011a). “The Place of Legitimacy in Legal Theory,” McGill Law Journal, 57:
1–35.
Priel, D. (2011b). “Jurisprudence and Psychology,” in Maksymilian Del Mar (ed.), New
Waves in Philosophy of Law. London: Palgrave Macmillan, 77–99.
Priel, D. (2012a). “Jurisprudence between Science and the Humanities,” Washington
University Jurisprudence Review, 4: 269–324.
Priel, D. (2012b). “Are Jurisprudential Debates Conceptual? Some Lessons from
Democratic Theory,” Osgoode Hall Law Journal, 50.
Priel, D. (unpublished). “Towards Classical Legal Positivism,” available at <http://ssrn.com/
abstract=1886517>.
Raz, J. (2009). Between Authority and Interpretation: On the Theory of Law and Practical
Reason. Oxford: Oxford University Press.
Schauer, F. (2005). “The Supreme Court 2005 Term Foreword: The Court’s Agenda—and
the Nation’s,” Harvard Law Review, 120: 4–64.
Shapiro, S. J. (2011). Legality. Cambridge, MA: Harvard University Press.
Sidgwick, H. (1897). The Elements of Politics. Second edition. London: Macmillan.
Stake, J. E. (2004). “The Property ‘Instinct’,” Philosophical Transactions of the Royal Society
London B, 359: 1763–74.
Stevens, R. (1978). Law and Politics: The House of Lords as a Judicial Body, 1800–1976.
Chapel Hill: University of North Carolina Press.
Stevens, R. (2005). The English Judges: Their Role in the Changing Constitution. Revised
edition. Oxford: Hart Publishing.
Stevens, R. (2009). “Torts,” in Louis Blom-Cooper et al. (eds.), The Judicial House of Lords
1876–2009. Oxford: Oxford University Press, 629–52.
Stich, S. P. (1990). The Fragmentation of Reason: Preface to a Pragmatic Theory of Cognitive
Evaluation. Cambridge, MA: MIT Press.
Tiedeman, C. G. (1890). The Unwritten Constitution of the United States: A Philosophical
Inquiry into the Fundamentals of American Constitutional Law. New York: G. P. Putnam’s
Sons.
Timmons, M. (1999). Morality without Foundations: A Defense of Ethical Contextualism.
New York: Oxford University Press.
350 Dan Priel
Tocqueville, A. de (2003). Democracy in America and Two Essays on America. Translated by
Gerald E. Bevan, edited by Isaac Kramnick. London: Penguin (first published 1835).
Van Fraassen, B. C. (2002). The Empirical Stance. New Haven: Yale University Press.
Waddams, S. (2003). Dimensions of Private Law: Categories and Concepts in Anglo-American
Legal Reasoning. Cambridge: Cambridge University Press.
Wade, N. (2009). The Faith Instinct: How Religion Evolved and Why It Endures. New York:
Penguin Press.
Wechsler, H. (1958). “Toward Neutral Principles of Constitutional Law,” Harvard Law
Review, 73: 1–35.
Wood, G. S. (2002). The American Revolution: A History. New York: Modern Library.
Zimmermann, R. (2005). “Characteristic Aspects of German Legal Culture,” in M. Reimann
and J. Sekoll (eds.), Introduction to German Law. The Hague: Kluwer Law International,
1–51.
Index
Abel, R. L. 328 authorization 78

acquiescence 153, 174–6 autonomy 140
Adler, M. D. 292
Alchourrón, C. E. 312 Bagehot, W. 335
Alexander, L. 52, 203 Bailyn, B. 322, 335
Alexander, L. and Schauer, F. 292 Barber, S. A. and Fleming, J. E. 334
ameliorative concept 232–5, 237–9, 243 Barry, Brian 343
importance of 247–8 Bentham, Jeremy 130, 281, 309, 343–4
analogy, see nature of law Berger, P. L. and Luckmann, T. 90, 92
analytical imperialism 284 Besson, S. 308
analytical jurisprudence 7–9, 231, 279, 281, Binder, G. 60
284–6, 293–4, 297–8, 333, 335 biological kinds 235
schools within xiii, xiv Birks, P. 341
see also legal system Bix, Brian 6–7, 231
anthropology 251, 268 Blackburn, S. 275
Aquinas, St Thomas 14–17 Blake, Michael 184
Arendt, H. 86 blasphemy 15, 260–1, 274–5
Aristotle 11–13, 202 Bobbio, N. 130
art, philosophy of 220–1 Boghossian, P. 276
Ascherson, N. 340 Bogus, C. T. 328
assertions 12 boxing 105–6
assurance Brazil 292
coercion and 191, 193, 198, 200–1 Bryce, J. 326
rights and 196–8 Buchanan, A. 293
Atiyah, P. S. and Summers, R. S. 342 burden of proof 160
Austin, John xvii, 50, 104, 130, 152, 155, Burge, T. 235, 242, 245
158–9, 161, 174, 183, 209, 213–16, Burgess, G. 92
231, 244, 248–9, 281, 283, 286–7,
326, 335, 343–4 Campbell, T. D. 79
weird account: Canadian legal system 286, 291
of commands 17 canons 316–7, 319
of sovereignty 17 Cardozo, Benjamin N. 103, 113–15
Australian Legal Philosophy Association 3 Carnap, Rudolf 237–8
authority, of law 96 Catholic religious practices 250
choice-of-evil defenses 59–67 see also religion
claim of preemption to law 51–73 Celano, Bruno xviii, 144
compliance 73 chemistry 216, 235
de facto authority 76–89 Chiassoni, P. 314
idea of law and 89–97 Chinese law 342
methodological fallacy of 76–80 choice of evil defenses 53, 59–67, 72
‘gappiness’ of the law 67–72 definition of 60
legal philosophy and 97 citizenship 333
legitimate 62, 76 coercion xviii, 103–4, 130, 153, 165, 167,
normative fallacy 75–9, 81–6 183–205, 220, 223
rational capacity and 73 coercive enforcement in modern judicial
Raz’s service conception of 53–9 practice 178–80
rightful authority 76–89, 96–7 entitlement to coerce 183–205
different concepts of 87–8 consequences of 205
methodological fallacy of 76–80 free-riding 199–203
semantic externalism 86–9 in ideal societies 191–8
see also de facto authority, theory of; justification for 191–3
idea of law moral error and compliance 194–6
352 Index
coercion (cont.) criterialism 245
in non-ideal societies 199–204 Culver, K. and Giudice, M. xxi, xxii, 279,
rights and assurance 196–8 282, 297
unjustified lawbreaking,
response to 203–4 Dancy, J. 260
law without coercion, conceptual possibility Darwall, S. et al. 343
of 185–90 de facto authority, theory of 75–6, 96
legal system, conditions for the existence concept of 88
of 186–90 idea of law and 89–97
see also legal system; moral error; power legal philosophy 97
cognitive dissonance 139 moral convictions 91–4
Cohen, M. R. 340 positivism and non-positivism 94–7
Coke, Sir Edward 112 practical matrix 89–91, 93–4
Coleman, Jules 155, 165 plurality of models 86–9
Coleman, J. and Simchen, O. 236, 245 rightful authority and 76–89
comparative jurisprudence, see jurisprudence; methodological fallacy of 76–80
United Kingdom; United States normative fallacy of Raz 80–6
compensation 117, 196–8 semantic externalism 86–9
compliance 175–6 see also authority, of law
conceptual analysis of law (CAL) xviii, 154 deductive argumentation 340
concept of law 210–1, 238, 251, 301–2 defect 4–8, 161, 178, 283, 310
conceptual analysis 209–13 non-defectiveness conditions 4
pluralism 239–41 deference 81–2, 88, 97, 175
marriage 242–3 definitions, philosophical 104, 107, 251
tardy 241 explicative definition 237
conceptual analysis, projects of 233–9 stipulative definition 238
ameliorative project, the 237–9, 247–8 democracy 95, 106, 211, 233, 332
conceptual project, the 233–5 see also politics
descriptive project, the 235–6 deontic logic 339–40
conceptual and descriptive strategies applied to see also logic
law 243–7 dependence thesis 53–4, 81, 84–5
etiology of ‘law’ 209 descriptive concept 246
manifest vs operative concepts of desuetude 179
law 232–3, 240 Dewey, J. 340
operative concept, indeterminacy of Dicey, A. V. 291
249–50 Dickson, Julie 7, 10, 106, 233, 236–7, 239,
philosophical analysis and 230–51 241, 245–8, 250, 303, 343
prescriptive conceptual analysis 232 disagreement, see legal disagreements
reduction, possibility of 214–17 disambiguation 237–8, 258
revisionist claims 212 dissonance 139
theorist, role of the 250–1 see also cognitive dissonance; conflict
use of the word ‘law‘ 93 distributive justice 205
conflict 139 doctrinal concept of law 185, 305, 315–17
see also dissonance see also interpretation
conformity 175–6 Dolcetti, A. and Ratti, G. xx, xxi
consent 78 domination 76, 80–2, 86
consequentialism 238 see also legitimate domination
constitutional ability, see weak Drewry, G. 328
natural law thesis Duhem-Quine thesis 318
constitutional conventions 291 see also Quinean philosophy
constitutional law 282–3, 334 Dummett, Michael 276
continuity problem 297 Dworkin, Ronald xi, xii, xx, 44–6, 92,
of legal systems 285–9, 297–8 94–6, 102–8, 111–12, 115, 184–5,
contradictory injunctions 140 203, 212–13, 218, 220, 223, 225–6,
conventionalism 105 233–4, 238, 240–1, 245, 247,
Cotterrell, Roger 283 249–50, 265, 267, 270–1, 273,
counterfactuals 305 302, 308–9, 312, 324–5, 328–34,
Crick, B. 332 336, 343
crime 244 disagreement, argument from 303–5
Index 353
Hart-Dworkin debate 304–6, 319 Foldy, M. 260
Hartian response to 46–9 formalism, see legal formalism
interpretive theory of law Foucault, M. 144
nature of law, on the 323–4 Frank, Jerome 103, 109
right answer thesis 323 Frankena, W. 163
Shapiro‘s response to 46 freedom 196–7, 337
Dyzenhaus, D. and Taggart, M. 94 free-riding 199–203
see also coercion
economics 111–12 Frege, Gottlob 211
economic analysis of law 329 French law 342
Edmundson, W. A. 51 Friedman, B. 332
egalitarianism 205 Fuller, Lon 15, 28, 92, 94, 103, 115, 119, 130,
Ehrenberg, Kenneth xiv, xv, xvi, 13 132–3, 141
Eklund, M. 260, 262, 274–5
eliminativism 238 gaps in law 121
Elstein, D. and Hurka, T. 259, 262, 265 conflicting laws 71
Elster, J. 143 gap theorists 70–1
eminent domain/ takings doctrine 117 vague laws 71
Endicott, T. 293, 331 see also legal system
enforcement Gardner, John 9, 19, 336
coercive 178–80 gender 233
England 160, 323 German language 107, 343
English law 323, 334 German law 342
English Revolution (1640) 92 God, nature of 341
see also United Kingdom God’s law 94
Enoch, D. 276 Goldsworthy, J. 19, 326
Enoch, D. and Toh, K. xix good faith 190–1, 193, 201
entailment 238, 272 Gough, J. W. 335
Entailments of Practices Conditional 33–4 Graber, M. 327
epistemic values 246 Great Britain, see United Kingdom
epistemology 343 Green, L. 53, 156, 175
equivalence thesis 160–1 Greenawalt, Kent 25, 30
Erehwon 313 Greenberg, M. 18, 218
error theory 18, 160 Griffith, J. A. G. 335
essentialism 296 grounds of law 304–6, 312
in perspective 222–5 Guastini, R. 309–10
ethical particularism 260 guilt, see rule of law: compliability
ethical positivism 238, 243 Guttentag, M. D. 344
ethics 79, 95, 112, 343
European constitutions 140 Habermas, Jürgen 308
evolutionism 316 Hacking, Ian 90
exceptio probat regulam in casibus non exceptis 122 Hare, R. M. 245, 269
exceptionalism 341 Harman, G. 343
exclusionary reasons Hart, H. L. A. 95, 103–4, 109, 114, 136–8,
expressivism 216, 268 183, 209, 212–16, 225–7, 233, 236,
externalism 235 244–5, 247–51, 258, 266, 269, 279,
natural kind 235 281, 283, 287, 290, 294–5, 302,
objective type 236 307, 309, 314, 324–6, 328, 330,
social 235, 242, 245 335, 337, 344
see also semantic externalism adjudication, discussion of 331
biography of 323
failure of closure argument 57 core/penumbra dichotomy 310
federalism, see legal system in defense of 22–50
Festinger, L. 139 Dworkin, response to 46–9
Finnis, John xiii, 7, 27, 92, 94, 130, 132, 233, existence of law 334
240–1, 249–50, 286, 302, 343 Hart-Dworkin debate 304–6, 319
Fiss, O. 333–4 idea of law 94
Flores, Imer xiii, xiv individualized reasons and social
Fodor, J. 343 norms 33–7
354 Index
Hart, H. L. A. (cont.) as exclusionary 166
norm guided vs habitual behavior 34–7 reason and 166
Shapiro’s examples 36–7 social obligation, binding character of 166–9
smoking, example of 35–6 social pressure, notion of 169–70, 173
wrong conditional 33–4 theory of legal obligation 169–72
internal point of view 268 first-order legal obligation, explaining 170–1
legal concepts, nature of 41–4 first-order legal obligation, positivist
legal obligations, alternative conception account of 177–8
of 43–4 legally obligated vs legally obliged 171–2
morality, dimensions of 41–3 modification of HTSO 172–7
legal norms 153 second-order legal obligation 169–70
legal positivism, see legal positivism Haslanger, Sally 230, 232–3, 235–8, 241, 251
legal system, existence of 311 hermeneutics 96, 246
methodological positivism 234 Hill, Justice 64–5
nature of law, on the 301, 323–4 Himma, Ken xvii, xviii, 296
obligation and authority, concepts of 41 history 111–12, 231
Rule of Recognition xi, 17, 22–5, 310–11 Hobbes, Thomas 86, 130, 183, 201, 279, 343–4
complex hybrid, as a 28–31 Hohfeld, W. N. 29
complex structure of 29–30 Holmes Jr., Oliver Wendell 102–3, 108–15,
conception of 23 123, 322, 340
differing addressees 27–8 Hom, C. 275
differing norms 27–8 Hughes, Robert
disagreement over 44–9 human nature
evaluation of 50 nature of law and 344–6
Hartian response to Dworkin 46–9 Humean philosophy 344
Hart’s own words 23–4 Hume’s Law 37–8, 40
Hart’s undeceivedness 25 Hurd, H. M. 136
legal validity in 24, 45 Hurley, S. L. 51
power-conferring character of 29–30, 50 hypothetical necessity
puzzlement 26–7 anti-positivist implications of 16–20
role/function of 24–5 metaphysical neccessity 17
rule-individuation, theory of 26 soundness of 20
rule-skepticism 311 weak natural law thesis and 11–16
unity of, overarching 26–8 see also weak natural law thesis
simulative vs internal statements 37–40
social norms 153 idea of law
social rules and regularities xvii, 31–3, 228 de facto authority and 89–97
theory of law 312 moral convictions and 91–4
undeceivedness of 25 positivism and non-positivism 94–7
United States, impressions of the 323 practical matrix of 89–91, 93–4
see also Hartian Theory of Legal Obligation; see also authority, of law
Hart’s theory of social obligation (HTSO) Identity Question 10, 14, 17–18
Hart, H. M. Jr. and Sacks, A. M. 95, 115 illocutionary acts 12
Hartian Theory of Legal Obligation 152–82 Implication Question 10
coercive enforcement, centrality of 178–80 imprudence 36–7
concept of obligation 161–5 inference 274, 276
binding force of 180–2 instrumentalism 237
mandatory prescriptions and 162 International Court of Justice (ICJ) 280
obligations as binding 164–5 interpretation
obligations as exclusionary 163–4 constitutional 334
obligations as reasons 162–3 interpretanda 318
natural necessity 158 interpretata 318
normativity 154–7 interpretative game 314
theoretical significance of of legal sources 314–17
equivalence thesis 160–1 meta-interpretative doctrines 318–19
obligation talk, centrality of 157–60 stricto sensu 315
Hart’s theory of social obligation theories of 329
(HTSO) 165–77 see also legal disagreements
social norms 165 interpretivism 238
Index 355
allure of 217–22 non-state legality 283
good law, making of 221 state-centered legal theory 281–5
see also reductionism state legitimacy 155
Italian legal system 305 legal authority 156
legal directives 63, 70, 72–3, 81, 84
Jackson, Frank 234 legal disagreements 302
Jackson, F. and Pettit, P. 339 background 301–2
Jori, M. 306 dual nature of law and 301–19
jurisprudence 111 Dworkin’s argument 303–5
integrative 115 evaluation of 319
naturalizing 219 general jurisprudence and 317–18
justice 113, 332 Hart’s legal positivism and 308–12
see also distributive justice interpretative disagreements 301–2
justiciability, doctrine of 326 legal sources, identifying and
justification defenses 59 interpreting 305–6
see also normal justification thesis (NJT) agreement on legal sources 313–14
interpretative disagreement on 314–17
Kahn, J. 340 taxonomy of 306–8
Kant, Immanuel 101, 107, 183, 196 axiological disagreements 308
Kantian philosophy 238 decision-making disagreements 307–8
Kar, R. B. 344 interpretative disagreements 307
Kelsen, Hans xix, 25, 94–5, 130, 214–15, 225, source-recognition disagreements 306–7
267, 272–3, 279, 281, 286–7, 317–18, legal formalism 108–9
335, 339 legal gaps, see gaps in law
Kennedy, Duncan 103 legal idealism 105
knowledge 233 legal indeterminacy 225, 297, 331
Kramer, M. xvi, 19, 22, 28–9, 37–8, 41, 45, 47, legal judgments 38, 41, 258, 264, 267–8, 271,
51, 130, 132, 308 273
Kurkchiyan, M. 340 external 268–9
internal 268–9
Lacey, Nicola 323 see also legal statements
Lamond, Grant 183 legal obligation, see Hartian Theory of Legal
Lance, Mark 4 Obligation
Langdell, C. C. 108 legal philosophy 94–5, 97, 219–20, 259, 277,
Langlinais, Alex 160, 176 324, 330, 335–7, 339–41, 345–6
language 112, 213 legal pluralism 79, 280–1, 284, 294–7
canonical referents 235 challenge of 281–5
language games 224, 227 legal positivism 8–9, 19, 82, 86, 92–3, 95, 105,
linguistic competence 223, 227 156, 171, 218, 226, 286, 302, 344
linguistic deferentiality 236 Austin’s account of xvii
linguistic usage of experts 236 authority and 86
natural language 227–8, 263–4 Dworkin and 47, 49, 105, 303, 308, 319
open-texture of 311 Exclusive Legal Positivism xvi, 22
philosophy of 96, 258, 272 first-order non-positivism xiii, xiv
shared conventions of 312 first-order positivism xiii, xiv
see also German language; pragmatics; founding fathers of (Hobbes and
semantics; thick concepts Bentham) 344
Lasswell, Harold 103, 115 Gardner’s account of 9, 19
law Hartian account of xviii, xx, 44–5, 95, 171,
as integrity 105, 111, 238 226, 302, 303
politics and 346 Inclusive Legal Positivism 22, 45
connection between 330–6 interpretivism and 218, 221–2, 226
legal validity 334–6 Kelsen’s account of 95
separation between, giving up of 328–9 legal disagreements and 308–12
see also idea of law; conceptual analysis of law legal obligation and 171
(CAL); nature of law; rule of law methodological non-positivism xiii, xiv, xv, xix
law-state methodological positivism xiv, xix,
law and state, relation between 290–5 234, 243–4, 251
legality, state and legal system 279–98 myths about 9
356 Index
legal positivism (cont.) sufficient conditions for the
natural law theory vs xii, 8, 19, 95, 266–7, existence of 186–90
286, 307, 343 synchronic/ diachronic dimensions of 314
non-positivism 94–7 theoretical disagreement 295–7
law as integrity, concept of 96 see also analytical jurisprudence; choice of law
law’s authority between 94–7 defenses; coercion; gaps in law; Italian
methodological xiii, xiv, xv, xix legal system; nature of law; see also under
political morality 92–3 individual countries
Raz’s authority-based argument xv–xvi, 44, legal theory
61, 82–3, 89, 94, 185 state-centered 281–5
reductionism and 209, 218 legal validity 155, 171, 226, 231, 251, 264,
Shapiro’s account of 43, 185 304, 313, 336
social facts and 258 see also thick concepts
soft (inclusive) and hard (exclusive) xiv, xv, legislation
xvi, 19, 95 coercive enforcement of 184
see also normative positivism creation and modification of 115
legal pragmatism 105 criticisms of 184
legal process school 103 formal features of 135–6
legal rationality inconsistent 145
ethical rationality 116 institutional use of 282
forces of law 113 judicial review of 330
legal-formal/systematic rationality 115 manipulative 136
levels/spheres of 115–23 role of 110, 135
linguistic rationality 115 United States 327
methodological approaches 114 see also power; rule of law
nature of law and 101–23 legislature, the 78
pragmatic rationality 116 legitimacy 76, 96
teleological rationality 115 moral 96, 218
see also moral importance; nature of law see also state legitimacy
legal realism 105 legitimate domination 76, 78
legal reasoning 109, 112, 123, 329 legitimation 92
see also reason see also domination
legal sources, see legal disagreements Legrand, P. 340
legal stability 93 Leiter, Brian xix, xx, 219, 231, 246, 250, 280,
legal statements 304, 308–9, 318
common observations about 257–9 lex mercatoria 292
see also thick concepts libertarianism 205
legal system 185, 230, 305–6 liberty 95
circularity 297 linguistics, see language; semantics; socio-
constitutions 282 linguistics; pragmatics; thick concepts
continuity of legal systems, problem of 285–9, Lippke, R. 204
297–8 Little, Maggie 4
existence of 311 Llewellyn, Karl N. 103, 109
federal systems 282, 292 location of evaluation 261
functioning of 313 Locke, John 183
identity of 293 logic 108–10, 112–13, 123, 137–8
immoral 271 deontic 137
indeterminacy 297 Lyons, David 156
international legality 283
law and state, relation between 290–5 MacCormick, Neil 28, 288
legal pluralism 281–5 Madison, James 322
legality, the state and 279–98 manipulation 131–3
non-state legality 283 of reasons 55
old jurisprudential questions, renewal of see also power
297–8 Markesinis, B. S. 340
penumbral examples of 72 Marmor, Andrei xv, xviii, 82, 85, 130, 141,
social practice and 185 155, 215, 221, 227, 341
specific/non-specific 276 marriage 105, 233, 242–3
state-centered legal theory 281–5 Marx, Karl 130
Index 357
Marxism 217, 226 natural language, see language
mathematics 107, 236 natural law 95, 296, 302
matrix 90 content of 104
see also de facto authority theories of 95, 105, 123, 286
maxims, see under individual legal maxims theorists 344
McCracken, Chad 101 see also weak natural law thesis
McDougal, Myres 103, 115 nature of law 241
McKeever, S. and Ridge, M. 260 American response to 327–30
Mercier, Adèle 235–6, 242 economic analysis of law 329
metaethics 216 law/politics divide, giving up of 328–9
analytic 343 theories of interpretation 329
metaphysics 343 analogy 104, 110, 113
metaphysical fiction 56 answers to 322–46
of social rules 32–3 background to 322–3
metatheoretical values 246–8 British response to 325–7
methodological positivism 243–4, 251 doctrine in legal education 327
Mexico 293 doctrine in legal scholarship 327
Mexican Constitution 117 government action, judicial deference to 326
Micronesia 292 judicial appointments, non-political process
Mikhail, J. 344 of 326–7
mischief rule 118 parliamentary supremacy 326
Montevideo Convention (1933) 292–3 concept of law 104, 106
Moore, Michael 15, 330–1 criterial concepts 105
moral error 194–6 definition of 337
see also coercion dual:
moral importance 102, 104, 106–7, 112, legal disagreements and 301–19
115–16, 118, 123 Dworkin and 323–4
see also legal rationality evaluation of 346
moral obligations 155, 157, 163, 186, fixity and flexibility 114
201, 203 Hart and 323–4
moral philosophy 79, 94, 96, 258 human nature and 344–6
moral pressure 168 implications for the search for 336–41
moral reasons, see reason interpretive concepts 105
moral repugnance 118, 121 jurisprudential fray 330–6
moral theory 79–80, 85, 97, 223, 225, 231 legal rationality and 101–23
see also political morality natural kind concept 103, 105
morality 45, 88, 94–5, 97, 110, 115, 123, 162, objections to 341–3
172–3, 238, 334–5 rule of law, dual approach to 324–30
dimensions of 41–3 single/ unique 336
law, relations with 89, 94, 346 sufficient conditions for law 103
moral convictions 91–4, 97 see also law and politics; legal rationality
moral duty 43 Nazi Germany 280
moral legitimacy 96 Nguyen, Thi 195
moral/non-moral dichotomy 41–4 nihilism 18, 79
morals 112–13 non-state legality 283
objective 97 normal case, the 90
true 97 normal justification thesis (NJT) 53–4, 60, 63,
universal 96–7 67, 69, 73, 81, 85
see also moral theory; political morality see also justification defense
municipal law 158, 168, 279, 295 normative fallacy 79–80
Murphy, Liam 231, 233, 237–8, 245, 249 of Raz 80–6
Murphy, Mark xi, xii, 4, 6 normative positivism 95, 97
Muslims 171 see also legal positivism
see also religion Normativity Thesis 154–5
North Korea 211
Nagel, T. 201 Nozick, Robert 184
natural kinds 103, 105 number theory 224
biological kinds 235 Nuremburg Trials 280
358 Index
Oakeshott, M. 325 political principles 332
obedience 175–6 see also democracy
Oberdiek, J. and Patterson, D. 231–2 positive law 244
obligation positivism, see legal positivism
binding force of 164–5, 180–2 Possibility Puzzle 33–4, 37
concept of 161–2 possible ‘legal‘ worlds 314
exclusionary, as 163–4 Postema, G. 292
mandatory prescriptions and 162 Pound, Dean 112
normative force of 164–5 Pound, Roscoe 103, 112, 114–15
reasons, as 162–3 power
sanction theory of 152–3 brute social power 129
see also Hartian Theory of Legal Obligation; children, over 148
Hart’s theory of social obligation; moral coercive 185
obligations; reason concept of 131
originalism 316, 329 law as 129–49
‘ought’ implies ‘can’ principle 141 manipulative 132–4, 136–7, 140,
ought-claims 159 142, 149
non-standard guidance 134, 143, 145–7
Paget, J. 61 pastoral 144, 148
Parfit, D. 15 political 132, 336
parliamentary sovereignty 326, 335 prescriptions 134–6, 138, 141–2, 144,
see also United Kingdom 146, 148
paternalism 149 inconsistency of 139–40
Patterson, D. 214 intentional structure of 138–9
Peacocke, Christopher 236 rule of law power 131–3
Peretti, T. 328 social power:
Permanent Court of International Justice ideological hegemony 130
(PCIJ) 292–3 rule of law, through lens of 130–1
Perry, Stephen 231, 233–4, 239, 243–9 social oppression 130
Petrasyski, Leon 118 standard guidance 149
phenomenology of law 93, 268 subjectivation 144
philosophical analysis threats 148
concept of law and 230–51 varieties of 131
manifest concepts 232 see also coercion; rule of law
operative concepts 232 practical matrix 89–91, 93–4
see also ameliorative inquiry; conceptual pragmatics 57, 261–3
analysis of law (CAL) see also language; semantics
philosophical anarchism 186 predictability 93
philosophy, discipline of 112 preemption thesis 52–60, 64, 67, 73,
see also art, philosophy of; language, 81, 84–5
philosophy of; legal philosophy; moral preemptive reasons 55
philosophy; political philosophy; science, prescriptions
philosophy of general 162
physics 216 mandatory 162, 164
Pinker, S. 344 particular 162
Planning Theory of Law 22, 45–6, 50 see also power: prescriptions
Plato 84–5 prescriptive statements, see power
Platonic realism 211 presupposition 246, 261–2, 269–70, 272–3
pluralism, see conceptual analysis of law (CAL): Priel, Dan xxii, 280, 332, 334, 342–5
conceptual analysis pluralism; legal pro tanto conditions xviii, 51, 62–3, 133, 146,
pluralism 189, 192, 198
political activism 329 property law 188–9, 192, 196, 200–1,
political discourse 184, 205, 332–3 203, 205
political morality, theories of 78–9, 84, 91–3, prudential reason, see reason
95–7 psychology 202
see also moral theory; morality public, the 78
political philosophy 205, 219–20, 339 publicity 130, 146, 149
political theory 231, 327, 346 Putin, Vladimir 340
politics 332–4, 336 Putnam, Hilary 235–6, 240, 245
Index 359
Quine, W. V. O. 237, 305 rule of law 221
Quinean philosophy 231 compliability (technology of guilt) 129–30,
see also Duhem-Quine thesis 134, 141–4
non-compliability 146–7, 149
Radbruch, Gustav 118 pervasive guilt 143
rationality in lawmaking conflict vs dissonance 139
ethical xiv consistency 129–30, 134, 137–41
instrumental 141 definition of 130
non-compliable directives 142–3 epistemic dimension of 143
strict liability 142 human behavior, guiding and
see also legal rationality subjecting 133–4
Ratti, G. B. 304, 313 human dignity, respect for 132, 145, 149
Rawls, John 26, 200, 308 legislative conception 135–6, 138
Raz, Joseph xv, xvi, 10, 18–19, 27–8, 34, objectivity in law 324–5, 333–4
38–41, 43, 51–9, 61–2, 64–8, 70, 72–3, permissive norms 137
77, 89–90, 94–5, 132, 136, 140, 147, power 131–3, see also power
154, 159, 161, 183, 185–6, 217, 219, central features of 145–9
222, 225, 239, 247, 258, 267, 269, instrumentally rational, as 146
271–3, 279, 281–3, 286–7, 289–91, non-paternalistic, as 147–8
293–4, 340 non-public, as 145–6
descriptive-social theory 86 requirements, institutional and
normative fallacy of 80–6 procedural 130, 132
rightful authority, theory of 80–7 rule of men vs 132
service conception of authority social power through 130–1
xvi, 53–9, 69 subjectivation 144–5
explanation of 63 two requirements of 129–49
see also de facto authority, theory of see also power
reason(s) 94, 112, 159 Rule of Recognition
legal 179, 181 American 30
moral 162–3, 179, 181–2, 189, disagreement over 44–9, 341
193–6, 226, 238 criteria of legal validity 45
motivational 160 Hartian response to Dworkin 46–9
normative 160 hyperbole 45
ought-claims, relationship with 159 Shapiro’s response to Dworkin 46
prudential 179, 181 see also under Hart, H. L. A.
social norm and 166 Russell, B. 317
see also obligation Russia 340
reason-holism 260 Ryle, Gilbert 210, 213
reductionism 209
anti-reductionism 225–9 Sachs, Albert 103
constructive interpretations 226 Sager, Lawrence 183
limits of 225–9 sanction 103–4, 172, 192, 199–200, 202
rationalization constraint 226 Santos, Boaventura de Sousa 292
reduction, possibility of 214–17 Schauer, Frederick 102–3, 232, 237–8, 251,
see also interpretivism 327
relativism 345 Schroeder, M. 260
religion 217, 226, 239, 242, 250, 332 Schwyzer, H. 224
see also Catholic religious practices; Muslims Sciaraffa, Stefan 3, 70, 273
representation 78 science, philosophy of 96
Riggio, Adam 101 Scott, John 144
rightful self-defense 192
authority 75–6, 96–7 self-government 93, 95, 328, 333–4
de facto authority and 76–89 self-interest 202
condition 197 semantic externalism 86–9
Ripstein, Arthur 183, 189, 196–7 see also externalism
Rodriguez-Blanco, V. 240, 245, 248 semantics 40, 57, 213, 227, 240, 261–3, 270,
Roman law 305, 313 276
Rosen, Arie xiv, xv, xvi criterial semantics 245
Ross, Alf 130, 335 semantic reduction 216
360 Index
semantics (cont.) disagreement on 314–17
semantic sting, the 105, 212–3, 245 identifying and interpreting 305–6
see also language; pragmatics interpretation of 314–17
separabilism 263, 265, 273 sources thesis 19
Separability Thesis 180 sovereign, the 78
service conception 53–9 sovereignty 287, 293
advantages of 60 see also parliamentary sovereignty
choice-of-evils and 67 Spinoza, B. 91
explanation of 63 Stainton, R. 242–3
legal authority under 69, 73 Stalnaker, R. 269–70
soundness of the 87 Stammler, Rudolf 104
Shapiro, Scott xvi, 8–10, 17–18, 22–3, 25–36, statistics 111–12
38–44, 46, 50, 55–6, 180, 183, 185, Stavropoulos, N. 218, 220, 234–6, 239–40, 244
218, 303, 308, 334 Stevens, R. 326, 341
combing one’s hair, example of 36 Stich, S. P. 343
Dworkin, response to 46 Stoljar, Natalie xviii, xix, 101
greetings between acquaintances, Strawson, P. F. 148–9
example of 37 supervenience 288–9
locking car doors, example of 36 syllogism 336
saving for retirement, example of 36
smoking, example of 35–6 Tamanaha, Brian 172, 279–80, 284, 292, 296
Sidgwick, H. 326 tardy 241
Simmons, A. J. 186 Tarello, Giovanni 309
slavery 196–7 taxation 197, 200–1
Smith, Munroe 113 textualism 329
smoking 35–6 theoretical disagreements xx
Soames, S. 57, 211 Rule of Recognition 44–9
social life 279 see also legal disagreements
social norms thick concepts 259–64
coercive enforcement of 177 entangled thick concept 263, 265, 266–7,
conformity to 170–1 273
as exclusionary 166 evaluation of 276–7
individualized reasons and 33–7 jurisprudential debate 266–76
absence of social norms 36 jurisprudential payoffs 266–7, 274–6
smoking 35–6 jurisprudential themes 267–74
social practices 34 ‘legal’ as a thick concept 264–6
legal norms and 288–91 see also legal statements
lake swimming 288–9 Thompson, M. 4
state/non-state norms 292, 295 Timmons, M. 343
systems of extralegal social norms Tocqueville, A. de 327
292–7 Toh, K. 266, 268–9
political and 290 tort liability 326
reasons and 166 trascendentalism 101
state-constituting xxi truisms 199, 203, 217
social obligations and 152–3, 165–9 Twining, William 279–80, 284, 296, 302
see also Hart, H. L. A.; Rule of Recognition; Tyler, Tom 203
Shapiro, Scott
social obligation ubi eadem ratio iuris, ibi eadem iuris
social pressure and 167, 169 dispositio 120
see also Hart’s theory of social obligation ubi lex non distinguit, non distinguere
social policy 334 debemus 122
social welfare 113 Unger, Roberto M. 103
socio-linguistics 87 United Kingdom 286
see also semantic externalism attitudes to law 341
sociology 112–3, 115, 202, 231, 245 Bill of Rights for Britain 324
sociological concept of law 185 British constitution 335, 341
Socrates 84–5, 231, 242 British courts 326
sources of law British exceptionalism 341
agreement on 313–14 British law 323–4
Index 361
Civil Service 326 Velleman, David 65
freedom in 337 Von Wright, G. H. 137–8, 141
fundamental law, decline of 335
House of Commons 335 Waddams, S. 340
House of Lords 333 Waldron, J. 92, 95, 132, 188
law and politics in 322, 324–7, 335, Walton, K. 270
337–8 Waluchow, Wilfrid J. 101, 118
doctrine in legal education 327 Warnock, Geoffrey 34, 227
doctrine in legal scholarship 327 weak natural law thesis
judicial appointments 326–7 anti-positivist implications of 16–20
judicial deference to government bad objections to 6–9
action 326 claims of 3–6
parliamentary supremacy 326 constitutional ability 13–14, 19–20
lawyers in 330 contribution of 9–11
political culture in 324, 331, 336 criticisms of 3, 9–11
wednesbury unreasonableness 326 defectiveness, notion of 4–8
see also England; parliamentary sovereignty explanatory role of 3–20
United States 35, 293, 326 good, notion of 5
American courts 327 hypothetical necessity and 11–20
American exceptionalism 341 moral natural law thesis 6
attitudes to law 341 natural law by determination 16
Constitution (1787) 46, 113–14, 117, 327, ought of kind-membership 4–5, 7
329–30, 333–5 practical rationality 15–16
Declaration of Independence (1776) 333 rational standards 5–6, 8, 11, 18–19
economic analysis of law in 329 role of 20
freedom in 337 sophistry 8
Hart’s views of 323 strong natural law theses 5–6, 9, 15,
judicial appointments in 326–7 17, 19
law and politics in 322, 327–9, 335, see also hypothetical necessity; natural law
337–8 Weber, Max 76, 130
lawyers in 330 Wechsler, H. 333
legal activism in 329 Weimar Republic 104
litigation in 328 Wertheimer, A. 204
municipal legal system of 160 Wilde, Oscar 260–1, 274–5
political culture in 322, 324, 331, 336 Williams, Bernard 259–60, 262, 301
Rule of Recognition in 45 Williamson, T. 263
smoking in 35–6 wills, submission to 78
state legislation 225 Wittgenstein, Ludwig 90, 210, 213
Supreme Court 183, 328–9 Wolff, Robert 175
Universal Declaration of Human Wood, G. S. 335
Rights 280 Wright, Orville 12–13
universalism 345
Yablo, S. 270
Van Fraassen, B. 301, 343
Väyrynen, Pekka 260–1, 272 Zimmermann, R. 340

Philosophical Foundations of The Nature of Law

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Philosophical Foundations of The Nature of Law

Caricato da

Copyright:

Formati disponibili

PHILOSOPHICAL FOUNDATIONS

List of Contributors vii

PART I. FURTHERING DEBATE BETWEEN

PART III. CONCEPTUAL ANALYSIS

In May 2011, the McMaster University Program in Legal Philosophy sponsored a

2 Cf. Leiter (2007: 171–3).

First-Order Positivism First-Order Non-

may be important or valuable because they are morally important or valuable.

derivative determinants. A soft positivist allows that moral principles may be

8 See e.g. Kramer (2004) and Waluchow (1994).

12 See e.g. Coleman (2002) and Leiter (2007: particularly, 175–99).

17 Cf. Sciaraffa (forthcoming).

Coleman, J. (2002). “Methodology,” in J. Coleman and S. Shapiro (eds.), The Oxford

My topic is a thesis of analytical jurisprudence that I have labeled elsewhere the

1. What the weak natural law thesis claims

5 See Murphy (2006: 8–20).

2. Some bad objections to the weak natural law thesis

6 I offer such defenses in Murphy (2006: 25–60), and in Murphy (2011).

8 Leiter (2011: 677). 9 Dickson (2001: 77).

10 Shapiro (2011: 408–9).

3. What a theory of law should accomplish

11 See Gardner (2001). 12 I discuss this point in more detail in section 5.

13 Raz (2004: 324).

That is sufﬁcient preliminary discussion of what a theory of law should be up to.

4. Hypothetical necessity and the weak natural law thesis

inventors whose designs were total non-starters never built an airplane.

5. Anti-positivist implications of the weak natural

31 Aquinas, Summa Theologiae IaIIae 95, 2.

46 Raz (1994: 202).

Positivists, see Kramer (2004: chs. 1–4).

1.1. Hart’s own words

4 Hart (1994: 95). 5 Hart (1994: 103).

1.2. Hart’s undeceivedness

6 Shapiro (2009: 246–7), footnote omitted.

2. The overarching unity of the Rule of Recognition

2.1. A puzzling footnote

9Shapiro (2009: 246), emphases in original, footnotes omitted.

2.2. Differing norms and differing addressees

13 Kramer (2004: 108–10).

3. The Rule of Recognition as a complex hybrid

3.1. Two preliminary points

3.2. A complex array

15 Shapiro (2009: 239–40), footnotes and paragraph breaks omitted.

3.3. No collapse into rules of change

4. Rules and regularities

Quite puzzlingly, Shapiro in Legality repeatedly accuses Hart of the category

20 (2011: 80), emphasis in original.

5. Individualized reasons and social norms

5.1. The wrong conditional

5.2. Norm-guided behavior versus habitual behavior

24 Shapiro (2011: 40). 25 Shapiro (2011: 103–4).

A lengthy quotation is warranted here:

5.2.1. The example of smoking

26 Shapiro (2011: 104).

5.2.2. Shapiro’s other examples

The one remaining example in Shapiro’s list is that of greetings between

6. Simulative statements versus internal statements

29 Shapiro (2011: 111–12). 30 Shapiro (2011: 112).

32 Shapiro (2011: 415 n. 44).

7. The nature of legal concepts

7.1. A conﬂation of two dimensions of morality

35 Shapiro (2011: 111).

38 (2011:114), footnote omitted.