The Ethics of Proportionate Punishment

THE ETHICS OF PROPORTIONATE PUNISHMENT
LIBRARY OF ETHICS AND APPLIED PHILOSOPHY

VOLUME 16
Managing Editor:
Govert A. den Hartogh, University of Amsterdam, The Netherlands
The titles published in this series are listed at the end of this volume.
THE ETHICS OF
PROPORTIONATE
PUNISHMENT
A Critical Investigation
by
JESPER RYBERG
Dept. of Philosophy and Science Studies,
University of Roskilde, Denmark
KLUWER ACADEMIC PUBLISHERS

DORDRECHT / BOSTON / LONDON
A C.I.P. Catalogue record for this book is available from the Library of Congress.
ISBN 1-4020-2553-X (HB)

ISBN 1-4020-2554-8 (e-book)
Published by Kluwer Academic Publishers,

P.O. Box 17, 3300 AA Dordrecht, The Netherlands.
Sold and distributed in North, Central and South America

by Kluwer Academic Publishers,
101 Philip Drive, Norwell, MA 02061, U.S.A.
In all other countries, sold and distributed

by Kluwer Academic Publishers,
P.O. Box 322, 3300 AH Dordrecht, The Netherlands.
Printed on acid-free paper
All Rights Reserved

© 2004 Kluwer Academic Publishers
No part of this work may be reproduced, stored in a retrieval system, or transmitted
in any form or by any means, electronic, mechanical, photocopying, microfilming, recording
or otherwise, without written permission from the Publisher, with the exception
of any material supplied specifically for the purpose of being entered
and executed on a computer system, for exclusive use by the purchaser of the work.
Printed in the Netherlands.

For Charlotte
CONTENTS
Introduction 1
1. Why consider proportionalism? 2
2. A brief overview 6
Notes 10
Chapter 1 Proportionalism and its Justifications 11

1. What is proportionalism? 12
2. The simple desert theory 14
3. The expressionist theory 19
4. The fairness Theory 36
5. A non-foundationalist approach 44
6. Conclusion 53
Notes
Chapter 2 The Seriousness of Crimes 59

1. The harm dimension 60
2. Culpability 68
3. Recidivism 77
4. Proportionalist answers 83
5. A fairness-theoretic approach 87
6. Conclusion 93
Notes 95
Chapter 3 The Severity of Punishments 101

1. The sensibility challenge 102
2. Delimitating punitive suffering 109
3. Resorting to mercy 116
4. Conclusion 118
Notes 120
Chapter 4 The Anchor Problem 123

1. Ratio, interval, and ordinal matchings 125
2. Anchor points and human dignity 131
3. Desert, prevention, and parsimony 142
4. Conclusion 148
Notes 150
viii Contents
Chapter 5 Proportionalism and Penal Practice 155

1. The challenge of self-defeatingness 156
2. Justice in an unjust society 166
3. Conclusion 178
Notes 180
Chapter 6 Relaxed Proportionality 183

1. Problems and promises 183
2. Modified proportionalism 189
3. Conclusion 195
Notes 197
Bibliography 201
Index 217
INTRODUCTION
The philosophical discussion of state punishment is well on in years. In contrast

with a large number of ethical problems which are concerned with right and wrong
in relation to a narrowly specified area of human life and practice and which have -
at least since the early 70’s - been regarded as a legitimate part of philosophical
thinking constituting the area of applied ethics, reflections on punishment can be
traced much further back in the history of western philosophy. This is not
surprising. That the stately mandated infliction of death, suffering, or deprivation on
citizens should be met with hesitation - from which ethical reflections may depart -
seems obvious. Such a practice certainly calls for some persuasive justification. It is
therefore natural that reflective minds have for a long time devoted attention to
punishment and that the question of how a penal system can be justified has
constituted the central question in philosophical discussion.
Though it would certainly be an exaggeration to claim that the justification
question is the only aspect of punishment with which philosophers have been
concerned, there has in most periods been a clear tendency to regard this as the
cardinal issue. Comparatively much less attention has been devoted to the more
precise questions of how, and how much, criminals should be punished for their
respective wrong-doings. This may, of course, be due to several reasons. The
traditional controversy between the utilitarian and the retributivist approaches to the
justification question may have made it less obvious to proceed into some of the
more detailed questions. Relics of the view that the question of the punishment
method and amount is a matter of pure positive law, which cannot be determined by
abstract ethical reasoning, or the contention that once the more basic justification of
a punishment system has been provided the more detailed questions would thereby
also be answered or at least be out of the hands of philosophers, may all be parts of
an explanation, the more precise content of which is not the present concern.
However, focus on the general question of justification and a more marginal
engagement in the problems of penal distribution has had the implication that
philosophical discussion has often appeared as purely academic manoeuvre far
removed from the realities of actual penal practice. Deep thoughts contributing no
practical guidance.
It is this traditional way of depicting philosophical work which has, in the
case of punishment, undergone a significant change over the last two or three
decades. The revival and development of the proportionality principle - with which
this book is concerned - marks a change in the focus towards an approach which
seeks to contribute directly to the construction and outfit of the punishment system.
Proportionalism does, as has been pointed out by one of its chief exponents, help in
closing the gap between philosophers who have concentrated on why the institutions
of punishment should exist at all, and penologists who have assumed punishment’s
1
2 THE ETHICS OF PROPORTIONATE PUNISHMENT
1
existence and have been concerned with the distribution of sanctions. That the
principle serves this purpose and that it has thus contributed to a dissolution of a
strict division of academic labour is, I believe, a noteworthy merit. But obviously
this is not the only thing which at first glance might direct attention to a closer
discussion of proportionalism.
In the following, a few preliminary words will be said about what has
motivated the present work and what I believe justifies devoting a whole volume to
an ethical scrutiny of this position. Furthermore, I shall present the reader with a
brief overview of the content of the ensuing chapters and of what will constitute the
main argument of this book; that is, it will be indicated why - though it is certainly
correct that there has been a gap between philosophers and penologists - I do not
believe that proportionalism is in the end the position that should bring philosophy
to the front-line of penal practice.
1. WHY CONSIDER PROPORTIONALISM?
The question of how punishments should be distributed in a society is obviously of

ethical importance. Even though one holds, as do most theorists, that a punishment
system can be justified, this is surely not tantamount to giving carte blanche to
punishing crimes in any possible way. Whether a perpetrator should be executed,
imprisoned or fined for a misdeed is in itself a substantive ethical question.
However, besides the fact that this is the kind of question with which the
proportionality principle is concerned, and which in itself makes it worthy of
attention, there are a number of reasons for focusing particularly on this view as the
candidate for an answer to the distribution question.
First of all, it is indisputable that proportionalism is at first glance
intuitively appealing. A reasonable interpretation of the frequently met sort of
statements complaining that a particular punishment is too harsh or too lenient for
the crime, or that the punishment for one crime is absurdly harsh or lenient relative
to the punishment for another, is to perceive them as expressing a devotion to
proportionality. As has also been pointed out, the approach to justice which the
proportionality view represents, can be found even in the way children object to
disparities in 2 the blame or punishment imposed on them for acts of similar
misbehaviour. In fact, theories which seek to explain the origin of the immediate
appeal of proportionalist judgements have been suggested. For instance, in one of
his last works Mackie points out the paradoxical character of retribution which in his
view consists in the fact that, on the one hand, a retributive principle of punishment
cannot be explained within a reasonable system of moral thought and, on3 the other
hand, that such a principle cannot be eliminated from our moral thinking. Mackie’s
answer to the paradox is to adopt a Humean approach according to which moral
distinctions are founded on sentiment, not on reason, and he supplies his position by
offering a biologically based explanation of such emotions. Retributive behaviour,
he suggests, can be seen as something which tends to benefit a retaliator by
discouraging an aggressor from repeating an attack. In creatures which possess a
sufficient capacity for emotion, retributive behaviour will naturally be accompanied
INTRODUCTION 3
by the development of retributive emotions. Whether it is correct that retributivist

behaviour and emotions can in this way be traced back to mechanisms of natural
selection is, of course, a controversial question and not one we need to be bothered
with in the present context. Neither is it necessary to consider other conjectures
concerning the genesis of such emotions. But what in the first place initiates the
development of such theories is the fact that we actually possess the kind of
emotions which are captured in a retributivist position and, more narrowly, in a
proportionalist approach to punishment. Now, the existence of such emotions does
not, of course, in itself show that proportionalism is an ethically valid position. But
it certainly provides a reasonable starting point and motivation for considering
whether the view can stand a closer scrutiny.
The second thing which motivates an ethical investigation of a
proportionalist distribution of punishment is that the view has in several countries
formed the ground on which convicts have been punished. The story of how
proportionalist ideals gained a foothold in modern penal practice has often been
4
told. Penological thinking in the 50’s and 60’s was predominantly consequentialist.
The criminal sanction was believed to control crime by its deterrent, rehabilitative,
and incapacitative effects. Retributivist concerns were to a large extent eschewed
and regarded as a reactionary approach to punishment. However, the late 60’s and
the early 70’s marked a period of growing discontent with the existing penal order.
The dominating rehabilitative ideal was attacked both theoretically and empirically.
Perpetrators should no longer be regarded as sick and as individuals in need of
treatment. The analogy between patients and criminals was rejected (and, even if
some criminals actually were sick, the criminal-justice system was no longer
regarded as capable of administering the requisite cure). The individualized
approach to punishment fostered by the basic rehabilitative idea that punishment
should be tailored to the needs of the individual criminal was accused of leading to a
lack of control and to arbitrariness in decision-making. And disillusionment with the
impact of rehabilitation on rates of recidivism prompted a “nothing works”
atmosphere.
Likewise, the use of incapacitation became widely criticized. The
assumption on which the incapacitation idea was based, namely, that inmates would
have continued to commit crimes had they been free, was attacked in several ways.
It was no longer believed that crime would be prevented by removing some
criminals from society (and, even if this should to some extent be the case, the price
was considered too high). The general frustration over inequity, injustice and
arbitrariness in the application of the law - or as one of the main critics put it: the
5
“lawlessness in sentencing” - formed the platform for an antithetical revival of
retributive ideals, now presented under the title of “just deserts”. By basing the
punishment system on proportionalism the problems were apparently avoided. The
practice of individualized and indeterminate punishments which had been a crucial
part of the treatment-oriented system would be abandoned and one would be
allowed to put aside many of the empirical questions on which consequentialist
punishing was based, and to which there were no clear answers.
During the 70’s and 80’s the new ideals led to a reform of the penal codes
in several countries: the USA and Europe, as well as in other parts of the world. In
the USA the work of sentencing commissions led to the implementation of
proportionate punishment schemes. Among the most significant attempts to reflect
such concerns were those of Minnesota and Oregon. The proportionalist guidelines
were presented in the form of sentencing grids: two-dimensional scales of sanctions
with a vertical axis grading the seriousness of various sorts of criminal conduct and
a horizontal axis rating the extent of the offender’s prior criminal record. In relation
to the Canadian sentencing system, the sentencing commission noted that “... the
paramount principle governing the determination of a sentence is that the sentence
be proportionate to the gravity
6
of the offence and the degree of responsibility of the
offender for the offence” . And Australian High Court decisions 7 pronounced
proportionality as the primary objective of sentencing in Australia. In Europe,
Finland amended its penal code to adopt a policy of proportionate sanctions as early
as 1976. It was specifically emphasized that a punishment should be measured for it
to be in “just proportion” to the damage caused and the guilt of the offender. And,
little more than a decade later, similar reforms took place in Sweden. In England and
Wales changes came about with the Criminal Justice Act of 1991. The 1990 White
Paper preceding the 1991 Act presented itself as offering “a coherent legislative
framework for sentencing with the severity of the punishment matching the
seriousness of the crime”, and pointed at desert as being the primary aim of
8
sentencing. Though specific guidance for the sentencers in the European countries
which underwent changes was not given by the kind of numerical guidelines which
were adopted in the USA but rather through statutory guiding principles, the
underlying rationale was still one of proportionality.
As this small sketch of a part of modern legal history indicates, the
abandonment of consequentialist ideas and the revival of retributive ideas in penal
9
practice is one of the most striking changes to have occurred over the last decades.
The mere fact that such reforms have taken place does not, of course, per se show
anything about the plausibility of the involved rival moral principles. However, the
fact that sentencing systems have been constructed in ways that attempt to reflect
the principle of proportionality, that is, that the principle is not merely a
philosophical abstractum but also a view on the ground of which persons convicted
of crimes have actually been punished, does certainly also make it reasonable to
consider whether the principle is one that we should in the end applaud and be
morally satisfied with.
The third and, indeed, the main motivating reason for engaging in such an
investigation does not concern the sketched reforms in penal practice but the
changes which took place in the academical thoughts on punishment. These changes
were certainly just as remarkable as the practical upheavals. In the period before and
during the middle of the 20th century there were very few who believed in
retributivism, and even fewer who openly defended it as the most plausible approach
to punishment. This is evidenced by the manner in which those theorists who felt
that there were substantial points to be extracted from the Kantian and Hegelian
thoughts on the matter exposed their viewpoints. In 1939 Mabbott opened his
defence of retributivism by claiming that he felt sure his enterprise would arouse
INTRODUCTION 5
deep suspicion and hostility both among those involved in penal practice and among
philosophers who regarded the retributive view as “the only moral theory except
perhaps
10
psychological hedonism which has been definitely destroyed by criticism
..” . Retributivism was regarded only as a polite name for revenge. A barbarous or
inhumane position far distant from what could possibly be regarded as a reflective
or enlightened approach to the issue. In that light it is not surprising that Mabbott in
a later comment on contemporary British philosophy noted that “retribution 11has
been defended by no philosopher of note [for over fifty years] except Bradley ..” .
During the 60’s a number of philosophers declared their approval of
retributivism. However, the dominance of consequentialist thinking was clearly
witnessed by the fact that the main focus for the retributivists was on pointing out
unacceptable implications of consequentialism rather than on elaborating on the
content of their own position. At this point things changed significantly in the 70’s.
References to the renaissance or revival of retributivism became part of the standard
refrain in titles and opening lines of works on punishment. And, in the 80’s, Gross
could without hesitation proclaim that “liberal opinion no longer need to be
12
ashamed to associate itself with concern about just deserts” . Today it would
certainly be a bad euphemism to talk of an incipient interest in retributivism. Rather
is it correct to claim with Davis that “.. today, the theory of punishment is largely
13
retributive theory” . However, this fact does not mean, as one might perhaps
believe, that theorizing about punishment is more or less over. On the contrary,
there is today much discussion for instance between retributivists and theorists who
only partly defend retributivist thoughts and, especially, between exponents of
different versions of retributivist theories.
The point that makes the revival and present dominance of retributivism
interesting in relation to this book is, obviously, that the proportionality view is
intimately related to retributivist thinking. Sometimes proportionalism is even
presented as a necessary condition for the classification of a theory as retributivist.
As will later be argued, I do not believe that such a classification is sound.
However, it is an indisputable fact that the proportionality view is always defended
as the retributivist answer with regard to distribution of punishment. And this is so,
even though these theories are in other respects very different. Even the theories
which are not genuinely retributivist in shape but which are more properly classified
as hybrids between retributivist and consequentialist concerns often incorporate
proportionalist considerations. The wide acceptance of the principle, combined with
the fact that relatively few have taken on the tasks of clarifying what it precisely
implies and of assessing the principle morally - since this is a book in ethics - is the
main motivation for engaging in an evaluation of the view.
In sum, the fact that the proportionality principle has some intuitive appeal,
that it has been applied as a basic principle in penal practice, and finally that it is
proclaimed to be morally sustainable is, I believe, what makes it reasonable for it to
be subjected to a thorough investigation.
2. A BRIEF OVERVIEW
Before embarking on the scrutiny, it is reasonable and hopefully helpful to say a few
words about the content of the following chapters and, more precisely, about what
constitutes the main arguments to be advanced. The discussion will proceed along
the following lines.
Chapter 1 begins by providing a more precise account of the content of the
proportionality principle. The principle is defined in a way which, I believe, is
uncontroversial and which manages to embrace some of the more detailed
disagreements which exist between varying retributivist interpretations of the view.
The main theme, which is then taken up, is how this principle can possibly be
morally sustained. The question is complicated by the fact that “retributivism” is a
label which covers several very different theories. The first theory which is
considered is, what I call the simple desert theory, according to which wrongdoers
deserve to suffer. Subsequently, the two most influential theories are considered:
firstly, expressionism according to which punishment can be seen as a
communicative process in which a perpetrator is, through the conveyance of an
appropriate condemnatory message, held accountable for his misdeed; and,
secondly, the fairness theory which perceives punishment as a way of restoring a
fair balance of benefits and burdens between the criminal and law-abiding members
of the society. Finally, comments are added on the possibility of providing a non-
foundationalist justification of proportionalism. The contention of the chapter is that
in the end neither of the different approaches succeeds in justifying proportionality.
Chapter 2 concerns the question which any proportionalist will have to
face, namely, what should determine the seriousness of a crime? If punishment
should be meted out in a way that is warranted by the seriousness of the crime that
has been committed, then obviously one needs an answer as to what makes one
crime more serious than another. The traditional reply consists in a dual-
dimensional account: seriousness is determined by harm and culpability. In a
background of what constitutes the most elaborate theory for gauging criminal
harm, some of the problems relating to the harm dimension of crime gravity are
discussed. And a longer passage is subsequently devoted to considering mens rea
and responsibility, both of which determine a perpetrator’s culpability. Several
proportionalists also believe that respect to a prior criminal record should be
payed in the final assessment of how severely a criminal should be punished.
Some of the arguments in favour of this view, along with some of the theoretical
problems which are led to by recidivism, are considered. It is argued, that the
different determinants of crime seriousness are confronted with what I call a
challenge of relative comparison and a challenge of absolute comparison. The
chapter ends with a discussion of a particular fairness theoretic account of crime
gravity which proclaims it is able to get around the outlined challenges.
Chapter 3 takes up an issue which is clearly of equal importance in a
discussion of the proportionality principle, namely, what makes one punishment
more severe than another? It is argued that a plausible account of severity, which is
immune to the Benthamite challenge that one and the same punishment may affect
those on whom it is imposed very differently in terms of what counts with regard to
INTRODUCTION 7
assessment of degree of severity, cannot be provided. Considerations on the possible

after- and side-effects which a punishment may have for the punished are also
presented. The discussion serves the purpose of providing a clearer idea of what
proportionality amounts to and challenges one of the suggested merits of the
principle, namely, its applicability in actual penal practice. A brief comment is
finally added on why the concept of mercy does not at this point provide a resort for
the proportionalist.
Chapter 4 concerns the question of how severely particular crimes should
be punished, or what is usually known as the “anchor problem”. Despite the fact
that proportionalism has sometimes been accused of leading to a toughening of
sentencing levels, the standard contention among adherents of the principle is that
the desert model certainly does not constitute a derivative of a “throw away the key”
approach to punishment. Proportionalism, it is typically underlined, is not a
draconian theory. However, too little attention has been paid to the theoretical
ground for the question of how different crimes should actually be punished. In
considering this question, I set out by outlining and evaluating different approaches
to what kind of matching there should be between a crime scale and a scale of
punishments. This leads into a more substantive discussion of how the two scales,
once constructed, should be anchored. The first conjecture to be evaluated is based
partly on considerations of the concept of human dignity. The second conjecture in
a subtle way seeks to combine considerations of desert, crime prevention, and
parsimony in punishing. It is argued that neither conjecture manages to provide
proportionalists with a theoretically well-founded guidance with regard to how
severely criminals should be punished.
Chapter 5 takes its point of departure in the uncontroversial assumption
that what we wish of a theory of punishment is not merely a theory which is
theoretically or formally sound for some possible world but is also a theory which
can guide us in the actual world. The question of the applicability of the
proportionality principle leads to a discussion of two problems. The first concerns
the practical consistence of applying proportionality as the governing principle of a
penal practice which is imperfect and fallible. The principle faces what I refer to as
“the challenge of self-defeatingness”. Several traditional deontological ways of
meeting this challenge are rejected, but in the end it is argued that the challenge
does not itself constitute a genuine problem for the proportionalist. However, it
generates a problem of priorities. The second problem that is brought forward
concerns the possibility of carrying out just punishments in an unjust society. It is
considered how different aspects of social justice affect the legitimacy of applying
the proportionality principle. It is concluded that the principle faces problems once
we take the vital step from penal theory to penal practice.
Chapter 6 begins by offering a summary of the main conclusions which
have been drawn in the foregoing chapters; it goes on to consider whether the
outlined problems which proportionalists are confronted with can be avoided by
adopting distributional principles which allow for deviances from strict
proportionality. A number of theorists have defended hybrid theories which in
different ways involve modifications of proportionality. Five versions of relaxed
proportionality are considered and it is argued that none of these conjectures

manage to avoid the basic problems which confront traditional proportionalism.
As this overview indicates, the view that is defended is that the
proportionality principle does not constitute a plausible candidate as to how
punishment should be distributed. The criticism which is presented can basically be
boiled down to the following three conclusions: firstly, that the principle lacks a
profound moral justification; secondly, that the principle is encumbered with a
number of theoretical problems which are not easily surmountable; and thirdly, that
the principle faces problems once we take the step from the ideal spheres of penal
theory to actual penal practice. Two comments of methodological shape should be
made in relation to the discussion of these controversial conclusions. What we are
considering in a discussion of the proportionality principle is certainly not - as has
in earlier periods often been assumed - a literalistic reading of the biblical demand
“Thou shalt give life for life, eye for eye, tooth for tooth, hand 14for hand, foot for
foot, burning for burning, wound for wound, stripe for stripe” . In the modern
retributivist epoch the content of the distribution view has been clarified and a
number of different answers to the questions which relate to the principle has
crystallized. In order to defend the three general conclusions it is, therefore, not
sufficient to consider one single approach to these problems. What this means is that
a large part of the ensuing discussion will consist in an outline and evaluation of
various different answers which recent proportionalists have provided. The days
when it was common to apply the same yardstick to all retributivists by rejecting
their outlooks on the ground of very general - and often caricaturing - counter-
arguments are certainly over.
Furthermore, and more importantly, it is aimed that the principle is
discussed in a way that is relevant for those who defend it. Retributivists have
sometimes complained that a part of the criticism which has been directed against
their viewpoints has been misplaced: it has consisted merely in emphasizing the
theory’s non-utilitarian character. Even though there are only a very few modern
deontological positions which are not at all sensitive to utilitarian or other forward-
looking considerations it is nevertheless clear that a criticism which consisted
merely in pointing out deviances from a utilitarian point of view would focus
precisely on what retributivists regard as part of the strength of their position and
15
would thus not contribute much, if anything, to the debate. The idea in this work
has been to assess what proportionalists themselves argue and therefore recognize as
being important parts of their theory and, as far as possible, to do so in such a way
that the outlined problems cannot merely be held to reflect differences in basic
methodological assumptions. For instance, this is the case in the evaluation of the
justification of proportionalism; this does not consist of a discussion of whether the
fairness theory or the expressionist theory of punishment is morally plausible - a
discussion which might lead into a methodological discussion of how ethical
theories can at all be validated - but merely in whether the theories, as held by their
exponents, actually succeed in justifying proportionalism. The accept-as-many-as-
possible-of-your-opponents-assumptions strategy has constituted the guiding idea.
This procedure, I hope, might help avoid a situation - which to often occurs in
INTRODUCTION 9
ethical debates - where conclusions that are reached through analyses seem pointless
from the criticized part’s point of view.
NOTES
1A. von Hirsch, “Proportionality in the Philosophy of Punishment: From ‘Why Punish?’ to “How
much?””, Israel Law Review, vol. 25, 1991, p. 580.
2See, for instance, A. von Hirsch, “Sentencing reform: its goals and prospects”, in A. Duff, S. Marshall,
R. E. Dobash and R. P. Dobash (eds.), Penal Theory and Practice, Manchester University Press, 1994, p.
28.
3J. L. Mackie, “Morality and the Retributive Emotions”, Criminal Justice Ethics, vol. 1, 1982.
4To mention a few references see, for instance, P. L. Griset, Determinate Sentencing, State University of
New York Press, United States of America, 1991; B. Hudson, Justice Through Punishment, Macmillan
Education, Hong Kong, 1987; A. von Hirsch, K. A. Knapp and M. Tonry, The Sentencing Commission
and Its Guidelines, Northeastern University Press, Boston, 1987; A. von Hirsch, Censure and Sanctions,
Clarendon Press, Oxford, 1993; M. Wasik and K. Pease, Sentencing Reform. Guidance or Guidelines?,
Manchester, 1987; M. Tonry and K. Hatlestad, Sentencing Reform in Overcrowded Times, Oxford
University Press, United States of America, 1997.
5M. E. Frankel, “Lawlessness in Sentencing”, Cincinnati Law Review, vol 41, 1972; reprinted in an
excerpted version in A. von Hirsch and A. Ashworth (eds.), Principled Sentencing, Hart Publishing,
Oxford, 1998.
6Canadian Sentencing Commission, Sentencing Reform: A Canadian Approach, Canadian Government
Publishing Centre, Ottawa, 1987, p. 154.
7See M. Bagaric, Punishment & Sentencing: A Rational Approach, Cavendish Publishing, Great Britain,
2001, p. 165.
8Quoted from I. Dunbar and A. Langdon, Tough Justice, Blackstone Press Limited, Great Britain, 1998,
p. 89. However, the 1991 Act was not interpreted as strictly in desert terms as desert theorists would have
wished; see A. Ashworth, “Four Techniques for Reducing Disparity”, in A. von Hirsch and A. Ashworth,
Principled Sentencing, Hart Publishing, Oxford, 1998, pp. 230-31.
9For a discussion of more recent developments see, for instance, C. Clarkson and R. Morgan (eds.), The
Politics of Sentencing Reform, Clarendon Press, Oxford, 1995.
10J. D. Mabbott, “Punishment”, reprinted in H. B. Acton (ed.), The Philosophy of Punishment, St
Martin’s Press, Great Britain, 1969, p. 39.
11Quoted by K. G. Armstrong in “The Retributivist Hits Back”, in H. B. Acton (ed.) ibid., p. 138.
12H. Gross, “Culpability and Desert”, in A. Duff and N. Simmonds (eds.), Philosophy and the Criminal
Law, Franz Steiner Verlag, Wiesbaden, 1984, p. 59.
13M. Davis, To Make the Punishment Fit the Crime, Westview Press, United States of America, 1992, p.
6.
14Exodus, XXI, 23-25.
15See, for instance, J. G. Murphy, “Three Mistakes about Retributivism”, Analysis, 1971.
CHAPTER 1
PROPORTIONALISM AND ITS JUSTIFICATIONS
Though the idea of proportionalism is susceptible to different interpretations and has

sometimes even been accused of being obscure - for instance, Bentham at one point
1
claimed that the term “proportionate” is more “oracular than instructive” - the way
the concept has been used in the retributivist tradition and the way it will be used in
this and ensuing chapters will not be controversial. Retributivists have, of course,
given a very different content to some of the more detailed sub-views inherent in
proportionalism, but the overall idea is relatively simple and can easily be spelled
out. In this chapter, I shall start by explicating more precisely what proportionalism
amounts to and then turn to the more controversial question of how the principle has
been morally justified by its proponents.
Readings in the early literature of the renaissance of retributivism might
leave the impression that the justification question somehow rests on a
misunderstanding. A position which, in the 50’s and 60’s, gave rise to much
discussion on the relation between guilt and punishment was that this relation is not
2
of an ethical but of a logical nature. With this view, the traditional dispute between
utilitarians and retributivists on whether it can ever be justified to punish an innocent
person was resolved. Punishment of the innocent would not be wrong but would
simply be a contradiction in terms. Likewise it might be claimed that ex definitione
punishment can only be proportionate to the crime. To claim otherwise, that is to
suggest that one should for some reason or another punish disproportionately, would
be to commit a logical failure. A demand of justification would thus be
misconceived. However, the problem with a definitional stop is that rather than
solving a question on punishment it merely transforms it to another question on
whether it can be justified to inflict, through legal mechanisms, harmful measures
3
which are not proportionate to a crime. Thus, recent proportionalists have rightly
regarded it as essential to provide justifications in favour of the proportionality
principle. In fact, what today complicates the justification question is that a number
of different sorts of justifications has been suggested. As anyone familiar with the
recent discussion on punishment will know, retributivism connotes a number of
theories which at the detailed level are quite diverse and which also provide very
different justifications of proportionalism. In an often quoted article from the late
4
70’s Cottingham pointed at the “Varieties of retribution” . And, as Walker’s recent
5
follow-up article “Even More Varieties of Retribution” witnesses, the number of
versions and suggested justifications has not declined during the past decades.
11
12 JESPER
THE ETHICS RYBERG
OF PROPORTIONATE PUNISHMENT
What I shall do is to distinguish three overall accounts of retributivism: the

simple desert theory, the expressionist theory, and the fairness theory. Though these
categories do not exhaust the range of possible versions of retributivism and
defences of proportionalism they, nevertheless, capture both the majority of theories
and indeed the most influential ones. The reason for outlining the theories is not
only to assess the justifications provided in favour of proportionalism but also to
provide a firm ground for some of the more detailed discussions in the following
chapters. In so far as there are positions which fall outside the three theories but
which offer interesting answers to some of the considered problems, these will be
brought forward as the discussion proceeds. For the present it is reasonable to start
by considering what is in the first place meant by proportionate punishment.
1. WHAT IS PROPORTIONALISM?
That the question of punishment distribution should be answered by adopting a

principle of proportionality might at first sight seem like nothing but a platitude.
After all, any moral theory of punishment includes some notion of proportionality,
prescribing that a punishment should be proportionate to what justifies it. For
instance, an adherent of rehabilitationism might suggest roughly that a punishment
should be proportionate to what it requires to rehabilitate a perpetrator, and a
deterrence theorist might likewise claim that punishments should, with the relevant
weighings, be proportionate to what is required to deter potential criminals.
However, when theorists consider proportionality in punishment what they typically
have in mind is a particular kind of proportionality. What the standard formulations
express, such as the claim that a punishment should “fit”, “match”, or simply be
“proportionate” to the crime, is a relation between certain aspects of respectively the
punishment and the crime, namely, the severity of the former and the seriousness of
the latter. The proportionality principle can be put as the view that a criminal should
be punished such that the severity of the punishment is proportionate to the
seriousness of the crime or, oppositely, that it is morally prohibited not to treat
criminals punitively in a way that is warranted by the gravity of their conduct. In
contrast to the former instances of proportionality, the standard interpretation is
characterised by its essential retrospective orientation. It is important to make clear
what this implies.
The view that a punishment should be proportionate to what is warranted
by the gravity of the criminal conduct interpreted in the sense that a crime of a
certain degree of seriousness should be punished more severely than another crime
of less seriousness, has now and then been defended on purely forward-looking
grounds. For instance, Beccaria held that if crimes of unequal seriousness - such as
assassination, poaching and forging - are punished equally severely this will
undermine people’s ability to distinguish between their seriousness. In the same
manner, Bentham advocated the view that if the state rates punishments according to
the gravity of the crimes then potential criminals will be induced to prefer the less
serious crime rather than one more serious. For instance, one of his arguments is that
“If then, for giving you ten blows, he is punished no more than for giving you five,
PROPORTIONALISM AND ITS JUSTIFICATIONS 13
the giving you five of these ten blows is an offence for which there is no punishment
at all: which being understood, as often as a man gives you five blows, he will be
sure to give you five more, since he may have the pleasure of these five for
6
nothing” . Other arguments in favour of utilitarian-based proportionality have also
7
been presented by more recent theorists. However, whether utilitarianism or other
consequentialist positions imply that more serious crimes should be more severely
punished is obviously an empirical question but not a question with which we shall
be engaged in the following (though it should be mentioned that some of the
empirical premises in Beccaria’s and Bentham’s arguments are far from being well
8
sustained). As proportionalism will be understood here - and indeed as it is
standardly interpreted - the principle has the form of a deontological constraint
characterized by an essential backward-looking orientation. The forward-looking
nature and the idea of trade-offs which characterize consequentialism is exactly
what proportionalists have emphatically objected to in their advocacy of the
9
proportionality principle. However, with these points about the justificatory
orientation and the form of the proportionality principle settled, we are left with the
question of what it more precisely means that a punishment be proportionate to the
gravity of a crime.
That a punishment should be proportionate to the seriousness of the
criminal conduct might include considerations of two sorts indicated in such
frequent complaints as, for instance, that it is morally unacceptable to punish a
brutal violent crime less severely than an economical crime, or that it is unjust to
respond to a theft with a punishment of ten year’s imprisonment. The first statement
can be made independently of knowledge of the actual punishment level, the second,
independently of knowledge of how other crimes are punished. The distinction thus
goes between a relative and a non-relative aspect of proportionate punishment or, as
10
might be said, between ordinal and cardinal proportionality. Ordinal
proportionality requires that a punishment should reflect the seriousness of the
crime, in the sense that its severity should comport with the severity of punishments
for other crimes. It is a purely comparative requirement. It implies that persons
convicted for crimes of different seriousness should receive punishments
correspondingly rated in terms of severity. In so far as theft is considered less
serious than burglary, the thief should be punished more leniently than the burglar. It
also implies that persons convicted of equally grave cases of criminal conduct are to
be allotted equally severe punishments. This implication, which I shall henceforth
refer to as the “paritycondition”, has often been particularly emphasized. Galligan
even says that a simple way of putting the proportionality principle is “like cases
11
should be treated alike” . This has often been pointed at as the reason for
implementing determinate sentencing systems though, of course, the claim that
crimes of different gravity should be correspondingly differently punished might
just as well provide this reason. As mentioned, ordinal proportionality requirements
can be satisfied independently of how the actual punishment level is set. A
sentencing system which imposes a minor fine for a rape or, alternatively, several
years of imprisonment for a parking offence might well satisfy ordinal
14 JESPER
THE ETHICS RYBERG
proportionality (though, of course, not if it does both). However, such punishments

might still in one sense be regarded as grossly disproportionate: they might be seen
as violating cardinal proportionality. Cardinal proportionality requirements concern
the way punishments non-relatively or intrinsically comport to specific crimes. They
deal with the question of how crime scales and punishment scales should be linked.
With the distinction between ordinal and cardinal proportionality in hand, it
follows that the proportionality principle can be given different interpretations. One
version of proportionality might be to accept ordinal proportionality and a strict
cardinal proportionality requirement according to which it is possible to mete out the
appropriate punishment for each particular crime. With this view, each crime itself
contains properties which makes it possible to determine directly, irrespective of
comparative judgements, the severity of the punishment that should be attached to it.
A principle prescribing harm-for-harm equivalence between crime and punishment
would constitute - given, of course, that harm is considered the parameter of
seriousness and severity - an example of this version of proportionalism. Another
version would be to maintain ordinal proportionality but to reject the
determinateness of a one-to-one interpretation of cardinal proportionality. Cardinal
proportionality requirements might be claimed to set only certain limits to what
should for each crime constitute an appropriate punishment. With this view, one
might judge a particular punishment cardinally disproportionate without being
bound to the claim that there is only one punishment which is proportionate. In fact,
one might go a step further by simply rejecting the existence of any sort of cardinal
proportionality requirements. It might be held that all that morality requires is that
ordinal proportionality be observed.
The label ”proportionality” has sometimes been reserved for the sort of
positions which do not contain a one-to-one cardinal requirement, i.e. for instance,
as a principle a retributivist might defend if he does not accept a strict interpretation
of lex talionis. However, in the following I shall use the term broadly, that is, as
embracing each of the outlined views differing with regard to cardinal
proportionality. In other words, the principle of proportionality covers views on
punishment distribution which at least requires that ordinal proportionality is
observed. Though most retributivists would certainly hold that some cardinal
proportionality requirements should also be observed, this interpretation of the
principle also covers possible hybrid theories which combine the ordinal
proportionality requirement with forward-looking reasons for punishment. Whether
a principle of punishment distribution which only partly prescribes ordinal
proportionality can avoid some of the problems which are brought forward in the
following chapters will be considered in a separate discussion in a much later
chapter. Now, with these prefatory definitional points settled it is time to turn to the
more substantive discussion of how proportionalism can possibly be justified.
2. THE SIMPLE DESERT THEORY
The cardinal concept in the various versions of retributivist theories of punishment

is “desert”. Whether this need be so if retribution is understood in its etymological
12
sense - as a pay back - can perhaps be discussed , but etymology is not the arbiter in
philosophy, and it is beyond dispute that desert forms the core in theories to which
the retributivist label has standardly been applied. In fact, desert is often regarded as
a defining characteristic of retributivism. For instance, Dolinko has suggested that
we should think of a person as retributivist simply if he justifies punishment by
“appealing to the notion that criminals deserve punishment rather than to the
consequentialist claim that punishing offenders yields better results than not
13
punishing them” . Unsurprisingly, the concept therefore also figures in the
expressionist theory and the fairness theory to which we shall return below.
However, as a point of departure it is reasonable firstly to consider a less complex
theory which I shall here refer to as the simple desert theory.
In general terms, desert claims ascribe desert to someone or something on
the ground of characteristics possessed or things done by the person or thing. As the
studies of such claims have clearly revealed, there can be large variations between
what can figure as the deserving part, on what grounds something is deserved, and
on what is deserved. For instance, though agents are perhaps what first come to
mind as the parties to which desert applies, desert claims in ordinary language have
a much wider scope. Artefacts as well as non-human objects can be said to deserve
something. It makes perfect sense to claim that “the manuscript deserves
14
publication” or to speak of Ayers Rock being deservedly famous. Moreover, it is
clear that though reward and punishment is often what is claimed to be deserved, the
two categories do not exhaust the scope of possible objects of desert. And there
seem to be no restrictions in principle on what can serve as the ground of desert,
except for the fact that there must be a base in virtue of which something is
deserved. To claim that something is deserved for no reason at all clearly contradicts
the logic of desert claims. For many ordinary language desert claims it is clear that
they do not have a moral content. However, a desert-claim which is regarded as
morally significant and which constitutes the core of the simple desert theory - as
has been advocated, for instance, by Mundle, Davis, Kleining and others - is that: a
15
wrongdoer deserves to suffer.
That one should treat people in accordance with what they deserve is
sometimes defended as a way of granting people the power to determine their own
16
fates. In a society where much depends on mutual cooperation, the practice of
acknowledging deserts gives people control over whether others will treat them well
or badly. However, when it comes to the view that wrongdoers deserve to suffer,
this is often regarded as something which is not instrumentally good but rather
something which is itself of basic moral value. This is clearly indicated in Kleinig’s
exposition of the view. He illustrates the point by imagining the case of a Nazi war
criminal who has found his way to an uninhabited island and has managed to carve
17
out an idyllic existence for himself. When he is discovered thirty years later, he has
no desire to leave or to cause further trouble. The question is whether he should be
punished. Kleining’s answer is in the affirmative. On his account the principle “that
the wrongdoer deserves to suffer seems to accord with our deepest intuitions
18
concerning justice” . Along the same lines, Davis imagines an old-style Hollywood
16 JESPER
THE ETHICS RYBERG
Western in which an irremediable wicked villain meets an unpleasant end. The

feeling that what happened to this person was altogether fitting does, in Davis’ view,
reflect the basic and widespread intuition that there is “intrinsic value in the
19
suffering of the guilty” .
In order to provide justification of a punishment system on desert-theoretic
grounds, there is a number of points which are in need of clarification. For instance,
an important question concerns the deontic implications of the desert principle. The
mere fact that “a wrongdoer A deserves a punishment P” does not in any
20
straightforward way entail that “someone ought to give P to A”. This is perhaps
especially relevant when the object of desert is suffering, because suffering is
something we usually have a duty not to inflict on others. Thus, if the desert
principle is supposed to justify institutionalized punishment it must apparently be
argued that the principle does impose obligations on others. Another well-known
point concerns the fact that no punishment system would punish just any kind of
moral wrongdoing. Thus, criteria for when the machinery of the legal system should
be put to work must be developed. Finally, even though it is agreed that the
wrongdoer’s desert should be observed, it is not necessarily clear that the object of
desert should be suffering. It might be held that a fitting response to wrongdoing
would be reproach, blame, reproof or criticism; in which case it would no longer be
obvious that punishment would be the appropriate instrument.
These points indicate that there are several challenges which adherents of
the simple desert theory must meet in order to provide a firm ground for a
punishment system. However, though the questions are obviously important I shall
not discuss them in further detail. Rather, what is of interest here is our cardinal
question, namely, whether the desert principle succeeds in providing a justification
of proportionalism. That is, does the fact that a wrongdoer, or more specifically a
criminal, deserves to suffer, justify the claim that the severity of the punishment
should be proportioned to the seriousness of the crime committed? Desert theorists
certainly believe so. The argument on which the view is based can, I believe, be
reconstructed as follows:
(1) A criminal deserves to suffer proportionately to the seriousness of the crime

committed.
(2) A punishment is more severe the more suffering it inflicts on the punished.
(3) Therefore, a criminal should be punished in such way that the gravity of the
punishment is proportionate to the seriousness of the crime committed.
Obviously, this is only a rough outline of how the desert-theoretic argument goes.
However, it does succeed in underlining two premises on which the argument is
based. Premise (2) simply concerns the relative ranking of punishments. Though we
have not yet considered this question (it will be thoroughly discussed in chapter 3),
it certainly seems reasonable to regard a punishment as more severe if it involves the
infliction of more suffering on a perpetrator. Let us therefore, for the present, regard
(2) as uncontroversial. Premise (1) states that one deserves to suffer more the more
serious a misdeed one has committed. There is apparently not full agreement
amongst desert theorists on how (1) relates to the basic desert principle: that a
criminal deserves to suffer. Some seem to believe that (1) follows from the basic
principle, while others hold that the two principles are logically independent, but
that (1) must be added to the basic principle in order to obtain a complete theory of
21
desert. Who is right in this respect need not bother us here. Let us simply assume
that if one accepts that a criminal deserves to suffer then one should also accept that
a criminal deserves to suffer more the more serious the crime he has committed.
Now, given the assumption that the basic desert principle is correct, does the
argument then succeed in justifying proportionalism? As mentioned there is a jump
from mere desert sentences to sentences which express that someone, e.g. a
sentencing system, ought to impose punishment on a perpetrator. However, even if
we accept this not-explicitly-set-out premise, the argument nevertheless suffers from
a serious defect which undermines the inference.
What links the premise on a criminal’s deserved suffering to the premise on
punishment is the fact that the punishment involves the infliction of suffering or
hardship on the one who is punished. However, suffering can be caused in a variety
of ways. A person might suffer from a painful disease, the loss of a friend or a close
relative, the loss of a job, and so on. Though not infinite, the list of possible causes
to suffering or hardship is obviously very long. But this is fatal with regard to the
justification of proportionalism. Suppose that A and B have each committed a crime
of the same degree of seriousness, and that it can be foreseen that A in the near
future will undergo severe hardship, while there is no reason to believe that this will
be the case for B. In order to make sure that both A and B undergo the suffering
which is warranted by the seriousness of the crime, A should, if at all, be punished
much less severely than B. The same might, of course, be the case even if A had
committed a crime which was more serious than the one committed by B. In order to
be valid, the argument in favour of proportionality would have to presuppose the
obviously false premise that the imposition of punishment is the only way in which
one can make someone undergo suffering.
To contend that proportionate punishment is necessary because one can
never be absolutely sure that a criminal will in fact undergo a predicted future non-
punitive suffering, is obviously not a plausible answer. In many cases, it is possible
to predict that a person will, in the immediate future, experience severe suffering,
e.g. if the person is ill. Moreover, the objection does not even have to involve future
suffering to undermine the proportionality argument. Consider, for instance, the case
of Dr. Bergman, an ordained rabbi acclaimed for his work in charity and
philanthropy; he pleaded guilty to defrauding the government by inflated claims for
medicaid payments to his nursing homes. The incident attracted enormous publicity
and the press vilified Dr. Bergman for a number of evils of which he was innocent.
The considerable humiliation Dr. Bergman suffered throughout his prosecution was
used by his lawyers as an argument against imprisonment: they contended that he
22
had already been punished enough. Likewise, it is easy to imagine situations in
which a perpetrator is racked by guilt or feels tremendous anxiety about a possible
prospect of imprisonment. There are thus several ways in which a criminal may
suffer severely after his crime is committed but before conviction; this means that,
18 THE ETHICS OF PROPORTIONATE
JESPER RYBERG PUNISHMENT
in comparison to other criminals who have perhaps committed equally serious

crimes but who have not undergone this sort of non-punitive suffering, he has
already paid part of his desert debt. Thus, prospects of future suffering need not be
involved to undermine proportionality.
As a matter of fact, there is even more to this objection to the
proportionality argument. In a somewhat ignored introductory article to the
anthology Philosophical Perspectives on Punishment, Ezorsky argues that, though
criminal A may have committed crime C at time t1 and deserves to suffer S and no
more than S for having done C, one cannot conclude that A deserves to suffer S at
23
t1. This is due to what Ezorsky calls “the whole life view” on criminal desert,
according to which not only the suffering which follows after a crime is committed
but also suffering prior to the crime should count in the final computation of desert.
Therefore, it is a non sequitur to conclude that A deserves to suffer S at t1, since a
possible pre-crime suffering may nullify the post-crime desert debt. If this is correct,
it adds a further dimension to the outlined challenge to the proportionality argument.
Should we accept the whole life view? Is there reason to reject the possibility that
present desert debts can be affected by pre-crime suffering?
If one looks into the logic of desert claims it may be observed that it is
24
often emphasized that such claims are always backward-looking. This refers to the
fact that desert bases refer only to the present or past features of a deserving person.
They never refer to features the person will have in the future. If a person deserves
to be punished, this is not because the punishment is expected to deter others or to
reform him, but because he has done something morally wrong. However, the fact
that the desert base must refer to past or present features in no way excludes the
possibility that the object of desert - in casu the suffering which a criminal deserves
- may be temporally prior to the desert base. Thus, whether pre-crime suffering
should count cannot be answered on purely conceptual grounds. In fact, according to
Ezorsky, an argument can be given in favour of counting in pre-crime suffering. She
considers a person who has served one year in prison, convicted for a crime he did
not commit, but who, after his release, decides to commit the very crime for which
he was punished. Now, on the one hand, the person deserves restitution for having
been undeservedly deprived of his freedom for one year and, on the other, he
deserves a year in prison for having committed the crime. What should this amount
to in the final computation of desert? Though Ezorsky admits that it would be
“moral madness” not to punish the person, because the consequence will be that any
person punished undeservedly would have earned the right to commit a crime, she
nevertheless believes that, with regard to what the person deserves, the case shows
that we should not determine the person’s desert irrespective of his past undeserved
25
tribulations. Whether or not one accepts this argument, it is a fact that proponents
of the proportionality argument must provide a reason as to why pre-crime suffering
should not count. This problem certainly does not admit of an easy answer.
There is, though, another way in which the challenge to the proportionality
argument could be met. If it is possible to argue that criminal desert is unaffected by
non-punitive suffering, then this will exclude references to the suffering which the
criminal either will undergo or has undergone in the past and which will be or was a
result of illness, natural catastrophes, stigma, or whatever other causes there may be
for hardship. (Strictly speaking, this is not itself sufficient to save proportionality
because, as indicated in Ezorsky’s example, there can be cases in which a person has
previously undergone undeserved punitive suffering, but the answer would certainly,
by excluding other kinds of suffering, avoid the most devastating objection to the
argument.) In order to defend this approach it is necessary to explain what exactly
constitutes the purpose of punishment if it is not merely the infliction of suffering on
perpetrators. In so far as it can be argued that punishment does serve another
purpose to which the infliction of suffering is perhaps only a possible mean, it may
be possible to retain an argument in favour of proportionalism. However, this will be
26
tantamount to giving up the simple desert theory.
In sum, what we have seen therefore is that seeking to justify a punishment
system and, more specifically, a proportionalist allocation of punishment, on the
ground of the simple desert theory, does not seem like a promising project. The
problem of non-punitive suffering simply undermines the justification. However,
this obviously does not imply that there is no room for desert claims in punishment
theory. It just means that proportionalism cannot in the outlined way be justified on
the ground of deserved suffering. This naturally brings us to some of the more
refined theories and justifications of proportionalism to which we shall now turn.
3. THE EXPRESSIONIST THEORY
A dominant view amongst theorists contemplating criminal sanctions has been to

stress the expressive character of punishment. A number of philosophers and legal
27
scholars have defended versions of expressionism. The thought that punishment
can be seen as a language, that is, as a way of communicating a message to the
criminal and perhaps other possible recipients, is not new. However, one of the
philosophers who have in the recent epoch drawn attention to the expressive element
in punishment is Feinberg. In an influential article, Feinberg points at a deficiency in
standard definitions of punishment, namely, that they generally ignore the fact that
punishment, in contrast to mere penalties, is a device for the “expression of attitudes
28
of resentment and indignation, and of judgments of disapproval and reprobation” .
Rather than being an evil simpliciter, punishment has, as Feinberg puts it, “symbolic
significance”. Corresponding thoughts have subsequently been developed in a
number of theories which do not merely consider the expressive element of
punishment in relation to a discussion of definitions, but rather sees it as the raison
d’etre of punishment, that is, as part of what in the end justifies a punitive response
to criminal conduct.
What is of interest in the present context is not all versions of the view that
punishment is a sort of communication but, more narrowly, those versions which
seek to provide a rationale for proportionalism. This delimitation excludes a number
of expressionist theories, namely, those providing a forward-looking justification of
punishment. Bentham, for instance, is apparently well aware of the expressive
aspect of punishment and emphasizes it in his discussion of “indirect means of
preventing crimes”. More generally, proponents of deterrence might rely on the
20 JESPER
THE ETHICS RYBERG
communicative dimension of punishment. With them, punishment might be seen as

29
a way of conveying a message like “Obey, or else!” to potential criminals.
However, given the definition of proportionalism, such accounts which incorporate
expressionism as part of a consequentialist justification of punishment are obviously
not relevant. In order to illuminate those versions of expressionism which proclaim
to support proportionalism a few initial questions naturally come to mind. What is
the message that is expressed? By whom, and to whom? And what exactly is the
purpose of the communicative endeavour?
Though adherents of expressionism do not always fully agree as to what
exactly it is that punishment expresses or communicates, a standard claim is that the
punishment of the criminal is a way of showing that he has performed a
reprehensible act and that he is disapproved of for having done so. Or perhaps more
precisely, that punishment expresses denunciation or condemnation of the criminal
misdeed. With regard to the question of whom addresses whom, the most usual
outlook is that the recipient is the criminal and that the condemnatory message is
delivered by those who officially impose the punishment on behalf of the whole
community. However, some adherents to the view also believe that third parties are
in different ways involved. For instance, a message is communicated to the victim,
namely, as Lucas puts it, that “the misdeed, although perpetrated by a member of
society is not to be construed as being in any way an action of society, and that
society identifies not with the criminal but with the victim and it is his right that it is
30
determined to uphold” . Moreover, several adherents of the view also claim that the
message, that the criminal conduct was reprehensible and that such actions should
be eschewed, is brought to bystanders or the public at large. However, though there
may be several recipients, it is usually underlined that the communication primarily
addresses the perpetrator.
The interesting question, of course, is what exactly is regarded as the
purpose of the expressive or communicative enterprise? That is, what should one
aim at by expressing or conveying the appropriate messages? At this point there is
divergence between different expressionist theories. One possibility would be to
contend that the purpose is merely to express denunciation or condemnation. That is,
once the expressive act has been performed the purpose is fulfilled. However, as
indicated, this is clearly not what modern expressionists usually have in mind. Mere
expression only involves someone who expresses, but nothing further. The talk
about a “recipient”, “conveyance of a message”, and “communication”, clearly
indicates that something further is aimed at. Nozick has, in this respect, introduced a
31
useful distinction between “teleological” and “non-teleological” retributivism. The
teleological retributivist aims for an effect in the criminal, e.g. that correct values are
recognized and internalized for future actions. In short, the goal is some sort of
moral transformation. In contrast, the non-teleological retributivist does not aim at
transforming the criminal. The goal is more modestly that of confronting the
criminal with a message, for instance, what count as the correct values. As an
illustration of the distinction, Nozick contends that the non-teleological goal
corresponds to that of making a recipient of a verbal message understand the
assertion, whereas the goal of the teleological retributivist corresponds to the

recipient’s accepting what is said.
Both the teleological and the non-teleological answers have been defended
by recent expressionists. Nozick himself believes that something valuable is
achieved when the non-teleological goal is fulfilled, even if the further teleological
result, the moral transformation, does not occur. Along the same lines, a leading
expressionist like von Hirsch also expresses affiliation to the non-teleological view
when he claims that, though some kind of moral response is expected from the
criminal (e.g. expression of concern or efforts at better self-restraint) when the
message concerning his wrongful act is conveyed, the censure “is not a technique
32
for evoking specified sentiments” . In his view, neither the repentant criminal who
has regretted his wrongdoing nor the defiant criminal who will not accept
judgements of disapproval, should be exempted from blame. Though they are both
incorrigible they are nevertheless capable of understanding another’s assessment of
their conduct. In contrast to this non-teleological aim other theorists adhere to
teleological versions or expressionism. For example, according to Duff, the purpose
is to bring the yet unrepentant criminal to repent of his crime. Duff believes, as we
shall see, that the aim is a “penitential reform”. Thus, at this point recent
expressionists are split between two approaches to what should be regarded as the
communicative aim.
Though we have not yet looked into some of the further details of the
different versions of expressionism, the previous outline does not make it hard to
imagine that the theory may have implications with regard to the distribution of
punishment. In fact, an argument in favour of proportionalism seems to follow
pretty straightforwardly. If we accept the claim that a criminal should be blamed for
his misdeed then it seems reasonable, and indeed in accordance with our ordinary
life comprehension of blame, to hold that he should be blamed more the more
reprehensible his crime was. Therefore, given the assumption that blame should be
communicated through punishment and that the degree of the blame that is
conveyed varies with the severity of the punishment, it follows that a more serious
crime should be punished more severely. This step from expressionism to
proportionalism is explicitly stated by von Hirsch in the following way:
1. The State’s sanctions against proscribed conduct should take a punitive form; that
is, visit deprivations in a manner that expresses censure or blame.
2. The severity of the sanction expresses the stringency of the blame.
3. Hence, punitive sanctions should be arrayed according to the degree of
33
blameworthiness (i.e. seriousness) of the conduct.
Despite the differences between the various versions of expressionism, the argument
does seem to provide a general framework for how the expressionist defence of
proportionality will go. But should we accept the argument? I believe there are
several reasons to be sceptical, even if one accepts the basic view that it is morally
valuable to censure wrong-doings.
22 JESPER
THE ETHICS RYBERG
An interesting preliminary question which naturally arises is, why should

the criminal be punished at all? What I have in mind here is not the question of why
a criminal should be blamed for his misdeed, but rather why, once we have accepted
that he should be blamed, this should be done through the infliction of punishment.
That is, why should a perpetrator undergo hard treatment? As pointed out by Hart,
the “normal way” of expressing condemnation is by the use of words. Just as we
may express admiration or gratitude to a person simply by thanking or praising him,
that is, by purely verbal means, it would seem that we could just as well use the
same means in cases where the message involves condemnation. Though it is correct
that we can communicate to other people by other than purely verbal means, there
certainly seems to be a tension between, on the one hand, the claim that we should
communicate our condemnation of his conduct to the criminal, and, on the other, the
34
claim that the criminal should be punished. While the simple desert theory, as we
have seen, had no problem in explaining why criminal conduct should be responded
to by punitive measures, namely, because punishment involves the infliction of
suffering on the wrongdoer, this is no longer obvious if viewed from an
expressionist perspective.
In fact, even if we accept that punishment is one way to convey a message
to the criminal, it would surely be a dubious moral principle that would prescribe
punishment of a person, if the desiderata of the communicative enterprise could just
as well be satisfied by other means not involving hard treatment. Thus, an
explanation of why hard treatment should be imposed on the criminal is required.
Moreover, the question is specifically crucial with regard to the justification of
proportionalism. If the condemnation need not take a punitive form, then obviously
it does not follow that a more serious crime should be responded to by a more severe
punishment. The perpetrator who has committed a more serious crime could be
blamed more than the one who has committed a less serious crime but neither of the
two would have to be punished. That the response to proscribed actions should, as
von Hirsch indicates in premise (1), take a punitive form, is therefore vital in the
defence of proportionality.
As far as I can see, the arguments which expressionists have presented in
favour of hard treatment fall into one of the following three groups: either it is
claimed that 1) hard treatment is necessary in order to make the criminal understand
the message that is conveyed; or 2) hard treatment is required in order to fulfil a sort
of reformative aim beyond the mere understanding of the message; or finally 3) hard
treatment is required for preventive reasons. As will be clear, it seems to me that the
arguments either fail to provide a convincing justification of hard treatment, or they
succeed in providing the justification but only at the cost of relying on premises
which themselves threaten proportionality. To see this, let us consider the three
arguments seriatim.
According to the first argument the infliction of a punishment is the only
way we can hope to address the criminal. Though adherents to this argument usually
defend a non-teleological view, that is, though the aim is not to transform the
criminal, hard treatment is, nevertheless, regarded as the only language the criminal
understands. For instance, Lucas contends that though the point of punishment “is to
make them [the perpetrators] understand that the reprimand is really meant”, some
kind of formal disapproval will not be sufficient because “some people are too
hardened to care much .... [o]n their scale of values they will have got away with it,
unless the reprimand is given tangible forms in terms which are meaningful to them.
35
Words mean little.” A related outlook is developed by Primoratz who claims that
“merely verbal condemnation is not likely to reach its immediate addressee and to
be fully understood by him. Regrettably, although perhaps not surprisingly, many
36
criminals are oblivious to mere words”. In my view these claims are not
convincing. If we are to take serious the view that - though a further result in terms
of a moral transformation might perhaps be hoped for - the aim is to make the
criminal understand the message expressing disapproval or condemnation, then
these claims seem simply to be false. Surely even criminals can understand a verbal
message that is put in an appropriate language. The picture of the criminal as a
creature incapable of being addressed in ordinary language is certainly naive. And
even if such non-verbal monsters do exist they are surely not representative of the
criminal in general.
However, perhaps this answer is too swift. After all, we do now and then in
our ordinary lives make statements along the lines “you do not fully understand until
you have tried it yourself”. For instance, it apparently makes some sense to claim
that one does not fully understand what it is like to deliver a child if one has not
been through it oneself. Or that one needs to have undergone a depression oneself in
order to understand what it really implies. What is at stake in such formulations
comes close to the traditional distinction between knowledge by definition and
knowledge by acquaintance. Thus, could hard treatment be defended as the only
means which make the criminal understand (by acquaintance) what he has done to
the victim? I think not. There are several problems with this suggestion.
Firstly, the plausibility of the argument would surely depend on what
exactly it is that is communicated to the criminal. Talk about understanding by
acquaintance does not seem plausible if the message, as expressionists typically
claim, is one of disapproval or condemnation. After all, the point is not to make the
criminal understand what it is like to be condemned, but rather to condemn him. It
makes more sense if the message, as Nozick suggests, is something like “this is how
37
wrong what you did was” . Secondly, it would still have to be established that it is
only by hard treatment that the relevant kind of understanding can be obtained.
Thirdly, there are many kinds of crimes with regard to which it is far from obvious
that one would get a clearer understanding (by acquaintance) of what one has done
by being inflicted with the hard treatment that punishment involves (e.g. does it
make sense to claim that hard treatment can make one understand (by acquaintance)
what it is like to be killed by a drunken driver?). Fourthly, even if one considers
violent crimes it may often well be the cases that the criminal has himself in his
lifetime already experienced something similar which means that hard treatment
would not be required to evoke the aimed understanding (by acquaintance). Thus, all
in all this interpretation does not seem to provide the argument with further
plausibility.
There are, however, some passages in the writings of the philosophers
quoted above, which indicate that something may be meant by “understanding” that
24 JESPER
THE ETHICS RYBERG
differs from what we mean when we say that we understand a simple message like
“you should not have acted like that”, and also differs from an understanding by
acquaintance gained through actual experiences. For instance, Lucas claims that the
criminal will be “unmoved” if a judge merely berates him. And Nozick who, as we
have seen, specifically claims that the aim of the non-teleological view corresponds
to that of understanding an assertion, in contrast to the teleological aim of causing a
moral transformation, nevertheless also holds that punishment makes the values that
38
the criminal has flouted have some “significant effect on his life” . Thus, it
sometimes seems as if there is a subtle distinction between merely understanding
39
and really understanding a message. If that is the case, then it must be regarded as
most unfortunate that nothing has been done to clarify what it means to “really
understand” something. As long as this is not made clear - and certainly it is far
from obvious what such a distinction implies - there is not much of an argument.
However, a final suggestion might come to mind. Perhaps the idea is not
simply that hard treatment is required in order to understand the condemnatory
message but rather to understand something further, namely, that the condemning
part is really meant. There could be some messages which are only believed by a
recipient to be really meant by the addresser if they are accompanied by certain
40
actions. Consider the following example which I owe to Baldwin. Suppose that a
lover is to communicate his love to the beloved. In that case, merely formulating the
appropriate words may not be sufficient to ensure that the content of the message is
really meant: certain acts must also be performed witnessing that this is the case. He
must spend time with the beloved and do what else is required to vindicate such a
claim. Now, could the expressionist resort to a similar sort of suggestion when it
comes to the justification of hard treatment? The argument would then have to be
that what is important in the way we should address a perpetrator is, firstly, that an
appropriate message is conveyed to him or her and, secondly, that the message is
conveyed in relation to a set of actions which are designed to ensure that the content
of it is really meant. While the first aim can be performed in normal ways of verbal
communication, this is not the case with regard to the second aim the fulfilment of
which requires hard treatment. Perhaps it is something along these lines which the
expressionists quoted above have in mind.
However, such an approach suffers from several weaknesses. A first
question is what the content of the message must be in order to fit into this kind of
suggestion. In the example it is the feeling of love that is communicated. If a person
sincerely expresses such a feeling then it is natural to expect that it will be
accompanied by certain actions. However, if what is conveyed to the criminal is - as
most expressionists hold - a condemnatory message then it is no longer equally
obvious that some sort of back-up actions are required to ensure that the message is
really meant. Furthermore, one might ask whether ensuring that a message is really
meant by the addresser is so important that it itself can carry a justification of the
imposition of hard treatment on criminals. However, what is more important is that
even if we accept this kind of justification it is not clear that this will work as part of
a justification of proportionality. The reason why a more severe punishment is
required for a more serious crime is usually held to be - as stated in premise (2)
above - that the severity of the punishment expresses the stringency of the blame.
But, according to the just outlined conjecture, hard treatment is not required in order
to communicate blame, or whatever the content of the message is, but only to ensure
that the message is really meant. However, it is not clear that the fulfilment of this
purpose requires more hard treatment, i.e. a more severe punishment, the stronger
the content of the message. In other words, it might well be sufficient to impose a
certain degree of hard treatment to ensure that the message is really meant, no matter
whether the message is more or less condemnatory. If A has committed a crime
which is a little more serious that the one committed by B, is it then not possible that
the requirements could be satisfied by condemning A more than B, in some verbal
or symbolic way, and then inflict the same degree of hard treatment on the two as a
way of ensuring that the condemnation is really meant? In my view it is hard to see
why this should not be possible. But that means that if hard treatment is not required
merely to communicate the appropriate degree of blame, but to ensure seriousness or
sincerity in the communication, then it is no longer clear that more severe
punishments are required as responses to more serious crimes.
Thus, in sum, the first suggestion, that hard treatment is the language of
criminals, no matter whether this is understood as the contention that hard treatment
is needed to make a criminal understand the message, or as the just considered
proposal that it is needed to communicate that the message is really meant, does not
succeed in providing a plausible justification of the necessity for hard treatment on
which a justification of proportionalism can be based. So much for the first
argument.
The second approach to the justification of hard treatment is to see it as
something that is required with regard to obtaining certain results which go beyond
the mere evocation of understanding in the criminal. According to this argument,
hard treatment plays a role in relation to some kind of reform of the criminal. At the
outset, a justification along these lines might strike one as surprising. There is not
much evidence in favour of the claim that, for instance, imprisonment should
contribute to a moral transformation of criminals. Rates of recidivism might even
witness the opposite. However, Duff has suggested a theory according to which hard
treatment punishment plays a reformative role by serving, if successful, as a
41
penance which the criminal comes to will for himself.
Duff agrees with other expressionists that the appropriate response to the
criminal as a moral agent is to censure him for his conduct. However, he also
believes that, though hard treatment can communicate censure, it can also be
conveyed by the conviction itself or by purely symbolic means. For him, however,
the purpose is not merely to condemn the criminal but to reach deeper than censure
does: the aim is to elicit attitudes of repentance in the criminal. Duff draws his
illustrations from religious contexts in which a sinner, by flouting shared values, has
evoked the bonds which tie him to a church community and to God. In such a case,
the community may subject the sinner to coercive treatment, not merely to inflict
suffering on him, but to bring him to understand and repent the sin and thereby to
restore himself to communion with God and his fellows. Correspondingly, Duff
contends that legal punishment should be understood as a secular species of
penance. The view is teleological, in the outlined way, by aiming at a moral
26 JESPER
THE ETHICS RYBERG
transformation of the criminal. The punishment should be a vehicle for repentance

and thus to reform in the sense that the criminal reconciles himself to the values
which his crime denied, and to the community to which he belongs.
However, Duff strongly emphasizes that the reformative aim is not one that
should be satisfied by whatever may be the most efficient means. Once we force or
manipulate the criminal’s attitude into conformity with some favoured set of values
we do not address him as a moral agent. Duff’s basic assumption is that any
tolerable system of punishment must respect others as rational and autonomous
moral agents. And this respect is not consistent with resorting to manipulative
techniques which might bring about a desired change in the criminal. However, if
we seek to change the behaviour of the criminal merely by trying to persuade him of
the correctness of the judgement that his conduct was wrong, then there is, so to say,
no hidden agenda in the way we face the criminal, and we would have addressed
him as a moral agent. In other words, the process of moral reform must be mediated
by the criminal’s own understanding. The criminal should through his understanding
be brought to repent his crime and to will and accept the punishment as an
appropriate penance. It is here Duff finds the reason as to why hard treatment is
required. The aimed penitential reform can “be achieved only by bringing the
42
offender to suffer for what she has done” . Thus, while mere censure can be
conveyed in other ways, hard treatment is on Duff’s account required to fulfil the
function of a sanction as a penance.
Should we accept this argument for the necessity of hard treatment? And
does a theory like Duff’s thereby bring us closer to a justification of
proportionalism? In my view, there are several reasons to be sceptical with regard to
whether this is the case. The main reason that I am inclined to be sceptical is that it
seems that such a justification will have to rely on a set of assumptions of human
moral psychology which strike me as highly dubious.
Firstly, I am not convinced that the function of a penance cannot be
fulfilled in ways which do not require hard treatment. According to Duff, a
penance serves several interrelated purposes: it focuses the wrongdoer’s attention
onto his wrong-doing; it symbolically portrays the character and the implications
of the wrong; it aims to make the criminal recognize and repent the wrong he has
43
committed; and it functions as a vehicle of self-reform. But is it impossible that
these purposes could be satisfied in ways other than through the infliction of
suffering on the wrongdoer? For instance, is it unthinkable that a criminal’s
attention can, through other non-manipulative techniques, be appropriately focused
on his misdeed (e.g. through concentration exercises, long conversations or
whatever)? It is hard to see why such possibilities are excluded. In fact, might one
not fear that hard treatment might tend to deflect the criminal’s attention from his
wrong-doing, focusing it on this current hardship rather than on what he has done?
Correspondingly, it strikes me as a dubious claim that it should be possible only
through hard treatment to present an appropriate symbolic portrayal of the
character of a certain misdeed. For the many other happenings in our lives it seems
that these may be symbolically portrayed in various ways, and it is hard to see why
the symbolic portrayal of a misdeed must require hard treatment. Moreover, it
seems to be a fact that various experiences which do not involve hard treatment
may work as initiators of some sort of self-reformative process. Now, the point of
these comments, it should be underlined, is surely not to deny that hard treatment
may serve the outlined purposes but simply to question whether the idea of
punishment as a penance succeeds in establishing the necessity of the infliction of
hard treatment on offenders.
Secondly, even if it is assumed that hard treatment is required in order to
fulfil the different purposes of a penance this does not necessarily mean that the path
to proportionalism is clear. The overall problem is that there might be a discrepancy
between what is required in order to communicate the appropriate degree of blame
to a person, and what it takes to give a person the opportunity to go through an
appropriate penitential reform. This discrepancy may manifest itself in different
ways.
As indicated, one of the ideas which underlines Duff’s theory is that people
are often unwilling to face up to their wrong-doings. If one has committed a wrong
there might well be a powerful temptation to evade the issue by self-deceptive
excuses or justifications. When hard treatment is inflicted on a criminal it may
therefore well be the case that the suffering is not readily accepted by the person as a
penance. The hard treatment will then function as a way of persuading the criminal
to accept as a penance the hard treatment that is imposed. The punitive suffering
which begins as a coercive attempt to attract the unrepentant criminal’s attention
should ideally become the penitential suffering which the repentant criminal accepts
for himself. However, the question is: what does a penance in this respect require?
To hold that the different purposes which a penance serve are somehow
instantaneously satisfied, that is, that it only requires a split second of suffering to
fulfil these purposes once one has accepted the suffering as a penance, does not
seem plausible. A psychologically much more plausible view is that a person who
accepts the inflicted suffering as a penance engages in a time-consuming process.
This is also indicated by Duff who says that “the task of coming to understand, to
44
repent, and truly to disown my crime may be a long and arduous one” . It is exactly
to this process the society should contribute by offering the criminal the requisite
suffering. But if that is the case then what should one do if one faces a situation like
the following.
Suppose that A and B have each committed a crime of the same degree of
seriousness and that the appropriate penitential process is estimated to require one
year of imprisonment. Suppose further that A is quickly persuaded to take on
himself the punishment as a penance, while it takes much more to persuade B. In
fact, we can assume that B is persuaded and thus realizes his need to engage in a
penitential process only at the very day before he is to be released. Now, if this is the
case, and if what basically matters is that society offers a punishment as a penance,
would the proper response then not simply be to punish A with one year in prison
while offering B a year in addition to the first he has already served? Whether or not
this would be morally plausible is not the present concern, but it is pretty clear that if
this is what the penance theory would imply then it would violate the proportionality
requirement.
Another question relating to the double purpose of conveying blame and

offering criminals the possibility of a penance concerns what one should do in cases
where a person during his punishment - e.g. a prison term - has willingly accepted
the punishment as a penance and has subsequently undergone the aimed penitential
reform. If this has been successfully carried out at a stage before the punishment
communicating that appropriate degree of blame has been completed - say, the
person has managed to go through the penitential process after half a year in prison,
while it is estimated to require one year in prison to communicate the degree of
blame that is warranted by the gravity of the crime - should the criminal then be
released before the term has been fully served (say, after half a year)? If that is the
case then this means that a person who, due to the fact that he is more easily
persuadable and possesses a psychological constitution which makes it much easier
for him to go through a genuine penitential reform, may be punished less severely
than another person who, in these respects, is differently constituted. And this might
be the case even if the two persons have committed equally serious crimes, or even
if the former person has committed a crime that is more grave than the one
committed by the latter. What this highlights is the obvious point that can be
directed broadly against teleological versions of expressionism, namely, that if the
purpose of punishment is to bring about a certain result beyond a criminal’s mere
understanding of a certain message, then it might well be the case that it requires
different degrees of hard treatment to reach this result with different persons.
The situations in which the fulfilment of the idea that punishment should
serve the function of a penance would lead either to disproportionate prolongation of
a punishment or to a punishment reduction would both violate proportionalism.
However, it might be suggested that both of the suggested implications of the
penance theory could easily be blocked, namely, by claiming that since such
prolongations or reductions would imply that the criminal would either be blamed
too much or too little relative to what was warranted by the gravity of the crime such
deviations from what would constitute the appropriate punishment in terms of the
communicated degree of blame would be unacceptable. Thus, even though Duff
45
himself contends that the penitential reform constitutes the “justifying aim” of
punishment, and though in some places he speaks as if the penitential reform may
determine punishment severity - e.g. he seems to accept that a criminal who has
46
genuinely repented his crime may be punished less stringently - it might
nevertheless be held that what basically matters is that the society conveys the
appropriate degree of censure. Thus, though it might be hoped that the criminal is
persuaded to accept the punishment as a penance and that he succeeds in going
through the reformative process, it is the conveyance of the appropriate degree of
blame that constitutes the primary purpose of punishing and which therefore
determines the degrees of punishment for different crimes. Given this view, the
theory would not lead to prolongations or reductions in the outlined cases. However,
as far as I can see, this answer is insufficient as an attempt to maintaining
proportionality in punishing.
The problem is that it may well be possible, within the framework of
Duff’s theory, to let the hard treatment which a penance requires for each individual
person determine the severity of a punishment without failing in the communication

of the appropriate degree of blame. Consider again the above situation in which the
undergoing of an appropriate penitential reform requires less punishment than what
is required to communicate the appropriate degree of blame for the crime. Suppose
that A and B have committed equally serious crimes but that, due to A’s and B’s
different psychological constitutions, it takes much less hard treatment for A than it
does for B to go through the appropriate reformative process and, moreover, that it
takes less than what is required to communicate the appropriate degree of blame to
A, while this is not the case for B. Now, even if the communication of the proper
degree of blame has primacy it does not in this case follow that A and B should be
equally severely punished. It would be sufficient to inflict on A the hard treatment
that is required for the penance and then subsequently supply this punishment with
an additional condemnatory message conveyed in a way that does not involve hard
treatment. If one accepts the idea of parsimony in punishing this alternative would
clearly be preferable. And the result might be that that while A and B had both
received a penance and had both been appropriately blamed for their misdeeds, that
A in the end had been punished less than B.
Now, the possibility of following this procedure would of course be ruled
out if it were not possible to supply hard treatment communication with
communication through other medias. However, as I shall argue below it is hard to
see why this should be impossible. Moreover, the possibility would be blocked if it
always required the same degree of hard treatment for different persons to undergo
an appropriate penitential reform, and if the degree of hard treatment required to
communicate an appropriate condemnatory message always coincided with the
degree of hard treatment required for a punishment to serve the purpose as a
penance. However, the first claim strikes me as clearly being false: different persons
would surely not always need the same degree of suffering to undergo the above
outlined purposes of a penance. And with regard to the second claim, no argument
has been presented in its support and it is certainly hard to imagine what could
possibly sustain it. Thus, in sum I believe that even if one accepts the basic idea that
a punishment should serve the function as a penance that is offered by the society to
the criminal it is, given the purposes of penance, still not obvious that hard treatment
is required and, what is more important, even if hard treatment punishment actually
is required it does not follow that the severity of the punishment should be
proportionate to the gravity of the crime that has been committed.
After having considered both non-teleological and teleological versions of
expressionism, starting from the question of how the infliction of hard treatment can
be justified, I shall now turn to the third argument, more precisely, the theory which
underlies von Hirsch’s argument for proportionalism. A proposal which has also
47
been defended by Narayan.
As their point of departure, von Hirsch and Narayan agree with other
expressionists that a criminal should be censured for his conduct and that this is the
primary function of punishment. If predatory conduct is dealt with through some
kind of neutral sanction (e.g. some sort of taxation) which does not convey
disapproval, this will be to deny the status of the person as being an agent capable of
moral understanding. Neutral sanctions would, in von Hirsch’s view, treat criminals
“as tigers might be treated in a circus, as beings that have to be restrained,

intimidated, or conditioned into compliance because they are incapable of
48
understanding why biting people (or other tigers) is wrong” . Only condemnatory
sanctions treat the actor as a person capable of choice and understanding. What is
interesting is that von Hirsch and Narayan believe - as did Duff - that reprobation
can be expressed not only through the visitation of hard treatment but also in a
purely symbolic mode. Given these alternatives the challenge, therefore, is once
again the same, namely, to explain why perpetrators should be addressed through
hard treatment. The answer given is that this way of conveying censure should be
preferred to other communicative means due to its crime preventive function. Given
the plausible assumption that crime prevention is valuable and that hard treatment,
in contrast to merely symbolically conveyed censure, will help to achieve this goal,
there is a reason to prefer hard treatment punishment.
At first glance, this answer might strike one as surprising since a
punishment system which is based on crime prevention might, depending on the
empirical conditions, prescribe sanctions which in severity differ radically from
what would follow from a censure-based account of punishment. Thus, the
challenge is to intertwine reprobation and prevention in a coherent theory.
According to von Hirsch and Narayan’s proposal, this is done by seeing prevention
as an element that holds within the censuring framework. What this means is that
only hard treatment which comports with the expression of appropriate degrees of
censure, is morally permissible. A person who is censured for a misdeed is conveyed
the message that his act is wrong and is thereby given a reason for desistence.
However, despite this reason the person may nevertheless be prone to temptation.
What hard treatment does, in contrast to other ways of expressing censure, is
provide the criminal with a further reason for resisting the temptation. Hard
treatment serves a special preventive function which provides a prudential incentive
for not breaking the law, and which thereby works as a supplement to the reason
conveyed by censure. That this bifurcated justification of punishment makes
prevention work only supplementary to censure, means that it cannot be justified to
increase, for preventive reasons, the severity of a punishment for a certain crime
beyond the level that is warranted by the seriousness of the conduct. This would, as
von Hirsch clearly stresses, be to express disapprobation to an extent which does not
correspond to the reprehensibleness of the crime. Thus, the elimination of
proportionality which would follow from a purely preventive justification of
punishment is, according to von Hirsch, avoided by giving primacy to censure and
49
letting prevention in only as an additional prudential disincentive.
Now, is this proposal more successful with regard to a justification of hard
treatment than the former proposals? And does it provide a sufficient background
for von Hirsch’s proportionality argument outlined above? The answer to the first
question is obviously empirically conditioned. Hard treatment will be justified only
in so far as it does in fact have a prudential disincentive function. Von Hirsch
himself contends that, if it does not have this function, then a society might still wish
to maintain some way to convey the requisite disapproval of crimes, but there would
no longer be need for so burdensome an institution as the criminal sanction. But, if
we accept the assumption that punishment is to some extent preventive, then von
Hirsch and Narayan have provided a justification of hard treatment. However, in my
view there is still reason to doubt that the theory succeeds in providing a plausible
base for proportionality.
As mentioned, von Hirsch strongly emphasizes that if one crime is less
reprehensible than another, but the first is nevertheless - for preventive reasons -
punished more severely, then this is objectionable because it will be to censure the
criminal to an extent not warranted by the reprehensibleness of the conduct.
However, the fact admitted by von Hirsch and Narayan - and I believe correctly -
that condemnation can be expressed in other modes than through hard treatment,
faces the two-pronged justification with a serious challenge akin to the one raised
against Duff’s theory. Suppose first, that of two crimes, C1 and C2, more hard
treatment is required to induce someone not to perform C1 than C2. That is, given a
purely preventive approach to criminal sanction, C1 should ceteris paribus be
punished more severely than C2. Suppose further, that C2 is in fact a more serious
crime than C1, and that a person should, therefore, be censured more for having
committed C2 than C1. Now, what would von Hirsch and Narayan in this case
prescribe with regard to the relative punishments of the two crimes? Due to the
primacy of censure over prevention the answer seems pretty straightforward: C2
should be responded to by a more severe punishment than should C1. However,
there is another possibility which would in fact be preferable according to the
theory’s own standards. That would be to censure the person who has committed C2
to the level at which the preventive aim is satisfied and then convey the remaining
censure in a mode that does not involve hard treatment. In that case, the
performance of C2 may, all in all, be censured more than the performance of C1, but
C1 may nevertheless be punished more severely, since it requires more hard
treatment to achieve the preventive goal with regard to C1. Therefore, given the two
assumptions which von Hirsch and Narayan accept, namely, that censure can be
conveyed by other means than hard treatment, and that there is no reason to (in fact
there is a reason not to) inflict hard treatment beyond what is required for preventive
reasons, there are cases of punishment which are not consistent with
proportionalism. And, what is important, now it can no longer be objected that the
more reprehensible crime is censured less than the one less reprehensible.
The argument, of course, rests on a number of assumptions. Firstly, it
presupposes that situations might occur in which it requires more hard treatment to
give someone an incentive not to follow a temptation to commit one type of crime,
than it requires to give an incentive not to commit another type of crime, even
though the latter crime is rated higher in terms of seriousness. Whether this is
sometimes the case is an empirical
50
question, but there is no reason to claim that it
simply cannot be the case. Secondly, the argument also presupposes that the
imposition of censure can be split up into parts and that the first part can be
conveyed through hard treatment while the second part is conveyed in another way
(e.g. in some “symbolic mode”). Considering the way we usually think about
censure in ordinary life, this assumption might seem strange. However, given the
way censure works in expressionist theories of punishment, there is no reason to
reject the premise. If a person receives two years’ imprisonment, while another only
receives one year then, according to the theory, the former criminal is censured more
than the latter. Without this assumption there would be no ground for adhering to
proportionality in the first place. But that means that the conveyance of censure is an
ongoing communicative process. And just as it is thinkable that a prison term can be
split in two (first you get the first year and after a break you get the final year), it
also seems possible to split up the censuring process in the way the argument
presupposes.
However, at this point some might still feel that there is reason to be
sceptical with regard to whether a certain degree of censure can be conveyed
through two separate medias. More precisely, it might be objected that it seems
implausible to assume that the conveyance of blame through hard treatment and
through symbolic means can function in the additive manner which the argument
presupposes. Rather, one might hold that if different communicative means are
applied the result will be that the blame is repeated and not that more blame is
conveyed. However, it is hard to see what reasons there could be for holding this
view. Consider the following example which concerns positive desert. Suppose that
A and B each receive 50$ as a reward for having performed an admirable act, but
that, since A’s act was even more admirable than B’s, he also receives some sort of
medal. Now, in comparing the messages conveyed to A and B, might not the overall
result be that since A is praised both in a non-symbolic and in a symbolic way he is
in the end praised more than B? In my view this might well be the case, and I see no
reason why it should not also be possible to “add up” when it is censure and blame
that is communicated. Of course, it might sometimes be difficult to compare degrees
of blame when it is communicated in different ways, but what is needed in order to
block the objection is an argument which establishes that this way of adding up
blame by applying different communicative means is, as a matter of principle, not
possible. And, as just indicated, there is no reason to assume that such an argument
could be presented.
A final more speculative objection to my argument might be that censure
conveyed through hard treatment always communicates more censure then if it is
conveyed in other ways. Thus, though censure and blame expressed in a non-
symbolic and in a symbolic way may function in an additive manner in a case where
the hard treatment inflicted on two perpetrators is the same but where only one of
the two receives an extra symbolically conveyed message, the blame conveyed in
the latter way can never outweigh even the smallest difference in hard treatment
communication. Hard treatment is, so to say, an infinitely more powerful blame-
expressing instrument. However, first of all it is far from obvious that this is actually
correct. But, more importantly, even if it is in fact correct, it would nevertheless not
succeed in blocking the argument. The way the argument was presented, I assumed
that crime C2 was more serious than C1, but that more hard treatment was required to
provide the preventive reason against committing C1 than against committing C2.
However, it would be sufficient to assume that the prudential disincentives to
perform respectively C1 and C2 are the same. In which case, the performance of the
two crimes should according to the theory be responded to with the same amount of
hard treatment, and the extra amount of blame which is required by a proper
response to C2 could then be conveyed in a different way. The punishment for the
two crimes would then be equally severe, though C2 is more serious than C1, which
contradicts the parity condition of proportionalism. Thus, not even this more subtle
answer blocks the argument.
Of course, one might respond by giving up the view that the severity of
punishments should be measured by the amount of hard treatment that is inflicted,
but rather in terms of the blame that is communicated to the punished. But this
would be to give up a central aspect of proportionalism and, as we shall see later,
this is not what von Hirsch and other proportionalists recommend. Thus, in short, it
seems that the two-pronged justification of punishment which von Hirsch and
Narayan have suggested faces a problem when it comes to the justification of
proportionalism.
What the previous considerations together indicate is that the justifications
of proportionalism which have been presented by the expressionist theories of
punishment are not as straightforward as von Hirsch’s proportionality argument
prima facie suggests. Of the three versions of expressionism I have considered, the
first non-teleological variant seemed to have problems in establishing the need for
hard treatment. A weakness which in itself is sufficient to undermine
proportionality. The teleological version defended by Duff, based the justification of
hard treatment on a number of assumptions of human moral psychology which were
not sufficiently unfolded to establish the need for hard treatment. And even if hard
treatment would in fact be required to fulfil the reformative purpose, the theory
faced further problems as a way of sustaining proportionality. Finally, I have
indicated that, though von Hirsch’s and Narayan’s proposal provides a convincing
justification of hard treatment, the step to proportionalism is still vulnerable to
objections. Though much more would have to be said in a complete analysis of the
different versions of expressionism, I believe that the criticisms advanced so far do
cast serious doubt on the validity of the expressionist justification of
proportionalism.
There is a final point worth mentioning which I believe gives further
support to this conclusion, or which at least indicates a lack in the development of
expressionism. Suppose that there is something you find important to communicate
to another person (or perhaps that you are under a certain obligation to
communicate). Suppose further, that after having addressed the potential recipient it
is obvious that he did not understand the message. What would (or should) you do?
The answer is pretty clear. Unless one has certain reasons to believe that the
conveyance is doomed to fail (e.g. if one has discovered that the person speaks
another language or is dead) one would certainly repeat the message. Now, why
should this be any different if we speak with the voice of punishment? If a criminal
fails to understand the censure that we seek to communicate, then why not repeat the
communicative act, i.e. why not punish the criminal again? If this is what
expressionism implies then obviously it will not be possible to maintain
proportionalism. If two persons, A and B, have committed the same crime they
might be punished differently if A at first understands the message while B does not.
The parity-condition is violated. Unsurprisingly, adherents of expressionism reject
that this would be morally acceptable. However, merely to refer to something like a
principle of double punishment, enunciating that one should never be punished for
the same crime more than once, is not satisfactory. There still remains a tension
between claiming, on the one hand, that the purpose of punishment is to convey a
certain message and, on the other, to adhere to a principle of proportionality which
excludes the possibility of repeating the punishment if the conveyance fails. Though
a number of answers to this problem can be imagined none of them are, in my view,
51
fully satisfactory.
Firstly, it might be responded that in actual life the communication simply
does not fail. In contrast to a verbal conveyance of messages, hard treatment is a
language which cannot be ignored or misunderstood. Therefore, there will never be
a situation which invites double punishment. However, this answer rests on an
empirical assumption which I see no reason to accept. Is is really impossible to
imagine someone who regards hard treatment punishment merely as something
which is unfairly inflicted on him; or someone who sees punishment merely as a
price that must be paid for a certain action he has performed; or, perhaps, one who
has been punished so often that he does not even give a thought to why he is once
again put into prison? Even if these examples do not seem convincing, it is
nevertheless a strong view to hold that hard treatment punishment, in contrast to
probably all other sorts of communication, is infallible. There is, however, a related
answer which does not rely on this dubious empirical premise. Rather than claiming
that situations which invite a double punishment will not occur, one might contend
52
that they cannot, for purely conceptual reasons, take place. If the communication
of blame is part of the definition of punishment, then a person would simply not be
punished if the message is not conveyed even though he has undergone a serious
hard treatment. Therefore, even if the hard treatment is repeated the perpetrator is
not punished twice or more. However, it seems that neither of these answers is what
expressionists have in mind. For instance, Nozick admits that a punishment can fail
53
“just as an ordinary act of communication can fail” . Similar claims, on the
fallibility of punishment, are made by other expressionists. This contradicts both the
empirical and the logical rejections of double punishment. Moreover, it would
certainly be a dubious view to hold that a person who has spent several years in
prison, but who has failed to grasp the condemnatory message, has not been
54
punished at all.
Secondly, another way to rebut the claim that expressionism may face a
problem of double punishment, would be to reject the idea that the expressionist aim
is basically communicative. If the purpose is merely to express reprobation of the
wrongful conduct rather than to convey a message to the criminal, then there will be
no ground for repeating the expressive act if it is not understood. To contend that the
expressionist goal has been fulfilled once an expressive act has been performed is
therefore a way to avoid the objection. However, though this reply is not
inconsistent with the claim that one hopes in addition that the expression is
understood by someone, the theory does make it harder to explain why the
expressive act in itself is so important, if the goal is not at least to make some -
the criminal or other members of the community - understand what is expressed.
Moreover, though not all proponents of expressionism are precise on the matter, the
general view clearly is that the purpose is communicative. As mentioned earlier,
Nozick underlines that the goal is to evoke understanding. Von Hirsch, Narayan,
Lucas and others apparently share this view. And Duff explicitly claims that it is
preferable to talk of punishment as communication rather than as expression,
because the idea of communication involves - as that of expression need not -
someone to or with whom one tries to communicate, that is, someone who receives
55
the message. Thus, it seems that expressionists neither should nor would accept
this answer to the problem.
Thirdly, it might be responded that the imposition of double punishment on
some criminals would be inconsistent with treating them as persons or moral agents.
As we have seen, Duff’s view is that, though the justifying aim is reform of the
criminal, it would be wrong to achieve this goal by manipulative means. Despite the
fact that the communicative process is coercive it should not force on the criminal
the desired change. The criminal must be free to choose the opportunity for
repentance and reconciliation which the punishment provides. To continue
punishing a criminal until he repents would, on Duff’s account, count as an
unacceptable attempt at coercive change. Now, could the same answer be given if
we continue punishing someone, not to reform him, but at a prior stage, to make him
understand what he may not have understood at the first communicative act (i.e. the
first punishment)? In short, would this just as well be inconsistent with the
autonomy of the agent? Would not it be to treat him like a tiger?
It is indeed hard to see that this can be plausibly argued. Even if we accept,
that the continued punishment of a criminal who has understood the message but
who will not accept the opportunity of self-reform is an unacceptable coercion, it
certainly does not follow that this is also the case if we repeat a punishment to make
him understand in the first place. On the contrary. If one goes as far as Duff as to
56
hold that the criminal has a right to punishment - that punishment is something we
owe to the criminal - or even if the view is put more modestly, that it is simply of
moral importance to convey a message concerning his wrongful act, it is surely hard
to see why it should be considered wrong to repeat the message if the
communicative endeavour fails at first. And even if one accepted that there would
be an element of unacceptable coercion involved if one went on repeating the
message over and over again if it was not understood by the criminal, it certainly
requires an argument to show that even a single repetition would be unacceptable.
After all, what we would be doing would be to treat the criminal - as specifically
57
prescribed by several expressionists - as a person capable of understanding. But, as
we have seen, this possibility is sufficient to undermine proportionality. Neither
therefore does this third answer seem convincing.
There is a further point to this problem of double punishment which is
worth noticing. Consider a criminal who actually understands that he is being
censured but who does not get the correct message as to the extent to which he is
being so. Is this a possible scenario? In order to defend proportionalism it must be
assumed that the severity of the punishment conveys how much a conduct is
disapproved of (this is stated in premise (2) in von Hirsch’s proportionality
argument). As mentioned it is assumed that the communication with the criminal is
an ongoing process; otherwise there would be no reason in keeping on inflicting
36 JESPER
THE ETHICS RYBERG
hard treatment on a criminal. But is it not possible that a criminal, who is put into
prison and who at the beginning understands that he is condemned by the hard
treatment he undergoes, after a while loses attention to the voice of punishment?
Though he probably looks forward to the day he is released, he has simply become
used to his time in prison and no longer functions as a recipient in a communicative
process. If this is possible or if it is in other ways possible partly to misunderstand
the conveyed message, then it will simply constitute a version of the double
punishment objection. If, of two perpetrators who receive the same time in prison, it
is only one who receives the whole condemnatory message then there would still be
reason to punish the other more in order to convey the same amount of censure on
the two. But once again this would contradict proportionality.
What these considerations highlight is the fact that not much has been done
to make clear exactly how the communicative process actually takes place. To say
like Feinberg that “the very walls of his cell condemn him” is certainly not very
clear. To what extent it is possible to overhear the message or to what extent the
communication may fail and, in this connection, what would count as sufficient
evidence for the fact that a communication has failed, is simply not made clear.
However, as the discussion indicates, these questions are of vital importance for the
expressionist defence of proportionalism.
4. THE FAIRNESS THEORY
The final influential retributivist theory to which we shall now turn is the fairness
theory (or the unfair-advantage theory). In contrast to the theories outlined in the
previous sections, the fairness theory does not consider the justification of
punishment in isolation from a more general theory of distributive justice. On the
contrary, the theory of punishment is often presented as part of a broader view of the
just distribution of benefits and burdens in a society. This means that the
philosophical discussion of the theory has had different focuses. Some theorists have
been concerned with the general distributive principles without paying any or much
interest to the question of punishment, while others have focused specifically on the
58
implications these principles have with regard to punishment. Since the focus in
the present context is, of course, narrowly on the question of punishment, I shall not
here spend time discussing the plausibility of the general distributive principle.
However, a short outline of the underlying principle is required.
In its most general form, the idea on which the fairness theory is grounded
is that, in a cooperative venture which involves costs and benefits of all parties
involved, there should be an equitable distribution of those costs and benefits. The
principle was famously articulated by Hart in his influential essay “Are there any
Natural Rights?” in which it is contended that: “when a number of persons conduct
any joint enterprise according to rules and thus restrict their liberty, those who have
submitted to these restrictions when required have a right to a similar submission
59
from those who have benefited by their submission” . What Hart in his essay
referred to as a “mutuality of restrictions” was later developed by Rawls and others
and acquired to name “the principle of fair play” or “the principle of fairness”.
At first glance the principle seems reasonable. If we think of a cooperative

venture in which all voluntarily agree to participate, the idea that there should be an
equitable distribution of burdens and benefits is intuitively appealing. That some
should gain only the benefits while others carry all the burdens would strike us as
unfair. The interesting question, however, is whether the principle can be applied as
a general principle of social justice. The idea is to consider a society as a
cooperative venture in which the members each enjoy a number of benefits which
are only available due to the cooperation. However, at least part of what creates the
benefits is that the cooperation imposes certain burdens on the members. For
instance, restrictions on each member’s liberty to do whatever he or she wants. The
controversial question has been to ask to what extent can the receipt of benefits
possibly generate an obligation to pay the burdens which the cooperation requires?
Or under what conditions is it justified to adopt coercive measures to prevent what
might be the rational course of action for each individual, namely, to withhold
cooperation - to be free-riding - whenever it is burdensome? As Nozick’s famous
criticism of Hart has shown, it is hardly sufficient to base an obligation not to free-
60
ride on the mere fact that an individual has benefited from others’ cooperation.
However, we can here put this traditional discussion aside and merely assume that
61
the principle of fair play, in one version or another, is plausible. The next
interesting question then is: where exactly does punishment enter the picture?
The fact that it is usually possible to receive the benefits without bearing
the burdens of cooperation is what leads to a discussion of a system of punishment.
However, even if we at the outset accept that benefits and burdens should be equally
distributed it is not immediately clear what role punishment should play in this
connection. An obvious thought would be that one ought to use punishment as an
instrument to prevent future imbalances of benefits and burdens, e.g. by deterring
potential free-riders. However, this would turn the fairness theory into a forward-
oriented theory of punishment and this is not what its proponents have in mind. On
the contrary, punishment is justified as a way of restoring the balance of benefits and
burdens once imbalances have taken place. The claim is that a criminal, on the one
hand, gains a benefit from others’ obedience to the law but, on the other, gains an
extra benefit by not restricting his action as does the law-abiding person. It is this
extra benefit which can be outweighed through punishment. Or, as put by Morris in
his classical modern exposition of the theory: “A person who violates the rules has
something others have - the benefits of the system - but by renouncing what others
have assumed, the burdens of self-restraint, he has acquired an unfair advantage.
Matters are not even until this advantage is in some way erased ....[H]e owes
something to others, for he has something that does not rightfully belong to him.
Justice - that is, punishing such individuals - restores the equilibrium of benefits and
62
burdens by taking from the individual what he owes, that is, exacting the debt.” .
That the fairness theory may have implications with regard to the
distribution of punishment is not surprising. If we hold that punishment is a means
to the end of restoring an equilibrium between benefits and burdens, then we are
surely not far from holding that a specific punishment fulfils its purpose only in so
far as its severity is warranted by the degree to which the equilibrium has been
38 JESPER
THE ETHICS RYBERG
disturbed. In other words, we can see why proponents of the fairness theory
standardly contend that the severity of a punishment should be proportionate to the
gravity of the crime. Roughly outlined, the argument takes the following form:
(1) The seriousness of a crime reflects the magnitude of the resulting disequilibrium
between benefits and burdens.
(2) The more a crime has disturbed the equilibrium between benefits and burdens
the larger is the burden which is required to restore the fair equilibrium.
(3) The severity of a punishment reflects the magnitude of a burden.
(4) Therefore, a crime should be punished more severely the more serious it is.
Though premise (1) may seem pretty obvious if one accepts a fairness theoretical
account of justice, it nevertheless covers a number of controversial questions which
have been the subject of some discussion. The first is whether there are degrees of
unfair benefits. The second concerns the plausibility of the crime scale which
actually follows if determined by degrees of unfair benefits. The answers to both
questions obviously depend on what it is that determines the size of an unfair
benefit; which again depends on what precisely an unfair benefit consists in. That
unfair benefits admit of degrees is something which is generally agreed upon by
adherents of the theory. However, Dagger has given an interpretation which
contradicts this view. According to him, the burden which a law-abiding member of
a cooperation carries is not the burden of obeying a particular law but rather the
burden of obeying the law in general. It is this general burden which the criminal
renounces and which thereby gives him an unfair benefit. What this implies is that
the benefits are the same independently of what sort of crime a perpetrator has
committed. As Dagger says: “the murderer and tax cheater should be punished to the
63
same extent for their crimes of unfairness” . This does not mean that the murderer
and tax cheater should receive that same punishment tout court. Rather, Dagger’s
conclusion is that the fairness theory does not tell us to what extent various
criminals should be punished. And that other theories must be added to the fairness
theory in order to answer this question. However, as indicated, Dagger’s
interpretation of the theory is exceptional. Most adherents emphasize that unfair
benefits do admit of degree. The problems related to this view is what most of the
64
critical discussion of the theory has been concerned with.
Suppose, for instance, that one accepts the interpretation of a burden which,
as indicated in the quotation above, is suggested by Morris, namely, that it is one of
self-restraint. And moreover, that the degree to which the law-abiding needs to
restrict himself depends on the strength of his inclination to commit a particular
crime. As Burgh has pointed out in his influential criticism of the fairness theory,
this implies that “a greater burden is renounced with regard to tax fraud than with
65
respect to murder” since we usually have a much stronger inclination to tax fraud.
In fact, most of us have no inclination at all to murder other people. If this sort of
criticism is correct it would obviously threaten the claim that crimes should be
ranked, and consequently punished, according to the degree of gained unfair
benefits. For the present, I shall not go deeper into this traditional discussion of the
theory, except by mentioning that it certainly seems correct that proponents of the
theory have often not been very precise in their analyses of the concepts of burdens
and benefits. And, consequently, nor with regard to the question of crime scaling.
There is, though, one exception with regard to the latter question. Davis has
concentrated his contribution to the theory specifically on the scaling challenge.
However, since the ranking of crimes is considered in the next chapter, I shall
postpone the discussion of Davis’ suggestion. For the present, it is sufficient to
notice that premise (1) covers a number of controversial questions.
Premise (2) is less controversial. If the balance between benefits and burdens
is upset because a criminal has gained an extra benefit which does not rightfully
belong to him, then it seems that an obvious way to rectify the advantageous position
of the criminal would be by imposing a burden on him. A burden the magnitude of
which corresponds to that of the unfair benefit. A point which has, in this connection,
sometimes been made, is whether the balance of benefits and burdens can actually be
restored through punishment. It has been suggested that a balance may be restored, for
instance, by returning stolen goods or by restituting the victim of a crime, but not by
punishing the perpetrator. Punishing a criminal and restoring a balance of benefits and
66
burdens are, or so it has been claimed, quite different from one another. However,
the objection rests on a misunderstanding of what an unfair benefit consists of. The
fairness theorist would reply that the unfair benefit is not a material good or whatever
else may be the concrete result of the crime, but rather that it consists of shirking a
burden carried by law-abiding members of the society. It is this particular benefit
which is removed by punishing the criminal and not merely by restituting the victim.
In this sense, a crime is regarded as a crime of fairness, whatever else it may be.
Moreover, once we accept the plausible assumption, which links premises (2) and (3),
namely, that a punishment is a burden to the criminal, it also seems reasonable to
accept premise (3), that the severity of the punishment determines the size of the
burden.
As indicated, the premises on which the argument is based - especially
premise (1) - have been the subject of some discussion. However, suppose we ex
hypothesi assume that the premises are in fact plausible. Does this leave us with a
plausible defence of proportionalism? The reason I believe that the question is likely
to be answered in the negative is analogous to the reasons considered in the
discussion of the simple desert theory and the expressionist theory. The problem is
that, even if it is correct that punishment can serve as an appropriate burden, the
question remains as to whether a burden can also be imposed in other ways than
through punishment. Some proponents apparently admit that this is the case, but, in
my view do not acknowledge the full implications of this possibility. Others
apparently reject the possibility, however, without providing convincing reasons to
the effect. For instance, in his defence of the fairness theory, Sadurski considers the
question of whether a punishment should be less severe if a criminal has already
suffered some burdens as a result of his crime. As examples he imagines a thief who
has broken his leg in the course of committing a crime, and a rapist who is caught
during his escape by a member of the victim’s family and is severely beaten. In
Sadurski’s view, the suggestion that these pains suffered by the criminal should
constitute burdens which reduce the overall amount of the benefit acquired through
40 JESPER
THE ETHICS RYBERG
the crime “may well be a correct implication from the ‘balance’ model of
67
punishment” .
However, this raises a question: why should we only consider the suffering
which a criminal has undergone during or in relation to his crime? Why not count in
the suffering of breaking a leg if it happened two days after the criminal committed
his misdeed? Or if it happened two days before the crime was committed? Or even
in his childhood? More generally put, why not simply adopt, as considered in the
discussion of the simple desert theory, a whole life view of burdens and benefits. If
the suffering of breaking a leg constitutes a burden in the relevant sense then it is
surely arbitrary, unless further reasons are provided, to let this burden count only
68
when it occurs in relation to a crime. However, once we allow for this possibility it
clearly undermines proportionalism in the sense in which it is usually understood by
its proponents. If we stick to Sadurski’s example, we can imagine that A and B have
each committed equally serious crimes but that A should be punished less than B
because A broke his leg as a child. This is surely not a consequence which
proportionalists would normally be happy to accept. Thus, it is strange that those
theorists who believe that proportionalism can be grounded on the fairness theory
have not really entered the discussion of this problem. To my knowledge, Sher is the
only adherent of the fairness theory who has tried to meet this challenge. What he
has tried to establish is - and this is probably the only way one could hope to avoid
the objection - that only punishment can provide the sort of burdens which are
required to restore a balance of benefits and burdens. This endeavour deserves a
closer scrutiny.
Sher’s version of the fairness theory differs from the standard versions,
defended by Morris and other theorists, in offering a different interpretation of what
an unfair benefit consists in. At the outset Sher agrees with the criticism made by
Burgh and others that, if an unfair benefit is interpreted as a freedom from, or lack
of, self-restraint, and if this benefit is determined by the strength of the law-
abiding’s inclination to commit a certain illegal action, then we end up with an
extremely counter-intuitive ranking of crimes since “most have a greater inclination
69
to cheat [on income taxes] than they ever have to murder” . Sher believes that this
problem of proportionality is nicely resolved by his interpretation of an unfair
benefit. On this account “a person who acts wrongly does gain a significant measure
of extra liberty: what he gains is freedom from the demands of the prohibition he
violates. Because others take that prohibition seriously, they lack a similar liberty.
And as the strength of the prohibition increases, so too does the freedom from it
which its violation entails. Thus, even if the murderer and the tax evader do
succumb to equally strong impulses, their gains in freedom are far from equal.
Because the murderer evades a prohibition of far greater force - because he thus
‘gets away with more’ - his net gain in freedom remains greater. And for that reason,
70
the amount of punishment he deserves seems greater as well” . In short, Sher
understands the unfair benefit as an extra measure of freedom from moral restraint.
What exactly is meant by a criminal gaining freedom from moral restraint
is not absolutely clear to me; however, for the present we can leave this out of the
71
discussion. What is important is that Sher believes that the interpretation avoids
the objection provided by a whole life view on benefits and burdens or, more
precisely, it avoids the implication that, for instance, the suffering from a broken leg
should affect the severity of a punishment a criminal receives. The reason is that the
criminal’s extra benefit is measured by his act’s degree of wrongness, whereas a
broken leg, or other similar burden, is measured on a scale of suffering (or, as Sher
suggests, a scale of “preference-(dis)satisfaction”). A balancing of an unfair benefit
and this kind of burden, therefore, is impossible, not because they stand in a wrong
temporal relation, but simply because they are incommensurable. On the other hand,
an unfair benefit can be balanced by the imposition of punishment, because what
characterizes punishment is that it is a performance of an ordinarily impermissible
act. And it is exactly this ordinary impermissibility which, on Sher’s account, makes
it a suitable way of restoring the balance that was disturbed by the criminal’s unfair
benefit. Or as Sher himself argues: “By treating the wrongdoer in what is ordinarily
a forbidden way, we strip away part of the protection that moral restraints on our
behaviour would ordinarily have afforded him. Thus, we remove precisely the sort
of advantage he has gained. Because the resulting disadvantage can be assessed in
terms of its usual moral wrongness, it can be weighed on the same scale as the
wrongdoer’s unfair advantage. Thus, it is commensurable with the wrongdoer’s
72
extra benefit as his previous hardships are not” .
The strength of Sher’s suggestion is that it succeeds in explaining why a
broken leg or suffering caused by a disease, by a natural catastrophe or other related
cause, does not count as a burden, at least not in the sense required to outweigh an
unfair benefit. However, it meets the challenge to proportionalism only by facing a
new challenge. As we have seen, something can count as a burden only if it is
measurable on a scale of moral wrongness. But it is certainly not hard to imagine a
criminal who has been wronged previously in his life. In other words, rather than
asking whether previous suffering caused by a broken leg can offset a criminal’s
unfair benefit, we can rephrase the question by asking whether the fact that the
criminal has previously been wronged by someone who intentionally broke his leg
in order to stop him at the football pitch can offset the criminal’s unfair benefit? Or
more generally put, whether the unfair benefit Y gains from wronging Z is
outweighed if Y has previously been wronged by X? If this is answered in the
affirmative it obviously constitutes a challenge to proportionalism. Once again, we
can imagine two persons who have committed equally serious crimes but who,
according to the theory, should be differently punished. Or even a situation in which
a person has committed a more serious crime than another person, but should
nevertheless be punished less severely.
Sher is actually aware of this challenge but believes that he is once again
able to resolve it. However, at this point in his reasoning he is definitely hard to
follow. What he contends is that: “even if X has previously wronged Y, it hardly
follows that a fair balance of benefits and burdens is restored when Y in turn wrongs
Z. If Y does this, then the original wrongdoer X is still left with the double benefit of
moral restraint upon others plus his own freedom from such restraint; and the
current victim Z is left with the double burden of moral restraint on his acts plus the
absence of restraint on the acts of (some) others. Thus, the original unfairness is not
73
removed but merely displaced” . The claim that the original fair balance of burdens
and benefits is not restored in this course of actions is apparently correct. However, it
does not justify Sher’s conclusion that “Y’s extra benefit in wronging the innocent Z
74
does not offset the extra burden that X’s earlier wrongdoing inflicted on Y” . In so far
as Sher’s theory is that a wronging is commensurable with an unfair benefit and counts
as a burden in the relevant sense, it follows that Y has first received an extra burden,
by being wronged by X, and later an extra benefit, by wronging Z; that is, at the end he
is neither in a particular advantageous or burdensome position. To punish Y would
simply be to inflict an extra unfair burden on him. Thus, Sher’s argument is a non
sequitur. Which means that the challenge to proportionalism has not been resolved.
In short, there seem to be three possible answers as to whether Y should be
punished for wronging Z when Y himself has previously been wronged by X. Either
one could claim that Y should be punished because he in the end possesses an extra
unfair benefit. But as we have seen, this possibility is excluded by Sher’s own
interpretation of a burden. Or one could draw the conclusion that Y should not be
punished. For many this is probably not an acceptable solution and, as mentioned,
Sher himself rejects it. And, what is more important, it is tantamount to giving up
proportionalism. The final possibility then is to contend that Y should be punished,
however, for reasons which go beyond the distribution of burdens and benefits. In
fact, Kershnar has suggested a reconstruction of Sher’s view according to which there
75
are two possible justifications for punishment. Either punishment is justified
because it offsets an unfair benefit. Or it is justified simply because a wrongdoer has
violated a moral norm, even if he in the end does not possess an excess benefit. On
this interpretation it would be possible for Sher to maintain his claim that Y should be
punished, even though it is not warranted by the calculation of benefits and burdens.
However, this position has its own problems.
To see this, it should be mentioned that Sher apparently believes that if X
wrongs someone and is then later wronged by another then X does not deserve
further punishment, because the way he was wronged counts as a punishment and
76
thus leaves him with no excess benefit. However, when this view is combined with
Sher’s claim that Y should be punished when he wrongs Z, even if he has himself
previously been wronged, there arise a problem in cases where we have, what
Kershnar calls, a “victim/victimizer circle”. Suppose we have the somewhat unusual
situation that X wrongs Y, then Y wrongs Z, and finally Z wrongs X. And suppose
that the wrongings are all of the same sort. Now, as just mentioned, Sher’s view
apparently is that X should not be punished for wronging Y because he has been
punished by being wronged by Z. On the other hand, Y should - according to Sher -
be punished because he, on the suggested interpretation, has violated a moral norm.
Thus, in sum, X should be punished while Y should not be punished even though
they have committed exactly the same wrongings and have been wronged in exactly
the same way. I am far from certain that Sher would accept Kershnar’s
reconstruction of his position. But if he should it would not help much. Kershnar
believes the problem generated by the victim/victimizer circle shows that the whole
position is implausible. But, even if one does not draw this conclusion, it does at
least contradict the parity-condition and thus proportionalism. Thus, the challenge to
proportionalism remains unsolved.
There is a final and more general comment worth making in relation to
Sher’s theory. As we have seen, Sher believes that his theory succeeds in avoiding
Burgh’s and others’ criticism, that the central male in se crimes, such as murder or
assault, should be punished less severely, or perhaps not be punished at all, than
male prohibita crimes such as tax evasion. However, by suggesting that an unfair
benefit consists in an extra freedom from moral restraint, it seems that Sher faces the
opposite problem, namely, that mere male prohibita crimes which would not be
wrong in the absence of a legal prohibition, should not be punished since they do not
77
impose a burden in the form of a moral restraint on the law-abiding. But this is an
implication which is hard to accept. There is, of course, one way to avoid this
objection. This would be by turning all male prohibita crimes into male in se crimes
by contending that it is morally wrong to break the law. In which case, tax evasion
would be morally wrong. However, this has the unfortunate implication that all male
prohibita crimes turn out to violate the same moral prohibition, namely, that one
should not break the law. But this means that all male prohibita crimes morally
restrain the law-abiding to the same degree, and that such crimes should therefore all
be punished equally severely. Strictly speaking, this does not violate
proportionalism because if all male prohibita crimes are equally serious the
principle implies that they should be equally punished. But it certainly questions the
view underlying premise (1) in the proportionality argument, namely, that crimes
should be ranked in seriousness according to the magnitude of the unfair benefit a
criminal gains from committing them.
Summing up, considerations of the fairness theory amount to the following.
As we have seen, the theory bases its defence of proportionalism on the assumptions
that the fair equilibrium of burdens and benefits is disturbed more the more serious
the crime, and that it consequently requires the imposition of a heavier burden on the
criminal in order to restore the initial balance. However, even if we accept the
plausible assumption that punishment does constitute a burden, the theory does not
succeed in justifying proportionalism. The problem is that it seems very hard to
avoid the claim that a person can face burdens in many other ways than through
punishment. But this means that other burdens which a criminal may have
undergone, even long before he committed his crime, will influence the equilibrium,
and thus affect the seriousness of the punishment he should receive. An implication
of this is that it is possible to imagine a person who has committed a much more
serious crime than another person, but who is nevertheless, due to experiences in
earlier parts of his life, punished less severely. This contradicts proportionalism. I
then considered Sher’s version of the fairness theory in some detail because he
78
contends that his interpretation is able to “resolve the problem of proportionality” .
However, as I have argued, this seems not to be the case. Thus, modestly put, I do
not think that proportionalism follows as easily from the fairness theory as its
adherents usually proclaim.
5. A NON-FOUNDATIONALIST APPROACH
The punishment theories which have been discussed in the previous two sections
can, I believe, rightly be regarded as the dominant positions in modern retributivist
thinking and thereby also as the most important approaches to a justification of the
proportionality principle. Even though the theories are, as we have seen, very
different - the former focuses on the conveyance of appropriate condemnatory
messages to responsible moral agents, while the latter points at the significance of
restoring a fair balance of benefits and burdens - they are, along with the simple
desert theory, similar in one important respect: all seek to provide a foundationalist
justification of proportionalism by recurring to more basic theories of the
justification of punishment. Even though this sort of reasoning is certainly
expectable when it comes to an issue as specific as punishment distribution, it does
not exhaust the range of justificatory options. Adherents of proportionalism
sometimes seem to find support for their viewpoints in a non-foundationalist
manner. In order to examine what this sort of justification may consist in we shall
have to tentatively depict the contours of the problem which has attracted most
attention in the modern philosophical debate on punishment and which has naturally
constituted the core of the show-down between the retributivist and the utilitarian
approaches to punishment, namely, the problem of the punishment of innocents.
That a terrible wrong is done if a person is punished for a crime that he or
she has not committed is something upon all people will usually agree. Voltaire’s
claim that “it is better to run the risk of sparing the guilty than to condemn the
79
innocent” or Blackstone’s almost contemporary consent to the view that “it is
80
better that ten guilty persons escape, than that one innocent suffer” are both
expressions of this frequently stated conviction. It is the prima facie plausibility of
this view which has formed that background of what has often been presented as a
devastating objection against the utilitarian theory of punishment and against
utilitarianism in general.
The first to point at this sort of objection, though without explicitly
mentioning the innocent, was apparently Kant who warned each man against
creeping through the “serpent-windings of utilitarianism to discover some advantage
that may discharge him from the justice of punishment, or even from the due
measure of it, according to the Pharisaic maxim: ‘It is better that one man should die
than the whole people should perish’.” For, as he puts it: “if justice and
81
righteousness perish, human life would no longer have any value in the world” .
One of retributivist theorists who in recent time have done most to emphasize the
importance of the objection is McCloskey. In several articles, written in the
predominantly utilitarian-oriented 60’s, McCloskey presented the objection in
slightly different versions which have become standard formulations of the
82
argument. McCloskey’s well-known example is the following.
Suppose that in a town with a mixed population a black man has raped a
white woman. Because of existing racial tensions it is reasonable to assume that the
crime will produce serious racial violence with many people killed and injured,
unless the rapist is apprehended. Suppose further, that the sheriff of the town can
prevent the violence only by framing an innocent black man who was seen close to
the place where the crime happened, and who would be believed to be guilty. Given
the possible alternatives, it seems that utilitarianism would imply that the sheriff
should frame and punish the innocent. However, this is exactly what McCloskey and
other critics regard as an unacceptable implication. Punishing the innocent would be
wrong.
Naturally the objection need not be formulated in terms of scapegoating.
The same problem can in principle be illustrated in cases involving deterrence or
incapacitation. What the objection shows is not - as is sometimes prematurely
contended - that there is, for the utilitarian, no connection between who is punished
and who is guilty. Usually there might well be such a connection (e.g. for a
utilitarian of the deterrent school it would normally be vital to maintain the
connection between punishment and crime in order not to lose the relation between
83
punishment and deterring the committing of crimes). But the fact which the
example points at is that this connection is not necessary. Now, what is important
here is not merely that the argument is supposed to constitute a reductio ad
absurdum of utilitarianism but rather that it is sometimes perceived as a key
argument in the dispute between utilitarians and retributivists exactly because it is
thought that only a retributive approach to punishment supplies one with grounds for
84
the stipulation that punishment must only be applied to the guilty. In that sense the
punishment-of-the-innocent argument indirectly supports retributivism.
Now, if there is something to this line of reasoning, could an indirect
argument along the same lines be then constructed in favour of proportionalism? At
first sight it is certainly tempting to answer this question in the affirmative. After all,
punishment of the innocent might reasonably be regarded as an instance of
disproportionate punishment. To this it might perhaps be objected that the fact that it
is unacceptable to punish the innocent person in McClosky’s example does not
commit one to proportionalism. As Hart noted in his discussion of punishment
distribution, “if in answer to his question [“Who may be punished?”] we say ‘only
an offender for an offence’ this admission of retribution in distribution is not a
principle from which anything follows as to the severity or amount of punishment
85
...” . In other words, it might be held that only the guilty should be punished while
at the same time defending a non-proportionalist approach to the “how much”
question. As a matter of fact, such a position has been defended by a few desert
86
theorists. It is not obvious, however, that such a position is persuasive. But even if
there is something to this objection it could easily be met by the proportionalist by
incorporating slight changes in the example from which the discussion takes its
departure. Rather than assuming, as does McCloskey in his example, that the person
who was seen close to the place of the crime was innocent, we might instead assume
that what he did at that place was perpetrate a minor crime, say a theft, but that
punishing him for the rape would still have the effects on others’ behaviour that
McCloskey imagines. In so far as it would still be unacceptable to punish the person
for the rape, the example would still constitute a reductio ad absurdum of
utilitarianism. And the further step, to perceive it as an indirect argument for
proportionalism, would no longer be blocked by a position which is retributive only

with regard to the “who may be punished?” question.
In fact, McCloskey himself indicates that such a modified version of the
argument would equally well constitute a reductio ad absurdum as the argument
based on the traditional example. That the force of the argument is not changed
despite the modification is also suggested, for instance, by Gross who emphatically
contends that “any punishment in excess of what is deserved for the criminal
87
conduct is punishment without guilt” . Thus, in sum, what we have is an argument
consisting of two steps: firstly, a step which seeks to establish that the utilitarian
theory of punishment is unacceptable and, secondly, a step which proclaims that this
provides support for proportionalism. I believe that if some people are reluctant to
abandon proportionality, even if the sort of foundationalist justifications earlier
considered cannot be provided, this reluctance might well be grounded on the belief
that it will have unacceptable implications to give up proportionality, that is, on
some sort of indirect argument as the one here outlines. Thus, let us consider each of
the argument’s two steps seriatim.
Does the argument constitute a genuine reductio argument? That is, should
we reject utilitarianism on the ground that it would imply that a person who has
committed only a minor theft should nevertheless be punished for a rape, if this
would prevent the described riots and lynchings? For the utilitarian who does not
accept that the traditional argument, involving punishment of the genuinely
innocent, gives us reason to abandon or at least modify his position, the answer has
usually taken two different forms which, of course, are equally relevant in relation to
the modified version of the example. The first approach has been to reject the
88
reduction’s being in the end absurd (the “outsmart” response) . Considering
McCloskey’s original example, the answer would be that though it is usually terrible
to punish an innocent person, the alternative, to allow a large number of people to be
killed in lynchings, is even more terrible. Therefore, though one should certainly
hope that the situation that is envisaged would never actually occur, the right answer
would nevertheless be to punish the innocent. This is the answer which was
89
suggested by Smart. The same answer could be given with regard to the modified
example, namely, that on reflection the excessive punishment of the thief is the least
unattractive alternative.
The second approach which utilitarians have traditionally adopted to
McCloskey’s argument has been to claim that the conclusion simply does not follow
from utilitarianism. More precisely, it has been suggested that in the real world the
incrimination of the innocent man in order to prevent riots is not what utilitarianism
would prescribe. Once one resorts to punishing an innocent person there is always
the risk that this will be found out. And if such punishment actually came to light it
would undermine all trust and respect for the law. Given the terrible consequences
which such a breakdown in confidence to the legal system would have, the chance
that it might occur would outweigh the good that might be obtained by preventing
90
the lynchings. This line of answer might, of course, also be given in relation to our
modified version of the reductio argument. Retributivists have typically responded
to this answer either by claiming that, since it might ex hypothesi be assumed that it
would never be discovered that the punished person was innocent, the answer does
not take the example seriously; or by constructing alternative examples in which it
seems more obvious that utilitarianism does imply that an innocent should be
91
punished. It has also been responded that, even if it is correct that utilitarianism
would not in the example prescribe that the innocent be punished, there must be
some cases in which this sort of victimization is what utilitarianism dictates. For
instance, McCloskey remarks that all that needs to be indicated is the “logical
92
possibility” of such an unjust system of punishment. However, in my view this
latter comment clearly weakens the argument. Obviously it is correct that there
might be situations in which utilitarianism would prescribe punishment of an
innocent. But it is far from obvious that all cases where this would be so would
strike us as equally absurd. In relation to our modified version of the argument, there
are certain cases where disproportionate punishment does not, or at least so I
believe, seem gravely counter-intuitive. Suppose that the only way one could save a
large number of persons from a terrible death was to punish a criminal one day more
in prison than he or she has deserved given the seriousness of the crime committed.
Or that the terrible outcome could be avoided only be requiring from a criminal a
fine which is slightly larger than the one deserved. To hold that this would be clearly
absurd does not strike me as convincing. Thus, merely to point at the logical
possibility is hardly sufficient. What is required to challenge utilitarianism is that it
can be show that it has unacceptable implications. This is why examples are needed
and why the discussion of them is important.
Whether such convincing examples can be provided or, more specifically,
whether the example of the disproportionate punishment of the thief succeeds in
establishing the absurdity of utilitarianism, however, is not a matter I shall pursue
any further. The primary purpose here is not to consider the plausibility of
utilitarianism. Thus, though utilitarians might respond to the argument in the
outlined ways and though I must admit that I do not regard the objection as being as
forceful as it is often assumed, we can here for the sake of the argument assume that
the argument really does constitute a genuine reductio ad absurdum of
utilitarianism. Even with this assumption there is still an important step missing in
order to reach a conclusion concerning proportionality. The question is, does a
rejection of the utilitarian view on punishment distribution establish the plausibility
of the rival proportionalist position? In my view, there are several reasons to be
sceptical with regard to this second step of the argument.
A first thing that should be noted is that in order for the example to give
any support to proportionalism at all, it must, of course, be assumed that the
proportionality view itself does not allow for the punishment which is imposed on
the thief. Even if one regards the punishment in the example as unacceptable, it is
not necessarily clear what precisely it is about it that is counter-intuitive. On one
interpretation, what may strike one as hard to accept is that a person is punished
very hard for having committed a minor crime. The reaction will be the same as the
one we might show when we are informed how people, in earlier centuries, treated
minor criminals, even if we are not told anything about how more serious crimes
93
were responded to. That is, on this interpretation what is at stake is a judgement of
cardinal proportionality: one should not allot that punishment for that crime, period.
But does proportionality proscribe that a thief receives a severe punishment? A mere
ordinal concept of proportionality obviously does not. The answer to this question,
therefore, depends on how proportionalists believe that scales of crimes and scales
of punishments should be connected. However, as I shall argue in a later chapter,
neither of the proportionalist answers to the anchor problem have succeeded in
identifying what should be regarded as the appropriate punishments for different
crimes, nor in setting clear limits as to what would count as acceptable punishments
for particular crimes. Thus, if what is unacceptable in the example is the cardinal
aspect of the punishment then, in order to support proportionalism it must be
presupposed that the principle excludes such punishments. And this is exactly what,
in the suggested (but not yet argued) absence of a satisfactory anchoring theory, has
94
not been shown.
Be that as it may, there is another more important problem concerning the
sufficiency of the argument as an indirect route to proportionality. In order to
establish the truth of the proportionality view it is not only necessary to show that
the view does not itself allow for the unacceptable punishment to which the example
draws attention. It is also necessary to show that this view is the only position which
is inconsistent with this punishment. But such is not the case. For instance, the tough
punishment of the thief might just as well be objected to on the ground of a radical
abolitionist position regarding all punishments as wrong. Or on the ground of a
principle which I shall refer to (and return to in the final chapter) as “negative”
proportionalism, according to which all that matters is that a perpetrator does not
receive a punishment that is more severe than the one that is proportionate to the
gravity of the crime. Therefore, even though one accepts the rejection of
utilitarianism, one is not forced to accept proportionality. The belief that there might
be an indirect way of establishing proportionality rests, to put it a little more
formally, in a confusion of contradictions: the argument, to go through, requires that
what is rejected is the contradictory opposite to proportionality; which obviously
utilitarianism is not.
There are two ways in which those who nevertheless feel that there is
something to the idea of an indirect way of justifying proportionality might reply.
One answer would be to add further arguments, besides the punishment-of-the-
innocent argument, with the intention of also showing the absurdity of other rival
approaches to punishment distribution. This is exactly what some theorists have
done in a more general defence of retributivism. For instance, after having directed
attention to the abhorrence of the “penal suffering of the innocent”, Moberly
proceeds with his defence of retributivism by setting forward a number of arguments
in favour of the view that the guilty should be chastised. In his view, the “deep-
seated sense of fairness which revolts against punishment of the innocent revolts
also against any treatment of the guilty which appears to confound guilt and
95
innocence” . However, even if one accepts Moberly’s appeal to the unfairness of
cases where a person’s grievous fault makes no difference to the treatment he or
she receives, and that such intuitions constitute an argument against the
abolitionist view, it is still obvious that this will not do with regard to establishing
proportionality. There would still be other views which would be consistent with
the rejection of punishing the innocent and not punishing the guilty (e.g. that one
should punish only those who are guilty but all with equally severe punishments, or
some sort of compromise position which combines utilitarian and retributive
96
ideals ). In other words, the problem with this answer is that in order to establish
proportionality indirectly one will have to show that all other views on punishment
distribution are unacceptable. To my knowledge, no one has tried to present such an
argument, and it is certainly not easy so see how it should be done.
Another answer might be to admit that an indirect proof of proportionality
is not likely to be provided, but to maintain that this does not imply that the above-
mentioned cases involving punishment and other corresponding cases do not support
proportionality. After all, strict indirect proofs in favour of a certain position are not
the sort of arguments which are usually found in discussions of ethical theory. But,
even if such a strict argument cannot be provided, it might still be possible to
support proportionality by reasoning which is non-foundationalist. This is what
97
Moore has suggested. His interest is not specifically proportionality but more
broadly the justification of retributivism. Moore bases his defence of retributivism
on examples which indicate the absurdity of punishing innocents and on examples -
like the case of Steven Judy who raped and murdered a woman, drowned her three
children, and afterwards said that he had not been “losing any sleep” over his crimes
- which strongly appeal to the unfairness of not punishing the guilty. His suggestion
is that the sort of justification we should be looking for is coherentist, that is, that we
can justify a moral principle by showing that “it best accounts for our more
particular judgements that we believe to be true”; and that our judgements in
examples concerning punishment of the innocent and lack of punishment of the
guilty are best accounted for in terms of a principle of punishment as just deserts.
Have we here got a track to a justification of proportionality which is not
foundationalist and which does not suffer from the same problems as the more rigid
indirect proof approach? I shall not here provide a definite answer but rather point to
a few facts which I believe may cast doubt on this proposal.
The first things that is worth noticing is that, even if Moore is right in
holding that the particular judgements which we believe to be true are most
coherently explained in desert terms, that is, that retributivism is in this way
justified, this is not tantamount to claiming that proportionalism is thereby justified.
As we have seen in the previous sections, some of the dominant retributivist
positions claim that a perpetrator deserves to suffer, to be condemned, or to be
inflicted with a burden but, as argued, these views did not on closer examination
lead to proportionality even though they would provide answers corresponding with
our judgements in the examples on which Moore’s argument relies. Briefly, if what
is justified in coherentist terms is a retributivist position akin to those positions
discussed above, then the distribution principle which follows is not proportionalist.
This is precisely what we have learned so far. Now, a possible answer to this
problem might of course be to contend that what we are concerned with here is not
the justification of retributivism but rather of the proportionality principle.
Therefore, one might ask whether proportionalism provides the most coherent view
on punishment distribution?
A justification along these lines would have to take several considerations
into account. Since the examples considered by McCloskey, Moberly and Moore
could also be accounted for by other than strict proportionalist distribution patterns,
further examples and judgement would have to be taken into consideration. It would
also be necessary to consider the strength of cases which speak against a
proportionalist distribution such as, for instance, situations where a slight departure
from strict proportionality would be of importance to the convict, in that this would
have an effect on his potential future criminal career, or where a departure would
have a significant effect on other people’s interests. Finally, there is also the
question concerning what kind of judgements should be included in a coherentist
justification. Since coherentist methodology is a complex matter there is no simple
answer to this question. However, at least on some accounts it would matter that
principles were mutually supportive and jointly illuminated by the moral concepts to
which we were appealing and, obviously, that they would manage to provide
answers to the problems to which we needed an answer. Whether the proportionality
principle would satisfy such requirements may be doubted if it is correct that a
theory of just deserts would not lead to proportionality. Moreover, some of the more
detailed problems which are illuminated in the following chapters may also cast
doubt on the possible success of this sort of justification. However, it should finally
be mentioned that whether a coherentist justification will support proportionalism is,
at best, a question to be left open since no one, to my knowledge, has endeavoured
to provide this sort of justification. Thus, for the present there is not much which can
be extracted from the idea of a non-foundationalist justification of the
proportionality principle.
6. CONCLUSION
Besides the introductory indication of what adherents of retributivism standardly

mean by the claim that the severity of a punishment should reflect the seriousness of
the crime committed, the discussion in this chapter has focused on the arguments
which different theories have provided in favour of proportionalism. Even if the
discussion may not necessarily exhaust the list of variants of retributivism, what has
been considered have certainly been the most influential theories. Especially have
the different versions of the expressionist theory and the fairness theory played
dominant roles in the recent discussion on punishment. Moreover, it is in these
theories that one finds the most explicit attempts to support proportionalism. It is in
this connection worth underlining that the arguments with which I have challenged
the attempts to defend proportionalism may constitute general challenges to theories
which subscribe to proportionate punishment.
As is clear from the discussion, the purpose has not been to consider
critically each of the basic theories on which proportionalism is grounded. Thus,
except for a few critical comments, I have not considered whether suffering can ever
itself be valuable, whether it is plausible to contend that condemnation of
wrongdoers is of basic ethical importance, or whether the fairness theory constitutes

an acceptable theory of distributive justice. Entering these discussions would require
far more extensive analyses. Rather, the procedure has been to assume the
plausibility of the theories and then consider to what extent they each succeed in
supporting proportionality. This gave rise to two general questions. The first,
whether punishment is sufficient to fulfil whatever is regarded as morally significant
according to a theory. For instance, as we have seen in relation to the expressionist
theory, if punishment is - at least not always - sufficient to fulfil the communicative
aim, that is, if a person could be punished without grasping the relevant message,
then this might lead to a problem of double punishment, which obviously contradicts
proportionalism. The second, and more interesting, question is whether punishment
is necessary with regard to whatever a theory prescribes. As we have seen with
regard to the simple desert theory, the suffering of wrongdoers could be caused in
many other ways than through punishment. The same was the case with regard to
the burdens considered in the fairness theory. However, in so far as punishment is
not necessary, the proportionality requirement would obviously be challenged. And
even if punishment to some extent is necessary to fulfil the aim of some of the more
refined versions of expressionism it would still not be necessary to the extent
required to maintain proportionality. Thus, both the inability to establish sufficiency
and the necessity of punishment may cause problems with regard to the maintenance
of proportionality and, therefore, constitute a general challenge that must be met by
any theory which proclaims that crime and punishment should be thus related. But
with regard to the theories under discussion, the challenges gave reasons to be
sceptical with regard to the justification of proportionalism.
Proportionalists can respond to the criticism in different ways. One might
perhaps suggest that, despite the objections, proportionality can be maintained as
long as certain conditions are satisfied. For instance, when adherents of the fairness
theory contend that punishment is a means to the end of restoring a fair balance of
benefits and burdens, they have to presuppose that the pre-crime balance was in fact
98
fair. Therefore, to object that the fairness theory cannot maintain proportionality
because a criminal may have experienced severe burdens in his life prior to the
crime, is simply to ignore the condition on which the theory bases proportionality,
namely, that the society from the outset is reasonably just. In similar ways it might
perhaps be possible, for the other theories considered, to condition the adherence to
proportionality in ways that would make it possible to avoid the objection rising
from the failure establishing the necessity of punishment. The other way one might
respond, could be to take another revisionary step by claiming that the failure to
establish the necessity of punishment does not at all undermine proportionality. If
one accepts the idea of a poena naturalis, or natural punishment, then one could
maintain proportionality by claiming that part of the punishment which a criminal
deserves may well have been imposed in other ways than through state punishment,
that is, for instance by an accidentally broken leg or through other kinds of
99
sufferings or burdens which he or she may have experienced during life. Thus, the
task of the punishment system would be to impose the amount of punishment that
remains, if anything remains, to give each criminal what he or she deserves.
What both of these responses do is to abandon proportionality as a simple

principle which in practical life prescribes guidelines for a state punishment system.
The problem with the first answer would be that in real life the condition on which
proportionality is based would hardly ever be satisfied. For instance, most people
have certainly experienced the suffering of burdens of varying degrees in their lives
prior to a possible crime. Thus, this would simply leave the question as to what one
in real life should do if the condition is not fulfilled. If this is answered by
advocating something like the second response, namely, by invoking the concept of
a poena naturalis, then this would imply that one could easily imagine that two
persons have committed the same crime but that they should, due to difference in
what they have experienced in their lives, be punished very differently in terms of
severity. This concept of proportionality would no longer provide a base for a
sentencing system of the kind which proportionalists have typically recommended.
On the contrary, in order to determine the punishment which each criminal deserves
one would have to include considerations of the criminal’s past and future lives.
Thus, the advantages of simplicity which in contrast to, for instance, a system based
on rehabilitationism are usually emphasized as a consequence of proportionate
punishment would certainly be lost. In fact, one might wonder whether it would ever
be possible to mete out a just punishment. Thus, though this revised concept of
proportionality would be consistent with the different theories it is surely not the
concept of proportionality which its adherents typically have in mind. However,
even such a modified idea of proportionality would not manage to avoid many of the
more detailed challenges to proportionalism to which we shall now turn.
NOTES
1J. Bentham, Principles of Penal Law, Works edited by J. Bowring, William Trait, Edinburgh, p. 399.
2See, for instance, M. Quinton, “On Punishment”, and other articles collected in H. B. Acton, The
Philosophy of Punishment, St. Martin’s Press, Great Britain, 1969.
3Moreover, even if it is correct that one can only be punished for a crime (or misdeed) this does not
necessarily imply the kind of proportionality which retributivists usually defend.
4J. G. Cottingham, “Varieties of Retribution”, Philosophical Quarterly, vol. 29, 1979.
5N. Walker, “Even More Varieties of Retribution”, Philosophy, vol. 74, 1999.
6J. Bentham, The Principles of Morals and Legislation, Prometheus Books, 1988, pp. 181-2 note 3.
7See, for instance, the discussions in C. L. Ten, Crime, Guilt, and Punishment, Clarendon Press, Oxford,
1987, pp. 141-46; N. Walker, Why Punish?, Oxford University Press, Great Britain, 1991, chapter 12 VIII;
or M. Bagaric, Punishment & Sentencing, Cavendish Publishing, Great Britain, 2001, pp. 184-87.
8See, for instance, A. von Hirsch, Past and Future Crimes, Rutgers University Press, United States of
America, 1985, p. 32.
9Of course, there might still be room left for trade-offs, in the sense that the proportionality constraint may
be a part of a threshold-deontological position. But, given the very high thresholds which deontologists
usually advocate (e.g. that constraints may be overridden only if this is the way to avoid genuine
catastrophes), the position still differs much in content from the weighing up approach which
consequentialists adopt.
10These terms were introduced by von Hirsch. Since they have become standard terms I shall use them
henceforth. See, for instance, A. von Hirsch, “Proportionality in the Philosophy of Punishment: From “Why
Punish?” to “How Much?”, Israel Law Review, vol. 25, 1991; or Censure and Sanctions, Clarendon Press,
Oxford, 1993, chp. 2.
11D. J. Galligan, “The Return of Retribution in Penal Theory”, in C. F. H. Tapper (ed.), Crime, Proof and
Punishment, Butterworth & Co., Great Britain, 1981, p. 165.
12See J. Cottingham, “Varieties of Retribution”, Philosophical Quarterly, vol. 29, 1979, p. 239.
13D. Dolinko, “Some Thoughts about Retributivism”, Ethics, vol. 101, 1991, p. 541-2.
14See, for instance, J. Kleinig, Punishment and Desert, Martinus Nijhoff, The Hague, 1973, p. 55; or D. E.
Sheid, “Constructing a Theory of Punishment, Desert, and the Distribution of Punishments”, The Canadian
Journal of Law and Jurisprudence, vol 10, no. 2, 1997, p, 456ff.
15C. W. K. Mundle, “Punishment and Desert”, in H. B. Acton (ed.), The Philosophy of Punishment, St.
Martin’s Press, Great Britain, 1969; L. H. Davis, “They Deserve to Suffer, Analysis, vol. 32, 1971-2; J.
Kleinig, Punishment and Desert, Martinus Nijhoff, The Hague, 1973.
16See, for instance, J. Rachels, “Punishment and Desert”, in H. LaFollette (ed.), Ethics in Practice,
Blackwell, Oxford, 1997, p. 473f.
17J. Kleinig, Punishment and Desert, Martinus Nijhoff, The Hague, 1973, p. 67.
18Ibid.
19L. H. Davis, “They Deserve to Suffer”, Analysis, vol. 32, 1971-2. However, talking about suffering as
being “intrinsically good” might well be intrepreted as a consequentialist view. See D. Dolinko,
“Retributivism, Consequentialism, and the Intrinsic Goodness of Punishment”, Law and Philosophy, vol. 16,
1997.
20For an instructive discussion on the deontic implications of desert claims see, for instance, D. Husak,
“Why Punish the Deserving?”, Nous, vol. 26, 1992.
21See, for instance, the discussion in C. W. K. Mundle, “Punishment and Desert”, in H. B. Acton, The
Philosophy of Punishment, St. Martin’s Press, Great Britain, 1969, p. 71ff.
22See D. N. Husak, “Already Punished Enough”, Philosophical Topics, vol. 18, no.1, 1990.
23G. Ezorsky, “The Ethics of Punishment”, Introduction to Philosophical Perspectives on Punishment,
State University of New York Press, Albany, 1972.
24See, for instance, D. E. Scheid, “Constructing a Theory of Punishment, Desert, and the Distribution of
Punishments”, The Canadian Journal of Law and Jurisprudence, vol. 10, no.2, 1997, p. 459.
25For a discussion, see W. A. Parent, “The Whole Life View of Criminal Desert”, Ethics, vol. 86, 1975.
26It is possible to imagine a theory which maintains the view that the suffering of wrongdoers is of
intrinsic value, but which at the same time avoids the problem of non-punitive suffering, by insisting that
there are other intrinsic values which can only be reached through punishment and not by other causes of
suffering, e.g. a disease. However, as far as I am concerned, no one has defended this sort of theory. And
it seems to me that the theory will face other problems when it comes to a justification of
proportionalism. Moreover, it is certainly not easy to imagine what this value, which can only be realized
through punishment, should consist of.
27See, for instance, R. A. Duff, Trials & Punishments, Cambridge University Press, Great Britain, 1986.
J. R. Lucas, On Justice, Clarendon Press, Oxford, 1980. A. von Hirsch, Censure and Sanctions,
Clarendon Press, Oxford, 1993. I. Primoratz, “Punishment as Language”, Philosophy, vol. 64, 1989. A. J.
Skillen, “How to Say Things with Walls”, Philosophy, vol. 55, 1980. J. Hampton, “An Expressive Theory
of Retribution”, in W. Cragg, Retributivism and its Critics, Franz Steiner Verlag, Stuttgart, 1992. R.
Nozick, Philosophical Explanations, Harward University Press, Cambridge, 1981. U. Narayan,
“Appropriate Responses and Preventive Benefits: Justifying Censure and Hard Treatment in Legal
Punishment, Oxford Journal of Legal Studies, vol. 13 no. 2, 1993.
28J. Feinberg, “The Expressive Function of Punishment”, Doing and Deserving, Princeton University
Press, 1970, p. 98.
29See A. Duff & D. Garland, Punishment, Oxford University Press, United States, 1994, p. 218.
30J. R. Lucas, On Justice, Clarendon Press, Oxford, 1980, p. 132.
31R. Nozick, Philosophical Explanations, Harvard University Press, Cambridge, 1981, p. 377ff.
32A. von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993, p. 10.
33Ibid. p. 15.
34The tension clearly is not resolved by adopting the view that the purpose is merely to express
condemnation. In fact, I believe that this view makes it even more obscure why hard treatment is
required, than the view that the purpose is communicative.
35J. R. Lucas, On Justice, Clarendon Press, Oxford, 1980, p. 133.
36I. Primoratz, “Punishment as Language”, Philosophy, vol. 64, 1989, p. 199.
37R. Nozick, Philosophical Explanations, Harvard University Press, Cambridge, 1981, p. 370.
38Ibid. pp. 376-7.
39For instance, Primoratz believes that punishment is required if criminals “are really to understand how
wrong their actions are”. However, what is meant by really understanding is not explained. I. Primoratz,
“Punishment as Language”, Philosophy, vol. 64, 1989, p. 200.
40T. Baldwin, “Punishment, Communication, and Resentment”, in M. Matravers (ed.), Punishment and
Political Theory, Hart Publishing, Oxford, 1999.
41R. A. Duff, Trials & Punishments, Cambridge University Press, Great Britain, 1986; “Desert and
Penance”, in A. von Hirsch & A. Ashworth, Principled Sentencing, Hart Publishing, Oxford, 1998; “A
Reply to Bickenbach”, Canadian Journal of Philosophy, vol. 18 1988; “Punishment, Communication,
and Community”, in M. Matravers, Punishment and Political Theory, Hart Publishing, Oxford, 1999.
42R. A. Duff, Trials & Punishments, Cambridge University Press, Great Britain, 1986, p. 262.
43See, for instance, R. A. Duff, “Desert and Penance”, in A. von Hirsch and A. Ashworth (eds.),
Principled Sentencing, Hart Publishing, Oxford, 1998, pp. 164-5.
44R. A. Duff, Trials & Punishment, Cambridge University Press, Great Britain, 1986, p. 289.
45Ibid. p. 262.
46Ibid. p. 289.
47See A. von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993. And U. Narayan,
“Appropriate Responses and Preventive Benefits: Justifying Censure and Hard Treatment in Legal
Punishment”, Oxford Journal of Legal Studies, Vol. 13 no. 2, 1993.
48A. von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993, p. 11.
49In chapter 5 I will give a more thorough presentation of what exactly it is that Hirsch means when he talks
about prevention. As the concept of an additional prudential disincentive indicates, it is not a traditional
optimizing view on prevention that he has in mind.
50Recall that when expressionists and other retributivists criticize preventive theories of punishment this is
often by pointing out that these theories may violate the idea of justice captured in the ordinal proportionality
requirement.
51For a more thorough discussion of the possible answers, see J. Ryberg, “The Expressionist Theory of
Punishment and the Problem of Fallible Communication”, Readings in Philosophy and Science Studies, vol.
1, 2001.
52See, for instance, J. R. Lucas, On justice, Clarendon Press, Oxford, 1980, p. 150.
53R. Nozick, Philosophical Explanations, Harvard University Press, Cambridge, 1981. p. 380.
54It might also be suggested that, if a person does not get the message the first time that he or she is
punished, then there is no reason to assume that the communication will succeed the second or third time
the punishment is repeated and that there consequently is no reason to re-punish the person. However,
this answer is also based on a very dubious empirical assumption. It is hard to see why hard treatment
communication, in contrast to all other sorts of communication, should be exhausted by the two
possibilities that the conveyance of a message will either succeed immediately or be doomed to fail
forever.
55R. A. Duff, “Punishment, Communication, and Community”, in M. Matravers, Punishment and Political
Theory, Hart Publishing, Oxford, 1999, p. 49.
56See R. A. Duff, Trials & Punishments, Cambridge University Press, Cambridge, 1986, p. 262f.
57See, for instance, A. von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993 p. 11.
58Amongst those who have defended versions of the fairness theory are: H. Morris, “Persons and
Punishment”, The Monist, vol. 52, 1968; J. Finnis, “The Restoration of Retribution”, Analysis, vol. 32,
1971-2; W. Sadurski, Giving Desert its Due, D. Reidel Publishing Company, Dordrecht, 1985; W.
Sadurski, “Social Justice and the Problem of Punishment”, Israel Law Review, vol. 25, 1991; M. Davis,
To make the Punishment Fit the Crime, Westview Press, United States of America, 1992; G. Sher,
Desert, Princeton University Press, Princeton, 1987; R. Dagger, “Playing Fair with Punishment” Ethics,
vol. 103, 1993.
59H. L. A. Hart, “Are There Any Natural Rights?”, in A. Quinton (ed.), Political Philosophy, Clarendon
Press, Oxford, 1967, p. 61f.
60R. Nozick, Anarchy, State, and Utopia, Blackwell, New York, 1974, pp. 90-95.
61For a critical discussion see, for instance, A. Ellis, “Punishment and the Principle of Fair Play”, Utilitas,
vol. 9 no.1, 1997, p. 90ff.
62H. Morris, “Persons and Punishment”, The Monist, vol. 52, 1968, p. 473.
63R. Dagger, “Playing Fair with Punishment”, Ethics, vol. 103, 1993, p. 484.
64See, for instance, A. Ellis, “Punishment and the Principle of Fair Play”, Utilitas, vol. 9, 1997; or D. E.
Scheid, “Davis and the Unfair-Advantage Theory of Punishment: A Critique”, Philosophical Topics, vol. 18,
1990.
65R. W. Burgh, ”Do the Guilty Deserve Punishment?”, Journal of Philosophy, vol. 79, 1982, p. 209.
66See R. Dagger’s discussion of this criticism in “Playing Fair with Punishment”, Ethics, vol. 103, 1993, p.
477f.
67W. Sadurski, Giving Desert its Due, D. Reidel Publishing Company, Dordrecht, 1985, p. 230.
68Moreoever, it will probably be very difficult to specify in a non-arbitrary way what exactly it means
that a burden occurred “in relation to “ a crime.
69R. W. Burgh, “Do the Guilty Deserve Punishment?”, The Journal of Philosophy, vol. 79, 1982, p. 209.
70G. Sher, Desert, Princeton University Press, Princeton, 1987, p. 82.
71For a discussion of this, see D. Dolinko, “Some Thoughts about Retributivism”, Ethics, vol. 101, 1991,
pp. 547-8.
73Ibid. p. 85.
74Ibid. p. 86.
75S. Kershnar, “George Sher’s theory of Deserved Punishment, and the Victimized Wrongdoer”, Social
Theory and Practice, vol. 23, 1997.
76See G. Sher, Desert, Princeton University Press, Princeton, 1987, p. 85.
77I owe this argument to D. Dolinko, “Some Thoughts on Retributivism”, Ethics, vol. 101, 1991, p. 547.
78G. Sher, Desert, Princeton University Press, Princeton, 1987, p.81.
79J. F. M. A. de Voltaire, Candide and Other Stories, Dent & Sons, London, 1962, p. 20.
80W. Blackstone, Commentaries on the Laws of England, 21st ed., Sweet, Maxwell, Stevens & Norton,
London, 1844, Chap. 27, p. 358.
81Quoted from A. von Hirsch, “Proportionality in the Philosophy of Punishment”, Crime and Justice,
vol. 16, 1992, p. 60.
82See, for instance, H. J. McCloskey, “The Complexity of the Concepts of Punishment”, Philosophy, vol.
XXXVII, 1962; or “A Non-Utilitarian Approach to Punishment”, Inquiry, vol. 8, 1965, reprinted in
G. Ezorsky, Philosophical Perspectives on Punishment, State University of New York Press, Albany,
1972.
83See, for instance, W. Lyons, “Deterrent Theory and Punishment of the Innocent”, Ethics, vol. 84, 1974.
84For reasons already explained in chapter 1, I shall here ignore the definitional stop approach according
to which a punishment can, for logical reasons, only be imposed on someone who is guilty of a wrong.
85H. L. A. Hart, Punishment and Responsibility, Clarendon Press, Oxford, 1968, p. 11.
86See, for instance, A. H. Mitias, “Is Retributivism Inconsistent Without Lex Talionis?”, Revista
Internazionale di Filosofia del Diritto, vol. 60, 1983.
87H. Gross, A Theory of Criminal Justice, Oxford University Press, New York, 1979, p. 436. See also his
“Culpability and Desert”, in A. Duff & N. Simmonds, Philosophy and the Criminal Law, Franz Steiner
Verlag, Wiesbaden, 1984, p. 65.
88In philosophical slang, to “outsmart” has become the term for embracing the conclusion of one’s
opponent’s reductio ad absurdum argument; see D. Dennett & K. Lambert (eds.), The Philosophical
Lexicon, 1978, p. 8.
89J. J. C. Smart and B. Williams, Utilitarianism: For and Against, Cambridge University Press, United
States of America, 1973, pp. 67-73. However, Smart admits that he himself would find it extremely
difficult or even impossible to sacrifice an innocent. See also his discussion in J. J. C. Smart,
“Utilitarianism and Punishment”, Israel Law Review, vol. 25, 1991.
90See T. L. S. Sprigge, “A Utilitarian Reply to Dr. McCloskey”, Inquiry, vol. 8, 1965. For an excellent
discussion of this traditional controversy, see also C. L. Ten, Crime, Guilt and Punishment, Clarendon
Press, Oxford, 1987.
91See, for instance, I. Primoratz, “Utilitarianism and Self-sacrifice of the Innocent”, Analysis, vol. 38,
1978; or his Justifying Legal Punishment, Humanities Press International, London, 1989, p. 44.
92See H. J. McCloskey, “A Note on Utilitarian Punishment”, Mind, vol. 72, 1963.
93See H. J. McCloskey, ”A Non-utilitarian Approach to Punishment”, in G. Ezorsky (ed.), Philosophical
Perspectives on Punishment, State University of New York Press, United States of America, 1972, p.
121.
94Obviously, one might object that the most reasonable interpretation of what is objectionable in the
example has nothing to do with cardinal proportionality, but with the fact that the thief is used as a means
only or that he is punished disproportionately in ordinal terms, that is, relative to how a person guilty of
rape is punished. However, since I do not regard the condition I have pointed at as the crucial objection to
the argument, I shall not here go further into the source of the counterintuitiveness.
95W. Moberly, The Ethics of Punishment, Faber and Faber, London, 1968, p. 80.
96In fact, most of the compromise theories which are considered in chapter 6 would be consistent with
Moberly’s conclusions.
97M. S. Moore, “The Moral Worth of Retributivism” in F. Schoeman (ed.), Responsibility, Character,
and the Emotions, Cambridge University Press, United States of America, 1987; and also his Placing
Blame, Clarendon Press, Oxford, 1997.
98I will return to a discussion of this sort of view in chapter 5.
99On “poena naturalis” see, for instance, N. Walker, Punishment, Danger and Stigma, Blackwell, 1980,
1
p. 130. See, for instance, A. Ashworth, Principles of Criminal Law, Clarendon Press, Oxford, 1995, p.
35f.
.
CHAPTER 2
THE SERIOUSNESS OF CRIMES
Since the claim of proportionalism is that the severity of punishment should be

determined by reference to the seriousness of the crime, the task of clarifying what
makes one crime more serious than another and how different crimes should be
scaled relatively to each other, is obviously of vital importance. Unless it is possible
to tell whether a rape is more serious than a burglary or whether theft is more
serious than reckless driving, proportionalism will be a vacuous view unable to
provide any practical guidance. Not only is some sort of ranking therefore a sine
qua non with regard to how these and other crimes should be punished, but different
degrees of seriousness may also have wider practical consequences for a number of
questions concerning, for instance, the legality of arrest without warrant, decisions
1
in trying a case at higher courts, or decisions to release prisoners on parole, etc.
However, I shall not discuss here these more detailed implications but stick to the
main question of crime comparison.
Though no final scale of crimes has yet been developed by those who have
considered the theoretical background for the proportionality principle, and though
some even admit that much work still has to be done, most adherents to the position
are nevertheless very optimistic. For instance, Primoratz even holds - on the ground
of his Hegelian version of retributivism - that the construction of a scale is “a
2
technical, not a philosophical question” . The purpose of this chapter is to consider
whether this is correct. More broadly, it will be assessed how far recent
proportionalists have come in the development of a crime scale and what problems
such a scaling raises. The first sections of the chapter are concerned with what
might be called the “harm-based” version of proportionalism according to which the
harm caused by a crime, the culpability of the criminal, and perhaps the prior
criminal record, are standardly regarded as the components which determine the
gravity of criminal conduct. The final section is devoted to the answers provided by
a particular fairness theoretic approach to the question which will be considered at
some length. As we shall see, the problem of comparing and ranking crimes faces
proportionalism with two serious challenges: a challenge of relative comparison and
a challenge of absolute comparison. The first challenge concerns the task of
clarifying each seriousness-determining factor in such way that it is, at least in
principle, possible to establish whether it is more present or less present in one
crime than in another. Even if this task were to be solved there would, however, still
remain the challenge of absolute comparison, namely, that of specifying how much
a certain factor contributes to the overall seriousness of a crime. This challenge
59
60 JESPER
THE ETHICS RYBERG
becomes more demanding the larger the number of different seriousness-

contributing factors.
Before entering the discussion, however, there is a possible outlook in need
of comment. It might be objected that the following considerations on whether
adherents of proportionalism have succeeded in outlining the ground for the
construction of a scaling of crimes may seem somewhat superfluous, since people
already have a clear conception of the relative gravity of crimes prior to any
philosophical enquiry. No one apparently doubts that a murder is more serious than
assault or that rape is more serious than theft. This view might find support in the
empirical work carried out by a number of criminologists and other social scientists.
The locus classicus is Sellin and Wolfgang’s The Measurement of Delinquency in
which the authors defend the claim that there is considerable agreement on the
relative seriousness of crimes amongst people from different countries as well as
amongst those belonging to different social groups within the same country.
3
Corresponding conclusions have been reached by other researchers. Moreover, it is
often emphasized that some of the earlier mentioned sentencing commissions which
have constructed scales of crimes have carried out their work without much
4
dissension and without running into other insuperable difficulties. Thus, assessing
the possibility of ranking crimes in gravity might seem to be questioning the
existence of something already there. A pointless academic enterprise. Though it
might be the case that such surveys have some role to play, several proportionalists
have, however, - and I believe rightly - disassociated themselves from scales based
5
solely on popular judgements. Even if it is correct that there is a general agreement
between people as to how the seriousness of different crimes should be rated, this
does not in itself show that the rating should be morally accepted. This would
require an independent argument. Moreover, it is generally agreed that there might
be a divergence between popular judgements and what is morally well-grounded.
The need for a theoretical enquiry clarifying what is morally relevant in the
comparison of crimes is, therefore generally acknowledged among proportionalists.
In this respect, considerations on crimes and punishment are no different from
problems in other areas of applied ethics, which are not solved by carrying out some
kind of poll or by surveying popular judgements, but by revealing and evaluating
the values involved in these problems.
1. THE HARM DIMENSION
The standard view among proportionalists with regard to crime seriousness is that it
should be determined partly by the harmfulness of the conduct. More precisely, the
claim is that, if all other things are equal, the relative gravity of a crime increases
with the degree of harmfulness. Not only is this a broadly accepted position, it is
also a view which is intuitively appealing. If asked why one regards assault as being
a more serious crime than a theft, the obvious answer is that the former misdeed
usually causes much more harm to the victim than does the latter. Moreover, the
view coheres well with the previously sketched arguments on which proportionalism
is held to be based. If one deserves to be punished for one’s wrongdoing then it is
THE SERIOUSNESS OF CRIMES 61
reasonable to claim that harm matters, since causing harm certainly constitutes a
standard example of wrongdoing.
A closer scrutiny into the position that harm counts in the computation of
seriousness requires, of course, an analysis of the concept of harm and of the
question of harm comparison. An important work in this respect has been provided
by von Hirsch and Jareborg in their article “Gauging Criminal Harm: A Living-
6
Standard Analysis”. The theory presented by von Hirsch and Jareborg is the most
elaborated suggestion yet with regard to the assessment of criminal harms and,
therefore, provides a picture of what has been achieved in this area. Moreover this
work, which is very often referred to in proportionalist literature, has been
characterized by other proportionalists as a “pathbreaking”7 contribution to the
discussion. However, though there is certainly no reason to believe that all
adherents of proportionalism will accept all the detailed elements in the analysis,
von Hirsch and Jareborg’s theory nevertheless demonstrates some of the more
general theoretical problems with which proportionalists will be confronted if one,
as a starting point, accepts that it makes sense to compare crimes in terms of harm.
What von Hirsch and Jareborg have developed are some guidelines for a
living-standard analysis of the impact different crimes have on the victims. Inspired
by Sen’s work, the theory is not directly concerned with the quality of life of the
individual victim but with the “means or capabilities” for achieving a certain quality
of life. Furthermore, the analysis is concerned with general judgements, in the sense
that the purpose is to provide guidelines for the estimation of the standard impact a
certain kind of crime has on the living-standard of a victim. Thus, though there are,
of course, large differences between how a crime will affect different people, it is
the normal impact of the crime - say a typical burglary or assault - that is
considered. The theory in this way is based on a considerable degree of
standardization which can hardly be avoided if the purpose in the end is to construct
a general scaling of crimes.
Now, what the theory does is to parcel out the most important kinds of
interests on which crimes typically intrude. The authors distinguish between four
“generic-interest dimensions”: physical integrity; material support and amenity;
freedom from humiliation; and privacy/autonomy. Naturally, the thought is not that
a crime necessarily affects all of these dimensions. While a residential burglary may
affect the material amenity dimension and the privacy dimension, a forcible rap
involves the physical integrity and the humiliation dimensions. With these
dimensions introduced, the next part of the procedure is to indicate the degree to
which a typical instance of a certain crime affects one or more of the dimensions.
Von Hirsch and Jareborg separate four living-standard levels: Level 1 (subsistence):
survival with maintenance of no more then elementary human functions; Level 2
(minimal well-being): maintenance of a minimal level of comfort and dignity; Level
3 (adequate well-being): maintenance of an adequate level of comfort and dignity;
and Level 4 (enhanced well-being): significant enhancement in quality of life above
the mere adequate level. As the final part of the machinery, the authors introduce a
harm-scale which grades harms from the very grave to the minor. With the purpose
of not given a misleading impression of precision, they separate five broad bands of
62 JESPER
THE ETHICS RYBERG
harm gravity: grave, serious, upper-intermediate, lower-intermediate, and lesser.

The way a certain crime maps onto the harm-scale is simple. If a crime intrudes
upon living-standard Level 1, it is rated in the category “grave”. If it intrudes upon
Level 2, it is graded as “serious”. If it intrudes upon Level 3, it is mapped as “upper-
intermediate”; and if it intrudes upon Level 4, it is graded as “lower-intermediate”.
Finally, if a crime has only a marginal effect on the living-standard, it is mapped as
“lesser”.
With these technicalities settled, the ground is provided for the
construction of the harm scale. All one has to do is to consider the living-standard
level upon which a certain crime intrudes and then place it into the corresponding
harm category. For instance, since a homicide destroys subsistence, which is
indicated at Level 1, it should be ranked as “grave”. While a petty assault, which
according to the authors affects the humiliation dimension to an extent
corresponding to Level 4, should be ranked as “lower-intermediate” on the scale. As
mentioned, a crime may also have an impact on two or more of the generic interest
dimensions. For example, assault and battery may affect both the physical integrity
dimension and the humiliation dimension. When this is the case, it is suggested that
one should, as a first step, identify the dimension that yields the highest harm-rating
(the primary harm) and subsequently identify the dimension that yields a lower
harm-rating (the secondary harm). When a crime involves secondary or further
harms these should be regarded as an exacerbating feature. The net harm, therefore,
is determined by adding a premium to the primary harm. How much this
exacerbation should amount to depends on the rating of the secondary harm. The
graver it is, the larger the premium.8
What is obvious with regard to this suggestion is - and this is clearly
pointed out by von Hirsch and Jareborg - that the theory is not a formula which
simply delivers a complete ranking of harms. For instance, there is no clear answer
as to how different crimes, each affecting more than one interest dimension, should
be ranked in relation to each other. Suppose that one crime intrudes on one
dimension to an extent corresponding to Level 1 and also on another dimension to
an extent corresponding to Level 2. While another crime has an impact on three
different interest dimensions to an extent corresponding respectively to Level 1,
Level 3, and Level 3. In such a case, there is no answer to how the harm of the two
crimes should be ranked relatively to each other, since there is no clear answer to
how much the exacerbation is when more than two dimensions are affected by a
crime. However, the method does not pretend to provide a strict metric but rather a
guide that may qualify our judgments or, as has been suggested, a stage of thought
through which it would be desirable if members of a sentencing commission passed
when making their decisions. However, even if we accept - which surely seems
reasonable - that at some point all we can do is rely on judgments when assessing
the harms caused by different crimes, this does not change the fact that a ranking of
harms faces the proportionalist with problems which require a theoretical solution.
Problems which arise independently of whether one accepts the more detailed
elements in the guidelines suggested by von Hirsch and Jareborg.
The first problem is part of a larger challenge confronting the claim that
seriousness is determined by harmfulness. The larger problem simply is that there
are several crimes which do not, at least in no straightforward way, involve harmful
conduct.9 A standard example is conduct which only risks or attempts harm. For
instance, the risk caused by reckless driving. Even if I drive very hazardously in a
crowded street there may nevertheless be no one who is actually harmed. Similarly
with regard to the inchoate crime of attempt. A planned crime may not succeed
simply because the person who has set out to commit it does not, for some reason or
another, perform all the acts necessary to bring it about (incomplete attempt). For
instance, a man who intends to shoot another may be caught by the police before he
gets the chance to pull the trigger. A person may also do all that is intended but
nevertheless not succeed in bringing about the desired result (complete attempt).
This would be the case if the man actually pulls the trigger but fails to hit the
potential victim. If one accepts that such conduct should in the first place be
criminalized and that it therefore deserves a punitive response, how serious should
these crimes then be regarded as being when they involve no resulting harm?
A possible way to respond, at least partly, to the problem is to adopt a
subjectivist point of view. The distinction between subjectivism and objectivism
constitutes a traditional dispute about the nature of a system of criminal law.
According to subjectivism, what matters is the harm related to the intended conduct
or, more broadly, to the conduct as perceived by the criminal. Motivated by the
view that the actual outcomes cannot serve as a proper base for blame, since they
may be a result of good or bad luck, subjectivists like Ashworth believe that “the
criminal law and the principles of sentencing ought in principle to hold him [a
defendant] liable for that which he intended, no more and no less”.10 This implies
that, in the case where the man shoots at another to kill him, it is irrelevant whether
the person is actually killed, whether the victim is only wounded, or whether the
bullet misses its target. What counts is the harm related to the intended crime in
casu the killing of a person. The subjectivist can in this way nicely account for the
question on attempts. They should be assessed on a par with a completed crime.
Similarly, what counts with regard to risks is the risk the defendant believed he was
taking. However, what if one does not accept subjectivism? What if one, like von
Hirsch and Jareborg and other objectivists, believes that it is the actual harm that
counts with regard to seriousness?
The way von Hirsch and Jareborg respond is to incorporate risk judgments
in the guidelines. What is suggested is a two-step procedure. Firstly, one should
determine the living-standard level that would have been affected by the completed
crime. Thus, at this step homicide, armed robbery and drunken driving are all
ranked at the level “grave” since they all affect the interest in subsistence. Secondly,
the net harm is estimated by risk-adjusting the harm identified at step one. This is
done by adding a discount to the non-adjusted harm. According to von Hirsch and
Jareborg this might, for instance, imply that certain attempts constitute a sufficient
high risk to keep attempted homicide in the “grave” range, albeit at a point below
the completed harm, while there might be a somewhat larger discount for the risk in
armed robbery, leaving it in the “serious” range, and an even larger discount for the
risk in drunken driving, placing it in a lower harm category.
64 JESPER
THE ETHICS RYBERG
However, this leaves a problem to which neither von Hirsch and Jareborg
nor other proportionalists have provided a clear answer, namely, what exactly is
meant by the risk of a harm? This is not at all obvious. In fact, von Hirsch and
Jareborg’s application of risks is not even clear. If what morally matters with regard
to seriousness is the risk-adjusted harm, that is, if what counts is harm risked times
the probability of causing the harm, then it seems to make no difference whether the
harm actually occurs or not. If there is a 90 % change that a harm follows from a
certain conduct, then the risk-adjusted harm of the conduct is the same (the harm *
0.9), no matter whether or not the harm takes place. It is therefore unclear why von
Hirsch and Jareborg hold that they “assume that attempts should be treated as less
serious then completed crimes11. To claim that there should be risk-adjusting when a
harm does not take place but not when it actually occurs seems somewhat
incoherent. At least some sort of justification should be required. However, even if
we leave these details aside, the problem remains to explain what exactly a risk
indicates. To what are we referring if we talk about the risk in reckless driving, or
the risk in an attempted killing in which one person shoots at another but misses?
One possibility is to rely on statistics. This seems possible in the case of
reckless driving. In fact, this has been suggested even with regard to attempts.
Husak has proposed that one might “calculate the percentage of attempted crimes
that actually succeed and then use [..] this ratio to discount the punishment for
12
unsuccessful attempts” . This kind of approach faces several problems. For
instance, statistics on reckless driving only contain the instances that are registered.
Similarly, statistics on attempts would only contain those instances that were
reported. But should the risk only be calculated on the ground of these cases, or is it
the thought that one should try to make some estimates of the total number of
violations? More importantly, there is a problem of multiple descriptions. What kind
of statistics should one confront in the calculation? Statistics on all sorts of reckless
driving or of a special sort? Should the statistic on attempts include all kinds of
attempts or, for instance, only attempted killing, or attempted killing by shooting, or
perhaps attempted killing by shooting with a certain weapon at close range? It is far
from clear what would count as a reasonable answer to these questions.
The lack of a clear concept of probability implies what I have introduced as
a challenge of relative comparison: the challenge of comparison in terms of more or
less within a serious-determining dimension. For instance, there is no ground for
comparing a certain harm which does not involve a risk (i.e. where the probability is
1) with a greater harm which does involve a risk. Thus, all in all it is clear that the
guidelines in von Hirsch and Jareborg’s analysis, or in possible alternative analyses,
require much further theoretical elaboration. With regard to the larger question, that
some crimes do not in any straightforward way involve a harm, it is moreover worth
remembering that von Hirsch and Jareborg’s theory is limited to criminal conduct
which injures an identifiable victim, i.e. a person. Thus, it is not constructed to
consider crimes against the state or against firms. At this point much work still
needs to be done.
A second problem of harm ranking, which to some extent relates to the
problem just considered, concerns what is sometimes referred to as “remote harms”.
Remote harms do not involve harms which are merely spatio-temporally distant. The
fact that a bomb which kills a person is timed does not seem morally relevant, and
could just as well be accounted for by von Hirsch and Jareborg’s method as any other
killing. Rather, what is meant is harms which stand in such a relation to a conduct that
it is not clear whether they should be ascribed to that conduct. Since this is obviously a
very vague definition, a few examples will be more illuminating. One example is
conduct which triggers a series of events that eventually have harmful consequences
13
and where the agent’s own or other people’s choices intervene in this series. This is
the case with regard to the possession of weapons, which is not in itself harmful but
which might have harmful consequences if the possessor himself or other people chose
to use the weapon. Another example which I believe worth mentioning is in cases
where a conduct triggers a series of events leading to harm but where there are no
intervening choices. The killing of a person leads to an immediate harm but it may also
be painful for relatives and thus harm them. Or the theft of a few coins might imply
that the person from whom they were taken is later unable to call an ambulance from a
phone box which eventually leads to the death of a person. A final example is
accumulative harms where a harm follows from an act only when it is combined with
similar acts of others. Conduct leading to environmental damages may be of this kind.
What is interesting about these cases is that, in so far as instances of such
conducts are criminalized, they raise the question of how much of the harm that is
triggered by the conduct of an agent should properly be held accountable for by the
agent? To claim that the question is not really relevant since what we consider in the
computation of seriousness is, as is the case in von Hirsch and Jareborg’s theory, the
standard harm, and that the standardization will eliminate the different examples, is
obviously not plausible. Firstly, the reason for standardization is not that it is the
standard harm of a crime that basically matters. Rather, standardization is a question
of adapting to what is practically possible in a functional sentencing system. But this
means that it still makes sense to ask whether the person who committed the theft of
a small amount of money should, in principle, be held accountable for the death of
the person who was not saved. Secondly, there are still many cases that would
include remote harms even on a standardized account. For instance, the possession
of weapons certainly involves a risk which would not be eliminated by standard
considerations. Crimes involving possession, as well as other kinds of crime, also
indicated that it will not do simply to object that, when intervening choices are
involved in a chain of events, nothing of the harm that eventually occurs should be
attributed to the initial triggering act. This would imply that there would be a
number of acts which though generally proclaimed to deserve a punitive response
would be no longer punishable. An example is incitement, that is, cases where
someone encourages or instigates another person to perform a harmful act. Finally,
one should not expect that mens rea somehow establishes which of the harms that
flow from an act should be attributed to the act and should thus figure in
computation of seriousness.
As mentioned, objectivists believe that it is the actual harm caused or risked
that counts. However, the question of what is a fair attribution of harm to a certain
act is not even answered by moving in a subjectivist direction by claiming that it is
only the intended or known harm that matters. If the object of the intention is defined
narrowly, that is, if the harm that counts is the one that the criminal specifically
intended to cause the victim, then too many acts which are usually regarded as worthy
of punishment will be excluded as not containing any intended harm. After all, in most
crimes the intention is not to cause a well-defined amount of harm to a victim, but
perhaps to gain some goods, or whatever. On the other hand, if what counts is the
harm related to the kind of act that was intended (e.g. to commit a theft) or the harm
that one knew would be risked by the act performed, then the question about remote
harms is still relevant. One would still have to consider whether the harm that is caused
to relatives of a person who is killed should count when assessing seriousness (for
obviously anyone who kills or risks killing someone usually knows that there is a risk
that this will harm relatives to the victim). Or to take another example, if it is correct
that children who have been sexually abused are more likely to become paedophiles as
adults, then a person who commits paedophilia might well have this knowledge, which
again raises the question of whether any future harms as a result of sexual abuse of
children should be added to the present.
Thus, it remains a genuine problem to clarify how much of the harm that
follows from a criminal act the criminal should be held accountable for. To contend
that it is all the harm that follows from the act is not plausible. This could imply that
possession of weapons should be regarded as causing the same harm as the killing
of a person. Thus, some other view is required. However, on this point
proportionalists have not had much to offer. One possibility perhaps is to introduce
some moral principles which are able to point out which of the harms that are
triggered by a criminal conduct should be attributed to the act. Yet what these
principles should consist in is far from clear. Another possibility, coherent with the
way von Hirsch and Jareborg account for attempts in their procedure, is to claim
that all cases involving different sorts of remote harms should be handled by
probabilistic methods. Thus, one should count in the risk of future harm caused by a
present act in all the different examples. This, of course, restates the question of
how risks should be calculated. Should statistics be involved? Should cases of
accumulative harm be accounted for by application of game-theory? It obviously
also raises the same problems as, for instance, the problem of multiple descriptions
of a certain conduct. On these points it is not unfair to claim that proportionalists
have not had much to offer.
Von Hirsch and Jareborg are aware that their method is not able to account
for some of the more sophisticated problems as, for instance, cases involving
accumulative harms. But it is worth noticing that even when it comes to crimes the
harm of which the method should be able to account for - that is, in cases of what
they call “ordinary victimizing offenses” - the calculations are more complicated
than they are depicted in the examples von Hirsch and Jareborg use. In cases
where a person is killed the net harm not only involves the harm that is caused to the
victim by taking his or her life but also the risk that this will harm the relatives to
the victim. Similarly, there might be many other crimes which, beside the harm that
is “directly” caused, also involve risks of more remote harms which should thus also
be counted in. With regard to attempts and other risks, the picture is also more
complicated. Not only does attempted murder or drunken driving constitute “risks to
survival” but they also constitute other risks, for instance, a risk that a victim will be
seriously injured, a risk that a victim will be less seriously injured, and perhaps risks
that several persons will be harmed in one way or another. In principle, all these
risks should be added in order to get a picture of the risk-adjusted harm caused by a
certain attempt or risk. This will also be the case with regard to many other crimes.
What all this shows is that, when it comes to the harm dimension of
seriousness, there is still much work to be done. This is not only a question of
carrying out the relevant calculations of harms but of clarifying the theoretical
background for performing such calculations. That is, clarifying what basically
matters. Thus, even if one believes, as I certainly do, that it makes sense to talk
about some crimes being more harmful than others and that harm is the least
problematic determinant of crime seriousness, a sufficient theoretical foundation has
not been developed. To this must be added a further comment which does not
concern the specific method applied in the computation of harms but the question of
exactly what it is that is estimated.
As has been indicated, the proportionalist discussion on how harms should
be gauged is not meant to result in a procedure that should be applied with regard to
the harm of each individual crime that is committed. Rather is it supposed to
estimate the typical harm caused by a certain kind of crime. This is explicitly
pointed out by von Hirsch and Jareborg. In their procedure, standardization is
perhaps even applied at several levels - both with regard to the estimate of the harm
of different crimes, and perhaps also with regard to calculations of the harm risked
by different sort of conduct. However, in real life the harm caused by a certain type
of crime obviously varies very much from one case to another. A person who is
physically, psychologically or socially vulnerable may suffer much more harm from
a certain crime than someone more resilient. And even if the intensity of a harm is
the same in two cases, one victim may, nevertheless, suffer the harm much longer
than another victim. What this means is that the harm that is attributed to a certain
crime by following the von Hirsch/Jareborg procedure, or by any other
proportionalist theory relying on standardizations, may be much different from the
harm which a specific instance of the crime actually causes its victim. The reason
why this is morally interesting is that standardization thereby seems to contradict the
core of proportionalism, namely, the claim that no one should be punished more
severely than what is warranted by the crime one has committed. When the
seriousness of a crime is, at least partly, determined on the ground of
standardizations one has opened up the possibility that the actual harm caused by a
person in committing a crime is less than the standard harm, and that the person will
consequently receive a punishment more severe than that deserved from the act
committed. In short, standardization might well lead to what is, from the
proportionalist’s own point of view, an instance of injustice. Nevertheless, it seems
to be broadly accepted that standardizations are required. In von Hirsch and
Jareborg’s view “criminal acts are too diverse to be rated on an individualized
14
basis” . However, if one for pragmatic reasons deviates from what is prescribed by
an ethical theory, then it needs to be established that the practical modification can
actually be justified within the framework of the basic theory. At this point
68 JESPER
THE ETHICS RYBERG
proportionalists have not been very careful in their considerations. I shall postpone a
related discussion to a later chapter. But it is worth noticing that, besides the more
technical questions on methods for gauging harms, there is this more basic problem
on the consistence of applying standardizations in the application of a principle
which prescribes proportionality between crime and punishment.
2. CULPABILITY
That the question of the allocation of punishment should be determined on the

ground of an interpretation of lex talionis (eye for an eye, tooth for tooth), which
considers only the amount of harm caused by a criminal act, is generally rejected by
recent adherents of proportionalism. Not only does lex talionis lead to a number of
both practical and theoretical problems - some of which will be outlined in chapter 4
- but it also suffers from the defect that it makes no allowance for the mental state of
the criminal or for the circumstantial aspects of the crime.15 In short, it simply
ignores the other major component of seriousness, namely, the criminal’s
culpability.
In consideration of the arguments which proportionalists have presented in
favour of the justification of punishment, the need for the culpability component is
obvious. To claim that a person should be blamed, or more generally that he
deserves to be punished, for a harmful act, independently of any considerations on
whether he, for instance, acted intentionally or was in some sense responsible for
the act, seems implausible. An indication of the role culpability therefore standardly
plays in proportionalist thinking is given by a simple formula suggested by
Nozick.16 If we, for a moment, ignore the specific content of Nozick’s own theory,
and instead apply his formula more generally, the seriousness of a crime can be
determined on the ground of the product C*H where H is the harm done or risked,
while C is the culpability of the criminal, indicated by the numerical values from
zero to one.17 What this formula illustrates is, firstly, that when there is no
culpability - that is, when C=0 - a defendant deserves no punishment and, secondly,
that culpability is a matter of degree. Though proportionalists agree on these formal
characteristics, there is much disagreement when it comes to the question of what
exactly determines the degree of culpability and to what extent. While the harm
component, as indicated in the previous section, has not been the subject of much
discussion amongst proportionalists, the culpability aspect of a crime has received
extensive attention. One of the things which the discussion has revealed, and which
certainly complicates the judgement on the degree of culpability of a certain crime,
is that the study of culpability is poly-dimensional enterprise. Not only does it
include considerations on mens rea, but it also involves considerations of personal
responsibility.
In the following, I shall start by briefly considering mens rea and then turn
to a discussion on responsibility as instantiated in the traditional theories of excuses.
When going through these dimensions of culpability, I will restrict the writing
strictly to what is of direct importance with regard to the cardinal question on the
possibility of ranking crimes in gravity. From this perspective the discussion reveals
that the different aspects of culpability face several instances of the challenge of
relative comparison. Moreover, the culpability dimension is also confronted with
what I initially introduced as a challenge of absolute comparison, which in this
context concerns how different degrees of culpability contribute to the final degree
of seriousness.
The first traditional aspect of the culpability of a criminal concerns the
mental states or attitudes a person holds when a harmful action is performed. The
law usually uses the term “mens rea” (the guilty mind) to connote these mental
elements. The conventional mens rea distinctions are between: intention,
knowledge, recklessness, and perhaps negligence. Other fault terms have sometimes
been used; however, with the purpose of this chapter in mind it is sufficient to focus
18
on this quadruple distinction. In short, and ignoring details, we can say that: a
person A did something intentionally if it was his conscious object to bring it about;
that is, if he did not bring it about he would regard himself as having failed in his
enterprise. A did something knowingly if he knew that his act would result in X
even though accomplishing X was not the objective of his undertaking; that is, if X
was for some reason was not produced he would not regard himself as having
failed. A did something recklessly if he consciously disregarded a substantial risk
caused by his act. And finally we can say that A did something negligently when he
19
carelessly disregarded a risk without knowing he was doing so. Much has been
written about these categories of mens rea in attempting to state more precisely how
the terms should be defined. As indicated, it is also a controversial question whether
holding a person liable for negligent acts is at all distinguishable from strict
20
liability. However, what is interesting here is how the different mental attitudes
which are part of an act with mens rea affect the degree of culpability and at the end
the seriousness of a crime.
The standard view is that the different kinds of mens rea reflect different
degrees of culpability. Whether this is the case with regard to the distinction
between intentional and knowingly done harm is a matter of dispute. Some maintain
that there is an important distinction, while others suggest that this distinction
21
should not imply different degrees of culpability. A clarification of this leads into
the traditional ethical discussion of the doctrine of double effect. However, ignoring
this question it is at least generally agreed upon that intentional harm ceteris paribus
typically implies a higher degree of culpability than recklessly caused harm, which
again implies a higher degree of culpability than harm caused negligently. It is
interesting to notice that this view is quite often not supported by more profound
arguments for why one mental state accompanying an act makes a person more
culpable. However, the distinctions certainly have an intuitive appeal. As Hart
illustratively remarks, it seems worse to break someone’s Ming china intentionally
than to knock it over while waltzing wildly round the room not thinking of what
might get knocked over. Though both actions may be blameworthy, it sounds
reasonable to hold that the person deserves a tougher treatment the more that
person’s mind is focused on bringing about the fatal result. Let us therefore assume,
as proportionalists usually do, that mens rea in the indicated way affects the degree
70 JESPER
THE ETHICS RYBERG
of culpability. This brings us to some of the more basic problems which confront the
proportionalist view on the matter.
The first thing which complicates the judgements on mens rea is that the
mentioned mens rea categories each cover a wide range of mental attitudes. A
person might act intentionally when he or she performs an act on the spur of a
moment, but also when the act is carefully planned in advance. Ashworth has in this
connection suggested a table of mental attitudes distinguishing five different forms
of intention. Similarly, recklessness can exist in varied forms depending on, for
instance, whether it involves a calculated or a spur of the moment risk. Each of
22
these variants implies different degrees of culpability. Thus, considerations of the
different kinds of intention, recklessness and so on is of great importance for the
proportionalist. This is further emphasized if one accepts Ashworth’s claim that,
though most cases of intention are worse than most cases of recklessness, there may
be some cases of recklessness which are more serious than some of intention. For
instance, a crime which involves planned recklessness may be more serious than
one committed intentionally but impulsively. Both of these observations face the
proportionalist with a challenge of relative ranking, that is, the challenge of
indicating how the variety of forms of mens rea should be manifested in a relative
grading of culpability. We need to know whether - and why - one form of intention
renders a higher culpability than another, and whether some forms of recklessness
imply a higher culpability degree than some of intention. Thus, the grading of
culpability on the ground of mens rea may be a much more complicated matter than
the quadruple mens rea distinction just outlined indicates at first glance.
The significance of this complexity, however, is best realized if we turn our
attention to the second and more serious challenge which mens rea generates: the
challenge of absolute comparison. The question is how to combine mens rea and the
harm scale to get an account of the seriousness of different crimes. Or put
otherwise, what exactly does it imply to say that a person is more culpable if a harm
is caused intentionally than if it is the result of recklessness? If this statement should
- in contrast to what is indicated by the Nozickian formula - be understood in purely
ordinal terms, that is, if all we can say is that it is a more serious crime if a harm is
done intentionally than if the same harm is caused recklessly, then there is no clear
way to the construction of a complete crime scale. For instance, let us assume that
the harm caused by a homicide is greater than the harm caused by an assault (we can
even assume that we can measure the harms in absolute terms), then how should an
intentional assault be ranked in comparison to a reckless homicide? As
indicated in the table, we can say that the intentional homicide is more serious
than the intentional assault, but there is no answer as to how reckless homicide
should be
CRIME HARMMENS REA RANKING
murder death intention murder

?
manslaughter death recklesness assault manslaughter
?
assault bodily indury intention
ranked relative to the assault. A similar problem, of course, rises with regard to
the comparison of all the other crimes on the scale. That is, in all other cases which
require the same kind of comparison between crimes of varying degrees of harm
and culpability. And it is important to notice that this is not simply a matter of lack
of precision but rather a matter of theoretical indetermination. However, if it is
alternatively assumed - now in accordance with the Nozickian formula - that the
different degrees of mens rea can be compared in absolute terms then it needs to be
clarified how this should be done. That is, we need an indication of have much more
serious a crime is when a specific harm is caused intentionally, recklessly or with
another mental attitude. This certainly emphasizes the significance of the fact that
each of these mens rea terms covers a wide range of attitudes.
The notion of culpability goes far beyond the concept of mens rea, as in the
sense just outlined. The second dimension of culpability which needs to be
considered concerns what is standardly referred to as “excuses” or “defences”. No
matter whether or not excuses are supposed to figure explicitly in sentencing grids,
they certainly play an important role in the complete evaluation of a crime. The
many different instances of excuses may roughly be classified in the following
groups.23 A first category consists of actions which are basically involuntary. This
might be due to external as well as internal causes. In a case where a person’s bodily
movements are part of a causal sequence bringing about a harm, the person might be
excused if the movements were for external reasons not under his control. An
obvious example is physical compulsion. The source of the lack of control is
internal, for instance, in the case of epileptic seizure. Whether both of these cases
are more properly described as not being actions at all but rather involuntary
movements is a question we can here leave aside. The important thing is that the
person should, despite the harmful result, not be regarded as having violated a legal
or moral norm. The second group of excuses includes cases where a person
performs a harmful act but does so under constraints from defects of knowledge or
defects of will. The first might be the case where a person shoots at a target on the
shooting range and kills a person hiding behind it. And the second where a person
does something under duress or as a result of provocation. Given these conditions,
the judgement might be that any ordinary law-abiding person would not have acted
differently. The excuses in the third group include cases where a person lacks
sufficient capacity to make judgements. Examples of this branch of excuses are
intoxication, infancy and insanity.
The examples in each of the above categories are not supposed to exhaust
the list of possible excuses. Neither should one believe there to be no disagreement
with regard to the legitimacy of different excuses. However, it is a fact that
proportionalists standardly accept the existence of excuses. This, of course, raises
the question of justification. Why should a defendant be excused by the
circumstances sketched in one or more of the three categories? The appealing
answer is that desert and blame presupposes personal responsibility. It would be a
matter of injustice to blame a person or in another way make a person pay for
something for which he or she is not responsible. As this indicates, excuses
instantiate an underlying view on responsibility. When it comes to a specification of
this theory, however, there is disagreement among different views. Theorists have
been divided between those defending a Kantian-inspired choice theory and those
advocating a Humean-inspired character theory of responsibility.24
The choice theory is traditionally the one which has had most adherents. In
this view a person is responsible for what he or she freely chooses to do, and not
responsible for wrongs he or she lacks the freedom to avoid doing. What is meant by
“choosing freely”, therefore, is the crux of the discussion. However, in a classical
modern formulation of the theory this is explained by Hart in the following way:
“What is crucial is that those whom we punish should have had, when they acted,
the normal capacities, physical and mental, for doing what the law requires and
25
abstaining from what it forbids, and a fair opportunity to exercise these capacities.”
The responsibility is thus conditioned both by the equipment of the actor and by the
situation in which the actor finds himself. For instance, in the case where a person
suffers from a fundamental deficiency of the mind it might be reasonable to claim
that he lacks a sufficient choosing capacity to be responsible. While in a case where
a person acts from necessity or under a significant pressure on his will it might be
judged that the situation does not present him with a fair chance to exercise his non-
defective capacities for choosing. Whether particular excuses are most adequately
explained as a lack of capacity or as a lack of possibility for exercising the relevant
capacity is not a matter that needs to be considered here.
What is important is that, though excuses might in some cases play a fully
exculpatory role, they usually function only as extenuating factors. This is not
surprising. It is not hard to imagine a sliding scale of intensity of the many factors
and circumstances which justify excuses, and this gradation is exactly what desert-
based theories should reflect by variable mitigations of the seriousness of a crime
and consequently of punishment. However, this leaves the choice theory with a
challenge of relative ranking manifesting itself in three questions on which
proportionalists have, with a few exceptions, been remarkably silent. In order to
estimate the extent of a wrongdoer’s responsibility, and thereby the degree to which
one excuse in comparison with another reduces culpability, it is, firstly, necessary to
clarify what exactly is meant by “choosing capacity”. Besides indicating that this
26
involves certain reasoning abilities, it is often not very well defined. Secondly,
further elaboration is needed on what it means to have a “fair opportunity for
exercising one’s choosing capacity”. An answer might either be to claim that this
somehow indicates the degree to which circumstances make a choice
psychologically harder for an agent, or it might be to adopt an objective criterion
according to which there must be some objectively regarded evil that one is avoiding
in order to lack a fair opportunity to avoid doing wrong.27 Given that sufficiently
clear conceptions of choosing capacity and fair opportunity have been developed
there is, thirdly, the question of how these two sources to excuses comparatively
affect the degree of seriousness of a crime. Suppose that an intoxication had some
impact on a person’s choosing capacity when he acted wrongly and that another
person acted wrongly under duress and therefore had an unfair diminution of his
opportunity to avoid the wrongdoing. If we assume that the wrongful acts are equal
in all other relevant aspects, then how do each of the two excuses affect the final
ranking of the two crimes? In other words, it is not sufficient only to consider
different degrees of capacity-excuses and fair opportunity-excuses: what is required

as well is reflection on which degrees of the first sort of excuses should correspond
to which degrees of excuses of the second sort. A point on which choice-theorists
have not had anything to say and which certainly does not invite an easy answer.
Thus, much is required to answer how an excuse in one case affects culpability
compared to another excuse in another case. The challenge is no less if we turn to
the rival theory.
Though most adherents of proportionalism have apparently relied on the
choice theory of excuses, a number of theorists have suggested that a more plausible
account of excuses is provided by the character theory.28 According to this theory,
ascriptions of responsibility are - as the name indicates - based upon judgements
about the character of the agent. However, this theory of responsibility apparently
opens up the existence of two different kinds of excuses which are often not clearly
separated by those who have presented the position. The first view (a) is that we are
excused from our wrongful actions when they are not determined by or expressive
of our character. This is what Nozick expresses when he says that “[a]n action is
done and its apparent explanation sees it as produced by a defect of character
(explicitly so characterized, or by traits that constitute a defect), the act being an
expression of that character disposition. Excuses undercut this explanation by
pointing to another explanation of the action that involves either no character defect
or a lesser one; this new explanation replaces the earlier one and its apparently
29
(more serious) character defect.” . Thus, with this view, excuses accord with the
attitudes we express when claiming of someone’s bad behaviour that “it was not like
him” or that “he acted out of character”. What they concern is the relation (or lack
of relation) between a person’s character and his actions. The second view (b)
relates excuses more basically to a person’s character and how this character came
about. More precisely, the view is that an agent is responsible for an action to the
extent that he is responsible for those aspects of his character which led to the
action. In this sense, responsibility for actions is derivative of responsibility for
character. As an example, Arenella mentions the Patty Hearst case in which the
eighteen-year-old Patty was kidnapped by revolutionaries.30 After weeks of
indoctrination and abuse she renounced her past and joined her kidnappers in their
criminal activity. When arrested she defended herself on the ground that she had
been brainwashed. According to Arenelle, the appeal of the brainwash excuse lies in
the claim that Patty was not responsible for those aspects of her character that
motivated her wrongdoing. With this view of excuses what matters is the impact (or
lack of impact) a person has on his character. Let us consider (a) and (b) in turn,
with regard to what each requires to meet a challenge of relative comparison.
What is required to give an account of the degree to which different
excuses diminish a defendant’s culpability on view (a)? This is not at all obvious.
Since the claim is that what matters is the extent to which a person acted out of
character, one possibility would be to determine how defective a person’s character
is, and how wrong his action was, and then consider the extent of how much each
judgement corresponds to the other. However, it is surely not theoretically clear
what this amounts to. In order to determine the extent of the defectiveness of a
character, it will probably be necessary to answer the intricate question of what a

character is.31
Suppose, first, that this is explained in purely behavioural terms. That is, to
ascribe a character trait like honesty to a person is simply a general description of
the person’s past honest actions. What we would have to do to establish that a
person has a bad character would be to determine whether his past actions have
been bad. This is obviously a difficult calculation since it involves actions differing
both in number and wrongness. However, there is an even more intricate problem
related to this view. What does it mean to say that a person’s former actions were
wrong? To say that this simply means that they were harmful would hardly make
the view plausible. This would imply that a person would have a bad character and
hence be (more) responsible for a present wrongdoing, even if the harm caused by
his former actions was a result of pure bad luck. This is surely hard to accept. The
obvious way to avoid this implication would be to suggest that the wrongness of
former actions is not only measured by the harm that was caused but that it is also a
precondition that the person was actually responsible for these actions. If the person
has been responsible for the harm he has caused in his life up to now this is what the
behavioural character-theorist finds important. However, this solution would lead
the character theory directly into a vicious regress.
Suppose that a defendant has performed the harmful action X0. Now, in
order to find out whether he is responsible for X0, we would have to find out
whether he was responsible for his former harmful action X-1. To claim that in
making this judgement we should rely on the choice theory of responsibility would
be odd. It would mean that the character theory had to presuppose the choice theory,
and it would certainly make it hard to explain why X0 should not be judged
according to the choice theory in the first place. However, if we base our judgement
on X-1 on the character theory in the behavioural interpretation, then we would have
to find out whether the defendant was responsible for an even earlier action X-2.
This again would presuppose that he was responsible for X-3, and we would have to
continue like this until we arrived at the first action X-n the person ever performed
which was harmful in the relevant sense. However, since the person had not
performed any harmful action before X-n, the person would not at that time have had
a bad character, which means that he would not be responsible for X-n. It was
performed out of character. However, this would imply that the person was neither
responsible for X-m, which would imply the same for all the actions all the way
forward to X0. In other words, the character theory which applies a behavioural
concept of character implies that a person is never responsible for an act no matter
what he has done before. Or put otherwise, one will always be fully excused for
one’s wrongdoing. But this is certainly not what was intended by the theory.
A more promising approach would therefore be to reject the pure
behavioural understanding of a character, in favour of a notion according to which
the character is prior to or a condition for actions. No matter whether this is done by
adopting a Rylean dispositional view of character traits or whether the character is in
another way regarded as what causes our behaviour one would not run into the kind
of problem just outlined. With this notion one might go so far as to answer in the
affirmative Dummett’s question on whether a man could have been courageous if he
in his entire life were never in a situation that called for courageous behaviour.32
However, though this answer escapes vicious regress, it is certainly also in need of
clarification. It is necessary to specify what a character in this particular sense
amounts to. And it needs to be indicated according to which standards a character’s
badness should be evaluated. Neither question admits of an easy answer.
A final comment is worth making in relation to view (a). It might be
suggested that, in order to establish the degree to which an action is expressive of a
person’s character, one does not need not go into difficult comparisons between
character and action. If it is assumed that a person’s actions are under normal
conditions expressive of his character, then all we have to consider to determine the
degree of an excuse would be the degree to which the conditions deviate from being
33
normal. The question would then be: what would count as instances of conditions
which make it unreasonable to attribute a wrongful act to a person’s character?
Though it can be disputed there is at least one answer that easily comes to mind.
That is, either cases where the person lacks capacity to avoid wrongdoing or cases
where the circumstances leave the person with an unfair capacity to exercise his
non-defective capacity. A character theorist like Fletcher explicitly states that “a
particular wrongful act is attributable either to the actor’s character or to the
34
circumstances that overwhelmed his capacity for choice.” . When the latter is the
case the actor is excused. However, given this answer, the character theory
apparently coincides with the choice theory. Thus, unless another specification of
“abnormal conditions” is suggested, the problem for this interpretation of view (a) is
exactly the same as the challenge facing the choice theory: that of specifying the
two possible causes of excuses and the relative degrees to which they excuse.
If we turn instead to view (b), the basic claim is that we are responsible for
our character. Thus, on this interpretation of the character theory, what matters is
not the relation between character and action but rather what goes before a
character. To contend that one is responsible for one’s character seems to imply that
we are either able to control how our character traits initially develop or, more
plausibly, that we have the power to revise these parts of our personality once they
are there. What this requires is explained by Arenella when he says that a character-
based theory “must presuppose that moral agents have some capacity for critical
self-reflection about those aspects of their character that make it difficult for them to
make the right moral choice. Moreover, moral agents must also have some modest
capacity for self-revision that permits them at least to modify the intensity of those
35
aversions and desires that impair their capacity to act like reasonable persons.” .
This, apparently, implies that a person is excused for a wrongdoing if his action was
motivated by a defective character trait but the person did not possess the capacity
for critical self-reflection or self-revision. However, it seems as if this view thereby
easily ends up faced with the same questions as those confronting the choice theory,
though the object of the choice is different. The degree to which a person is excused
would depend on his capacity to choose critical self-reflection and self-revision and
perhaps the lack of a fair chance for exercising this capacity. Thus, besides being
complicated by the fact that there are here two necessary objects of the choice, the
76 JESPER
THE ETHICS RYBERG
challenges which confront view (b) seem to equal those already outlined in relation
to the choice theory.
Now, what these different considerations on both the choice and the
character theory clearly indicate is that much clarification and theoretical work is
required in order to enable the proportionalist to meet the challenge of relative
comparison. However, suppose that a closer scrutiny would make it possible to
provide a plausible account of responsibility which would make clear what exactly
we should be looking for in order to determine whether the reduction of someone’s
culpability is larger in one case than in another when excuses are involved. This
would nevertheless still not answer the second and more serious challenge on how
much an excuse in a particular case diminishes culpability and thereby the
seriousness of the crime. As a simple example, suppose that we are comparing an
assault, a robbery and a minor theft and, suppose further, that the assault is more
harmful than the robbery which again is more harmful than the theft. If the person
who committed the assault acted under some degree of duress and if there are no
CRIME EXCUSE RANKING
assault duress ?
robbery
robbery no ? assault
theft
theft no ?
excusable circumstances related to the robbery or the theft, then, as the following
table indicates, it is simply not clear how the three crimes should be ranked
with regard to seriousness. Obviously this challenge of absolute comparison rises
independently of whether one believes that seriousness of crimes is measurable on
an ordinal or a cardinal scale. As long as one holds a view which requires some kind
of grading of crimes a theory will be required which can somehow meet the
challenge. However, on this point proportionalists have been remarkably silent.
In fact, the problem is even more complicated. In so far as proportionalists
accept that there are not one but several factors which affect the culpability of a
criminal, it needs to be explained how these factors should be combined. That is, it
will be necessary to indicate how culpability and hence seriousness are affected
when both different sorts of mens rea and different excuses are involved in criminal
behaviour. It is not hard to imagine an extended version of the above table in which
the compared crimes also differ in mens rea as, for instance, if some were done
recklessly while others where carried out intentionally. Given the complexity this
will contribute to the question of justifying one ranking rather than another, it seems
a little strange that, for instance, a careful thinker like von Hirsch repeatedly claims
that a “rulemaker should have no difficulty in scaling reckless conduct below
purposeful and in providing for reduced sanction for the partially coerced or the
36 37
provoked” or, more modestly, that it “should not be too difficult in principle” to
develop, for a sentencing doctrine, more refined distinctions concerning
purposefulness or carelessness and to develop theories of partial excuses in order to

determine the extent of an actor’s culpability. As the previous considerations have
indicated, the more reasonable conclusion seems to be that proportionalists in their
applications of the concepts of culpability as a determinant of crime seriousness are
theoretically on somewhat slippery ground.
3. RECIDIVISM
The final dimension that needs to be considered in order to provide an account of the
seriousness of criminal conduct is recidivism. The discussion on the significance of
recidivism differs, to some extent, from the discussion of the dimensions outlined in
the previous sections. While most proportionalists seem to accept that both harm and
culpability have an impact on the seriousness of a crime - though there are, of course,
disagreements when it comes to the more detailed discussions of these dimensions -
the views on recidivism differ more radically. Some proportionalists - such as Fletcher
and Singer - contend that prior record of the involvement in criminality should not be
38
considered at all. Whether a criminal has prior convictions should not, in their view,
affect judgement of the seriousness of the present crime. However, a number of other
adherents of proportionalism have defended the view that a prior criminal record does
39
enhance seriousness and, hence, does provide a basis for additional punitive severity.
This, of course, gives a reason for considering what challenges confront this
suggestion. Moreover, a further reason is provided by the fact that a prior criminal
record actually plays an important role in the sentencing systems which have adopted a
proportionalist rationale. One of the things which has been retained in the
transformation which several sentencing practices have undergone, from
rehabilitative systems to determinate sentencing systems, is the criminal’s prior
record as a factor in determination of a punishment. As mentioned earlier
proportionalist American penalty scales have taken the shape of a two-dimensional
matrix in which the vertical axis is the crime score indicating the seriousness of the
current crime, while the horizontal axis represents the number of previous
convictions. This raises the question of how such a practice can be justified. Thus,
has the repeater done something which, everything considered, is more serious than
what the first-time criminal has done, even if the current crime is in both cases the
same? Should there be what Fletcher calls a “recidivist premium”? And, in that
case, how should it affect the judgement of a crime’s seriousness?
While it is not hard to imagine a justification for letting prior criminal
record count if one holds a forward-oriented view on the justification of punishment,
it is at first sight less obvious that it matters from a backward-looking desert-based
point of view. After all, if a criminal has already received appropriate punishments
for his previous misdeeds it is not clear why this should affect the evaluation of a
current crime. On the other hand, this argument certainly does not a priori exclude a
prior criminal record as a factor affecting the seriousness of criminal conduct. What
proportionalists who defend the importance of prior record will have to assert is that
previous convictions affect either the harm or the culpability of the current crime or
that it in itself constitutes a further dimension contributing to the seriousness. The
78 JESPER
THE ETHICS RYBERG
first of the three options seems easily excludable. The harm caused to a present
victim is simply not enhanced by the harm that was caused to previous victims. This
leaves culpability and a further dimension. One of those who have persistently
defended the significance of prior criminal record is von Hirsch.40 In his first
defence von Hirsch contended that prior convictions do affect a criminal’s
culpability. Though he later rejected this position there is, nevertheless, a point in
shortly outlining the argument since I believe that it is relevant with regard to the
evaluation of the more recent defences. What von Hirsch claimed was that when the
first crime was committed the criminal was only one of a large audience to whom
the law impersonally addressed its prohibitions. However, once one is formally
censured for misconduct through punishment it is like “having one’s nose rubbed in
41
what one has done” which means that one is now fully aware of the prohibition
and therefore more culpable if the crime is repeated. As he himself later noticed, this
argument suffers from the flaw that, though some first-time criminals may be
ignorant in the relevant sense, there are certainly also some who are not ignorant.
Thus, the argument simply does not give support to the claim that first-timers should
always be regarded as less culpable.
In his most recent works von Hirsch has suggested another theory. This
time it is no longer the culpability that is affected. Rather, criminal record is a
further dimension which should be considered independently of harm and
culpability in the final evaluation of a crime. What he, along with Ashworth and
others, advocates is the theory of progressive loss of mitigation.42 Which means that
a first-time criminal should be given less than the full measure determined by harm
and culpability, while the repeater should receive the full deserved punishment. In
that sense, what is at stake is not a recidivist premium but a non-recidivist discount.
The question that needs to be answered, therefore, is why the criminal convicted for
the first time should receive a discount and why - and this is the second part of the
proposal - this discount is gradually lost when crimes are repeated. To the first
question von Hirsch says: “The respected process, on account of which the discount
is also granted, is that by which a person can attend the disapproval visited upon
him and alter his conduct accordingly. In viewing the person as a moral agent, we
initially assume him capable of such a response and thus give him his ‘second
43
chance’.” . Or as it is also put: “The first-offender discount reflects ... an ethical
judgment: it is a way of showing respect for any person’s capacity, as a moral agent,
44
for attending to the censure in punishment.” . This still leaves the question of why
the discount is given up after a number of repetitions. To this von Hirsch answers:
“It is because that respected process has not occurred. The person has chosen to
disregard the disapproval visited on him through his punishment, and thus seems not
45
to have made the requisite additional effort at self-restraint.” .
According to the first of von Hirsch’s arguments, the thought apparently is
that the discount is based on a certain respect. The object of the respect is a capacity
to reflect on the wrongness of one’s deed - as communicated to one through
punishment - and to modify one’s future actions accordingly by exercising a
sufficient amount of self-restraint. Why exactly this respect should be manifested
specifically in a non-recidivist discount is not further explained, though it can hardly

be regarded as self-evident. However, more important is that, if the view is that
respect for the mentioned capacity should result in a discount, then this apparently
has implications with regard to recidivists as well. After all, a person who has
several convictions might just as well possess this capacity. The question that needs
to be answered therefore is why a recidivist should not have the same discount. To
respond that the fact that a criminal has several prior convictions simply shows that
the person does not possess this capacity is not a plausible answer. If the recidivist
does not have the capacity then he probably did not possess it even after the first
conviction, which means that neither was he at that stage entitled to a discount.
Furthermore, von Hirsch seems to believe that every human being actually
possesses this capacity.46 Thus, in order to answer the question we will have to
confront the second of von Hirsch’s arguments outlined above.
According to this argument, the discount is lost after a number of
convictions because the criminal has chosen not to apply the capacity, including the
capability of self-restraint. If this argument is considered in isolation then it is not
sufficient to justify the discount theory, since one might just as well claim that the
first-time criminal neither chose to apply his capacity nor show self-restraint, and
that neither the recidivist nor the non-recidivist therefore should have a discount. To
respond that the first-time criminal was not fully aware of what was wrong, and that
therefore he did not have the same background for exercising self-restraint, would
lead directly back to von Hirsch’s above-mentioned culpability argument which he
himself rejected. Though this might be the case in some situations, there are
certainly also situations - as perhaps when the crime is murder or rape - where the
first-time criminal is fully aware of what is wrong and therefore ought to restrain
himself accordingly. Thus, in order to provide a justification of the view that there is
at first a discount which is lost when crimes are repeated, the two outlined
arguments must be combined. What we end up with is thus a moral principle which
amounts to something like the following: one ought to show a discount in response
to respect for people’s capacity to reflect on wrongs and to restrain themselves
accordingly, but the respect should not be maintained if people on a number of
occasions do not apply this capacity.
The question which a principle like this immediately gives rise to, of course
is, why should we accept it? Certainly the principle can hardly be claimed to be so
intuitively appealing or self-evident that it is not in need of justification. However,
von Hirsch does not provide any reasons that give this kind of support. Though the
outlined argument apparently is von Hirsch’s main argument in favour of the
discount view, he has also suggested some additional arguments. For instance, he
claims that, though a criminal should be blamed for his first misdeed, we should also
“accord him some respect for the fact that his inhibitions against wrongdoing have
functioned on previous occasions.”47. In so far as this is meant as an independent
reason for a non-recidivist discount, it certainly requires further elaboration. If the
point is that the first-time criminal should have a discount because he has previously
resisted temptations to commit crimes, then a general discount presupposes that
everyone has actually had these kind of temptations. Is that so obvious? If, on the
80 JESPER
THE ETHICS RYBERG
other hand, what should be respected is the mere fact that a person has not
previously committed crimes, even if the person has not had any temptations to do
so, is this really something that deserves a certain respect?
Moreover, it is not clear what precisely it is that should be respected. If it is
the number of times a person has had inhibitions against wrongdoing then this will
probably vary from one person to another, which apparently would imply that people
should have different discounts. If, instead, it is the mere fact that one has had
inhibitions then this might also be the case for the recidivist in between the crimes that
were committed: which would mean that recidivists should also have a discount. The
argument also presupposes that all those who have not had inhibitions against crimes
have actually been convicted. Otherwise, they would get a undeserved discount when
punished for the first time. But it is well known that not all criminals are convicted.
Finally, it is unclear why the proclaimed respect should manifest itself specifically in a
discount. Exactly as was the case in von Hirsch’s main argument the most interesting
premises here are missing.
The final argument von Hirsch has suggested is concerned with human
frailty.48 Against a background of prior compliance, a transgression should be
regarded as a lapse which should be judged less stringently than if the transgression
had occurred against a background of other transgressions. The view is that we
should “show some sympathy for the all-too-human frailty that can lead someone to
such a lapse”49. This is done by showing less disapproval for the first misdeed; that
is, by giving a discount. The argument does have some appeal. Talk about human
frailty, as something for which tolerance and understanding should be shown, does
seem to represent a specious point of view. However, as Durham has correctly
warned, it is easy to feel comfortable with a notion of “human frailty”.50 What
remains is the task of showing how this consideration can lead to a theory of
progressive loss of discount or to any other theory which implies that a first-time
criminal and a recidivist, where everything is considered, deserve different
punishments. Might one not agree that we should show sympathy for human frailty
but at the same time claim that recidivism is simply a result of frailty? It could even
be claimed that the fact that a person has committed several crimes strongly
indicates that he is even more frail than a person who commits only a single crime.
This would of course contradict the discount theory. Whether it is correct that only
the first few crimes can properly be regarded as a result of frailty or whether
repeated crimes might just as well - or even better - be a witness human frailty is a
question that can only be resolved by clarifying what “frailty” actually means.
However, on this point von Hirsch does not have much to offer. The closest he
comes to a suggestion is that frailty has to do with failing “in a moment of weakness
52
or wilfulness”51 or “exposure to pressures and temptations” . But this is obviously
not sufficient to solve the problem. Might one not several times perform misdeeds
due to weakness or temptations? As is the case with regard to any argument which
proclaims that the first-time criminal and the recidivist should be treated differently,
it has to be established that there is a morally relevant difference justifying the
unequal treatment. If it can at all be shown that frailty constitutes this kind of
difference, it certainly requires another and much more detailed analysis of what
frailty consists in than the answers indicated by von Hirsch. Without this kind of
analysis the frailty-argument is without the proclaimed strength.
As indicated, I do not find von Hirsch’s arguments convincing. Though he
apparently is the proportionalist who has done most to explain why recidivists
ceteris paribus deserve more severe punishments than first-time criminals, his
arguments do not place the view on a solid moral ground. However, the main
purpose here is not to thoroughly assess the outlined arguments but rather to
consider the implications when we turn to our cardinal question concerning the
53
construction of a crime scale. No matter whether one accepts von Hirsch’s theory
or any other theory which holds that prior criminal record should count, a scaling of
crimes in gravity require answers to a number of more specific questions. Both a
challenge of relative comparison and a challenge of absolute comparison can be
raised in several respects. Let us first consider the former kind of challenge.
The first question concerns the way the number of prior convictions should
affect the seriousness of a current crime. Should the fifth crime ceteris paribus be
regarded as being as serious as the tenth crime? According to a purely accumulative
point of view, there should be no upper limit to what is deserved. The larger the
number of prior convictions is the more serious the current crime becomes.
However, the view which is apparently preferred not only by von Hirsch but also by
other proportionalists is, as we have seen, a principle of progressive loss of
mitigation. But this means that in order to construct a scaling of crimes it needs to
be determined at which number of convictions the discount should be fully
exhausted. A rationale which could give an indication of whether one should get the
full measure after three, eight or fifteen convictions has not even tentatively been
54
developed.
The second way in which a theory of recidivism faces the challenge of
relative ranking is with regard to the question of whether the time that has passed
since the previous conviction(s) should be taken into account. If a person has
committed a current assault, does it then make any difference whether the previous
assault the person was convicted for occurred a month or ten years ago? According
to von Hirsch, the temporal span should make a difference. As he says: “the longer
the stretch of time prior to the current act during which the defendant has led a law-
abiding life, the less plausible it becomes to claim that the current misdeed is,
indeed, typical or characteristic of the way he has been behaving”55. As was the case
with regard to the concept of “frailty”, the concept - as well as the moral
significance - of an act being “characteristic” needs clarification. Almost no matter
how the concept is defined this suggestion does, however, bring in other factors
beyond the temporal distance. For example, some crimes require special
circumstances in which to be performed. Circumstances which are not always
present. For instance, perjury certainly requires very special conditions. But that
means that whether something can reasonably be regarded as characteristic does not
necessarily depend on the time span, but rather perhaps on the number of
opportunities one has had to perform the crime in question. To claim that whether
something should be regarded as characteristic of a person’s behaviour should only
depend on the time that has passed, seems arbitrary. Bringing in the different aspects
that may have an impact on whether an act is characteristic certainly complicates the
judgements. More generally, adherents of the view that the time that has past since
the previous convictions does matter, will have to clarify how it affects seriousness.
It is hardly defensible to claim that there is a certain number of years or months
beyond which prior convictions suddenly stop counting. But, alternatively, if one
claims that the significance of prior convictions diminishes with the temporal
distance, then it needs to be explained how this more precisely functions. Does the
time span have a large or only a minor impact on the role that the prior criminal
record plays? With regard to this question neither von Hirsch nor other
proportionalists have had anything to offer.
A third important question that must be considered, if a prior criminal
record is regarded as important, is how the seriousness of the previous crimes
affects the evaluation of a current misdeed. Does it make any difference to the
seriousness of a current assault whether the recidivist’s prior conviction is for
murder or illegal possession of drugs? In von Hirsch’s view the answer should be in
the affirmative. It does make a difference whether the previous crime and the
current crime are very different in kind or whether they to some extent belong to the
same category. In the latter case the prior conviction is more important. It is worth
noticing, though, that von Hirsch understands a “category “ in a broad sense such
56
as, for instance, “intentional victimizing crimes” . In so far as one accepts that the
seriousness of previous crimes is significant the more general problem, of course, is
to specify how exactly the seriousness of these crimes should affect the computation
of current desert. Obviously the problem is not simply that of suggesting some kind
of metric but, more fundamentally, to provide good reasons as to why an answer
should point in one direction rather than another. No such reasons have been
provided.
The lack of theoretical underpinning is no less if we turn the focus to the
challenge of absolute comparison, that is, to the question of how a prior criminal
record scores on the crime scale. A crucial question, of course, is how large the
discount on the first crime should be, compared to the full measure of the crime in
question (or alternatively, how large recidivist premiums should be if one defends a
principle of progressive gain in aggravation). If the change from the first crime to
the crimes where mitigation is lost should be a gradual change, the question can be
repeated for each of the number of recorded crimes until a ceiling has been reached.
Thus, it needs to be answered how we should, with regard to seriousness, compare a
murder without prior convictions to an assault with three prior convictions or a theft
with six prior convictions. The mere claim that discounts should be large or small is
not in itself interesting. What would be interesting, of course, would be the reasons
that could be given for claiming the one rather than the other. With regard to this
challenge, as well as to the former one, no convincing answers have been given.
All in all, it does not seem premature to conclude on the ground of the
previous considerations that, in so far as proportionalists believe that a prior
criminal record should be a further dimension which matters with regard to the
scaling of crimes and thus in the computation of desert, the principles that underlie
this dimension are - if not defective - theoretically under-determined.
4. PROPORTIONALIST ANSWERS
What has been indicated in the previous sections is that the project of ranking or
comparing crimes in seriousness confronts the proportionalist with a large number of
difficulties. The problems occur at several levels. One category of problems relates to
the clarification of each of the dimensions on which seriousness varies. The properties
which are determinant with regard to the computation of seriousness are so little
clarified that there is often no basis for judging whether they are more or less present
in different crimes. That is, it is not possible to establish the different degrees of the
seriousness-generating properties within each dimension. Though the harm-dimension
in this respect is probably the least problematic there nevertheless are problems, for
instance, when it comes to the comparison of crimes involving risked harms. The
problems are even more significant with regard to the culpability. This dimension
raises both the problem of assessing different sorts of mens rea and of indicating
whether some excuses are more or less extenuating than others and, furthermore, the
problem of combining varying degrees of mens rea and excuses into different degrees
of culpability. Finally, corresponding problems exist with regard to the significance of
prior criminal record when it comes to the factors determining degrees of either non-
recidivist discounts or recidivist premiums.
Even if all these problems were solved, one would still be left with the
second category of problems concerning the way the different dimensions should be
combined. That is, it needs to be indicated how different degrees of harm, culpability
and - in so far as it is regarded as relevant - prior criminal record should be worked
together in a final judgement of the seriousness of a particular crime in comparison to
other crimes. This problem is perhaps even more complex than what should be
expected from the foregoing discussion, since it includes - at least according to some
adherents of proportionalism - even further aspects than the ones hitherto outlined. An
example is multiplicity of crimes. That is, cases in which a person has committed
several crimes before he is convicted and punished. The obvious question is how such
cases should be assessed from a proportionalist point of view. The easy answer, of
course, is to add up the seriousness of each of the crimes committed and thereby get a
final judgement of the seriousness of the total number of misdeeds. However, several
proportionalists regard this is unacceptable. The feeling is that a number of minor
crimes cannot add up to a very serious crime.57 However, if one believes that there
should be some kind of discount for bulk offending then it needs to be argued why,
and it has to be specified how exactly this discount works. As we have seen in the case
of recidivism, a discount theory raises several difficult questions. However, even if we
ignore this more tentatively discussed - but certainly practically important - issue, the
question of how the outlined dimensions of seriousness should be combined does itself
constitute a genuine problem.
The problem is put into even more perspective if we briefly touch upon a
question which has not yet been considered, namely, what kind of scales are at stake in
the measurement? Suppose it is claimed that the dimensions of harm, culpability
and prior criminal record are each measurable only on ordinal scales. In such a case,
it is certainly hard to imagine how these dimensions can, in a non-arbitrary way, be
84 JESPER
THE ETHICS RYBERG
worked together into a scaling of crimes. Suppose alternatively, that the dimensions
are measurable on ratio scales. In such a case it might seem more straightforward to
construct a ranking of crimes. At best, one might hope it could be done - as devised
by Nozick’s formula - by multiplying the degrees of harmfulness, culpability and
recidivism. However, what is perhaps here gained in simplicity with regard to the
jump from each of the dimensions to the judgement of the seriousness of a crime, is
now lost with regard to the measurement within each dimension. That one should be
able to establish not only that a criminal is, for instance, more culpable if he
performs a wrong intentionally, though partly excused, than if he does it recklessly,
and unexcused, but also how much more culpable that criminal is in the former case
compared to the latter, certainly throws the outlined difficulties with regard to
clarification of each dimension into relief.
That there are problems in the comparison of crimes is obviously not
something that proportionalists have left unnoticed. Hart clearly recognized the
issues when he asked if ”negligently causing the destruction of a city [is] worse than
58
the intentional wounding of a single policeman?” The fact that a solution has
nevertheless not been suggested, combined with the other fact that some sort of
comparison is a sine qua non for proportionalist theories, naturally raises the
question as to whether there are ways to get around the problem. Or more strongly
put, whether the way the problem has been posed until now rests on a
misunderstanding. If one confronts the literature there are certain claims that point
in this direction though, however, not in a very convincing way.
Scheid apparently believes that some misunderstandings have inflicted the
discussion. Though he does not contend that the problem of ranking crimes is
resolved, he nevertheless believes, more modestly, that “some confusion on this
topic could be avoided ... if a distinction between the legislative and judicial tasks
were kept in mind”59. The point is that the legislative task is to assign different
punishments to different types of crimes, while the judicial task is to sentence
individual criminals for particular instances of crimes. All the legislator needs to
do is to assume some standard level of culpability and then consider standard cases
of each type of crime. Thus, when considering the ranking, “the question of how
60
harm and culpability should be combined is not a concern for the legislative task” .
Scheid may be right in his opinion that, in so far as some believe that the ranking of
crimes with all possible combinations of harm and culpability should take place at
the legislative level, confusion has inflicted the discussion. However, with regard to
the main question of how culpability and harm and other possible dimensions
should be combined, it is hard to see that the distinction between legislative and
judicial tasks has anything substantive to offer. Even if it is correct that the
combinationproblem does not arise at the legislative level, it still - as Scheid is
obviously well aware - exists at the judicial stage at which different degrees of
culpability should be taken into account. In so far as the arguments given in favour
of proportionalism - or any other thinkable arguments - really establish that it at
least cannot be justified to punish a criminal more severely than the seriousness of
the crime warrants, the ethical problem of comparing crimes in gravity remains
intact independently of the division of labour within a sentencing system.
Another comment on the ranking problem is provided by Ten. He believes

that the construction of a scale of crimes is “a project that seems capable of being
carried out” and he supports this claim by reminding his readers of an analogous
case. In his view, the problems faced by proportionalism in comparing crimes in
gravity are not very different from those confronting teachers in ranking essays:
“When tutors and teachers rank the essays of their students, they do not have only
one relevant feature to look for. There are a number of different features -
originality, understanding of the issues discussed, lucidity of presentation, etc. -
which each makes a contribution to the quality of the essay. An essay may be strong
in one dimension but weak in another, and yet it is possible to make an overall
assessment of the essay as being better or worse than another”61. Even though Ten’s
claim certainly has some appeal, I must admit that I am not quite sure what it
manages to establish. If one merely contends that we are now and then faced with
cases which require some sort of ranking but where the decisions we make are not
supported by good reasons, that is, they are totally arbitrary, then that is probably
correct. But this is hardly relevant. No one would deny that sentencing commissions
or judges can make decisions on the relative gravity of crimes. However, the
interesting question obviously is what these decisions should be like if we do not
applaud arbitrariness. Therefore, if Ten’s point is that even though teachers do not
possess a very strict or simple underlying rationale in the evaluation of essays their
judgements are nevertheless qualified - in the sense that they express certain shared
intellectual values with regard to what should count in the evaluation and with
regard to the extent to which different intellectual requirements should be satisfied
and perhaps be weighed against each other - then I believe that Ten is right (indeed I
hope he is!). However, this does not answer the question as to what in the case of
crime ranking corresponds to the rationale that makes the teacher’s judgement
qualified. That is, we are still left with the open question of how the comparison of
crimes in gravity should in the end be determined if we wish to avoid mere
arbitrariness in cases like those illustrated in the earlier sketched tables.62 Of course,
Ten’s point might simply be that there is nothing that prevents the possibility of
there being some rationale for a scaling of crime just as is the case with regard to
essays. In that case the claim is correct. There is nothing in what we have seen until
now which a priori excludes this possibility. The point rather is that no rationales
have been suggested and that it is indeed hard to imagine what a plausible rationale
should look like.
However, there is still a possible way to answer this problem. This would
be by denying the continued request for a justification. Such an answer, which
perhaps constitutes the most attractive response to the ranking-problem, can be
summarized in the following manner. The main problem in the discussion has been
to provide some kind of justified rationale underlying the way the different
determinants of seriousness should be combined. But is this not to ask for too much?
Rather than requiring something like a justified metric, would it not be plausible,
once a sufficient level has been reached in the reasoning process, to base the finer
details on some kind of general opinions or intuitions? Isn’t it reasonable to admit
that intuitions must at some stage enter the considerations and to hold that, once it
86 JESPER
THE ETHICS RYBERG
has been argued that at least harm and culpability should count, the problem of
combining the dimensions must be solved by what could be called our “common
sense”, that is, our more intuitive judgements on what is reasonable? And does the
way the challenge of absolute ranking has been presented therefore not rest on a
misunderstanding? Does the whole problem not simply arise as a result of the
excessive interest of philosophers in searching for well-argued rationales in places
where the appropriate attitude would be to listen to what we feel?
Though proportionalists have often not been very clear in their discussion
of this issue, there are at least some claims which point in this direction. The most
explicit formulation of the view is given perhaps by Primoratz who, after having
explained that harm and culpability matter, specifically introduces the “society’s
63
evaluations” in the comparison of crimes. The question that remains is only
whether this should be done by using the technique described by Sellin and
Wolfgang or by some other procedure. It is from this point of view Primoratz
contends - as quoted at the beginning of this chapter - that all we are left with is a
technical, not a philosophical question. An alternative position, rather than including
the society’s evaluation, would be to base the final scaling of crimes on the
judgements of sentencing commissions or to regard the final weighing as a purely
judicial decision.
Whether we should be satisfied with this kind of answer is a question
which to some extent touches upon some of the most complicated discussions
within ethics concerning the basic question of what we should in the end expect
from an ethical theory and what role intuitions should play with regard to the
assessment of a theory. I shall not here enter a discussion of these questions but
merely indicate why I am sceptical with regard to the outlined way of getting around
the problems.
Firstly, it not clear what it exactly means to hold that we should rely on
intuitive judgements in the weighing of harm, culpability and respects to criminal
record. If the contention is that we should follow our intuitive judgements with
regard to what constitutes a reasonable general weighing principle, then it is far
from obvious that we have this sort of intuition. If one asks oneself that what one
believes in general is an acceptable way of balancing the different determinants,
then I simply do not believe that we have very clear or shared intuitions. If, on the
other hand, the suggestion is that we should rely on our more intuitive judgements
with regard to whether one crime seems more serious than another, then it is surely
much more reasonable to assume that we actually (at least in some cases) have
certain intuitions. But would it be plausible to simply reconstruct the weighing of
harm, culpability and prior criminal record in a way that simply matches such
intuitions on comparative gravity, even if this would imply that the three
determinants of seriousness are given totally different weight when we compare one
set of crimes than if we compare another? In my view this would be extremely
arbitrary and there is, I believe, no guarantee than our overall judgements of crime
gravity will happen to follow a consistent weighing principle.
Secondly, though we certainly do have intuitions when it comes to the
comparison of, for instance, grave assault and minor theft, it is much less obvious
that we have clear intuitions if comparison is made between crimes which scores
differently in harm, mens rea, responsibility and prior record. In such cases, some
sort of guidance would be valuable.
Thirdly, and perhaps most importantly, it is questionable whether we
should in the end be satisfied with a theory which in itself has nothing to offer with
regard to a weighing of the different determinants. The significance of a lack of a
rationale is obviously best demonstrated by some radical examples. Consider, for
instance, a reckless killing of several persons and an intentional theft of goods worth
10$ from a shop. The one crime scores more in harmfulness, the other more in mens
rea. Which should be considered the more serious? Obviously, in this case most
people would surely consider the reckless killing the more serious of the crimes. But
the point is that proportionalism, without a more precise answer to the challenge of
absolute comparison, can provide no good reasons in support of this ranking. If one
regards explanatory power as a theoretical virtue then proportionalism does not - as
a theory of principled sentencing - in this respect reach a high score. Thus, though
these considerations obviously lead into basic methodological considerations to
which, admittedly, we cannot claim to be on firm ground, it seems to me that neither
this final answer nor those considered above have satisfactorily managed to answer
the - both theoretically and practically - important question of the comparison of
crimes in seriousness.
5. A FAIRNESS-THEORETIC APPROACH
Those adherents of proportionalism who have done most to answer the question of
how different crimes should be ranked in terms of seriousness usually defend a
harm-theoretic approach to the question. On this point, however, there is a marked
exception. One of the theorists who has most persistently defended a fairness
theoretic approach, namely, Michael Davis, has suggested an alternative method for
the ranking of crimes. A method which in fact handles some of the more
complicated questions which the ranking-problem gives rise to, such as how
attempts or strict liability crimes should be ranked and how recidivism should affect
the determination of punishment. Moreover, the method is characterized by
apparently being relatively simple in practical application, a feature which, in the
light of the complexity of the harm-theoretical ranking procedure, is indeed
remarkable. This in itself gives a good reason for taking a closer look at Davis’
theory.
The central claim of the fairness theory, as we have seen, is that a criminal
by breaking the law gains a certain unfair advantage over law-abiding citizens. It is
this advantage the criminal law is supposed to remove or nullify by punishing the
criminal for his misdeed. The larger the unfair advantage, the more serious is the
crime and the more severe the punishment that is required to take back the
advantage. In principle, the fairness theory therefore provides a simple answer. All
one has to do to compare crimes in gravity is to provide a method for measuring the
unfair advantages gained by different crimes. Though this might seem a manageable
task and though Davis’ final procedure is simple, the argument which sustains this
procedure has a complicated structure. In short, the argument amounts to something
like the following.
88 JESPER
THE ETHICS RYBERG
Davis sets out by suggesting a seven-step procedure for the assignment of

punishments of crimes. In short, the procedure prescribes that crimes be ranked
according to what people would mostly fear would happen to themselves. Despite
the complications, to which we shall return shortly, this seems like a simple method
for a ranking. Since people usually prefer a theft to a murder, the murder should be
ranked higher on the scale. However, the obvious question is, why should crimes be
ranked thus? What is needed is a justification, or more precisely, it needs to be
established whether this has anything to do with unfair advantages. In order to do
this, Davis firstly suggests an auction model which in principle allows us to gauge
unfair advantages. Since this model is not thought of as a real possibility but rather
as a “heuristic devise”, the next step is to show that the results gained by the auction
model are the same as those produced by the easily applicable seven-step procedure.
In short, since the auction model measures unfair advantages and since there is a
structural isomorphy between what follows from the auction model and the simple
ranking procedure on the basis of fear, all we will have to do in practice is to apply
the latter procedure. The structure in this piece of reasoning raises three questions
each of which I shall consider in turn. Firstly, what do the prescriptions of the
seven-step produced precisely amount to? Secondly, in what way does the auction
model gauge unfair advantages of crimes? And thirdly, is it correct that the seven-
step procedure and the auction model produce the same results?
The seven-step procedure is a descendant of the method that was originally
proposed by Mabbott. The procedure is summarized by Davis in the following way:
1. Prepare a list of penalties consisting of those evils (a) which no rational person
would risk except for some substantial benefit and (b) which may be inflicted
through the procedures of the criminal law.
2. Strike from the list all inhumane penalties.
3. Type the remaining penalties, rank them within each type and then combine
rankings into a scale.
4. List all crimes.
5. Type the crimes, rank them within each type, and then combine rankings into a
scale.
6. Connect the greatest penalty with the greatest crime, the least penalty with the
least crime, and the rest accordingly.
7. Thereafter: type and grade new penalties as in step 2 and new crimes as in step 4,
64
and then proceed as above.
Thus, the procedure not only concerns crimes but also the scaling of punishments
and the way the two scales should be anchored. However, in the present context the
interesting part of the conjecture is step 5. In a detailed explanation following each
step, we are told that crimes are typed by “the minimum object they would normally
have in view”65. For instance, theft and blackmail belong to the same type because
the minimal aim is the same in both crimes: to get another’s property. The only
reason Davis gives as to why crimes should be grouped in this way is that the
potential criminal will be provided with a reason to “choose the lesser crime rather
than the greater when he chooses his crime”66. Thus, the typing seems more like a
question of practical design than something which in itself is vital to the ranking.67
Once the crimes have been typed, each of the crimes within a group should be
ranked. This is done by placing lowest in the list the crime most people would
prefer to happen to themselves (or someone or something they care about) if forced
to choose between that and any other crime of that type.68 Finally, each of the types
should be connected into an ordinal scale which Davis claims would resemble
something like a map of a complex subway system (where crimes correspond to
stops, and types correspond to lines).
Though this procedure is supposed to be a quick way to achieve results of
the ranking in terms of unfair advantages, the method is not, however, as
straightforward as it might at first appear. For instance, what precisely is meant by
the claim that crimes should be ranked according to what people would prefer to
risk given a choice between different crimes within a type? As Davis points out, the
thought is not that what is feared is states of affairs as such (e.g. death or loss of
property) but acts (e.g. being intentionally killed or deprived of property). However,
it is not obvious what this implies. For instance, what exactly is it that one should
consider with regard to reckless driving? It is hardly how much one usually fears to
be killed or injured by a reckless driver. Since this would be dependent on the
likelihood of the crime, it would apparently imply that a minor theft should be
regarded as more serious than a serious blackmailing because most people regard it
as much more likely that they will be subjected to a minor theft than that they will
be blackmailed. Similarly, most people probably have a greater fear of being the
victim of a minor assault than of a special kind of torture, simply because the former
calamity is much more likely to occur. On the other hand, what one should consider
cannot be how much one would fear if one was actually killed or injured by a
reckless driver. That would imply that there would be no difference between the
reckless driving of a car and recklessly riding a bike. If I were to be actually killed
or injured I would not care whether this event were caused by a bike or a car.
However, Davis would probably agree that the reckless driving of a car is more
serious than reckless riding a bike. But in that case, what is it that counts when one
considers the ranking?
Another problem concerns that fact that it is not all crimes that everyone
can become the victim of (e.g. a blind person cannot be blinded). In order to make
up for this, Davis adds that one should consider how much one fears each crime
being committed against oneself or someone or something one cares about.
However, this additional plea is not sufficient to account for the fact that there are
crimes which lack an easily identifiable victim. What would the procedure prescribe
with regard to, for example, tax fraud, espionage, bribery, or perjury? Davis
apparently believes that we should also include in our considerations how much we
69
fear crimes that happen to “a government we care about” . But what exactly does
this imply? Should I consider how much I fear that I myself would suffer from
administration by corrupt officials or should I consider how much I fear corruption
which happens in a state I care about? That is, should one consider how much one
fears crimes that happens to a state or how much one fears oneself suffering from
90 JESPER
THE ETHICS RYBERG
crimes against a government? What these problems illustrate is that, though the
seven-step method is supposed to be the quick route to a ranking, the instructions
given in the procedure are not sufficiently clear.
The seven-step procedure considers the seriousness of crimes in terms of
how much different crimes are feared. However, as mentioned, it says nothing about
the crucial concept of an unfair advantage. The second question Davis’ argument
raises is on how the actual determinant of seriousness, namely, the size of the gained
unfair advantage, should be measured. As we have seen in chapter 1, there has been
some discussion on the question of what exactly an unfair advantage consists in.
70
Davis’ view is that an unfair advantage is a “cheater’s advantage” . That is, the
advantage one gains in comparison to others by having improved one’s chance of
obtaining something that is valued. As mentioned, Davis sets up an auction model to
measure such unfair advantages. What we are to imagine is a government that sells
a limited number of licences to commit specified crimes, that is, a kind of pardons-
in-advance to would-be criminals. The thought is that the prices at which the crime
licenses are sold provide an index of the value of the unfair advantages a criminal
takes by committing the related crime. However, this presupposes that certain
conditions are fulfilled.
For instance, the number of licenses would have to be limited. The
limitations would be determined by the amount of each sort of crime the society
would be willing to tolerate. For example, we are told that the society might offer
only 1000 robbery licenses each week but 10000 burglary licenses. Another
precondition is related to the discussion of what Davis calls “poaching”. A question
which naturally rises in relation to the auction model is why should anyone be
expected to buy a license? If a certain conduct is not already illegal and therefore
has some punishment attached to it, it is obvious that no one could be expected to
spend money on a license. It therefore seems that the auction model, in order to
provide a ranking of crimes in gravity, must already presuppose the existence of a
punishment for each and every crime to motivate the bidding at the auction in the
first place. Some critics believe that this renders Davis’ entire auction model
incoherent from the very beginning. However, Davis is aware of the problem but
thinks that it can be solved by adding some assumptions on poaching (crimes
committed without license). In order to avoid prices at the auction being biased by a
pre-auction view on the seriousness of different crimes, Davis suggests that either a)
all instances of poaching receive the very same severe punishment; b) that poaching
is impossible; or c) that no one ever poaches.71 Critics have remarked that some of
these assumptions are wildly unrealistic.72 However, since Davis repeatedly
underlines that the auction model is not intended as a real possibility but as a
hypothetical model, I do not believe that this is a problem. There is, however,
another problem related to two of the three possible assumptions, namely, that they
are inconsistent with the way Davis holds that the measurement of the unfair
advantages of attempts can be coped with by his auction model.
In the same way as one can bid on a license to get away with a crime if it
succeeds, the thought is that one can just as well bid on a license that will pardon a
crime that does not succeed. In other words, one can get away with an attempt if one
has bought a license to fail. Davis imagines that the thoughts going through the head
of a would-be criminal bidding on this kind of license might be something like: “if I
73
fail I am safe ... and I am willing to take the chance if I succeed” . He also believes
that the licences to fail will fetch a lower price at the auction than the licenses to
succeed with regard to a crime, which means that attempts should be regarded as
less serious than completed crimes. However, if a person only has a license to fail
but no license to succeed with regard to a certain crime, but nevertheless succeeds in
performing the crime then that would be an instance of poaching. But then it seems
that, in order to expect that anyone will bid on a license to fail, it would have to be
possible to poach because, as Davis himself is aware, no one would try to commit a
crime with the purpose of failing. If, as Davis suggests, the bidders are to be found
amongst those who are not able to get a licence to succeed, then poaching must be
possible, because otherwise it seems that no one would bid on these licenses, which
is tantamount to claiming that attempts do not deserve a punishment at all (a
74
possibility Davis himself regards as implausible). Thus, the assumption which best
fits within Davis’ theory of attempts is a), that poaching is possible but should be
punished with a very severe punishment. But even this has its problems. For
instance, it may make it very unlikely that anyone will actually bid on a licence to
attempts with regard to certain (minor) crimes. So much for the poaching
assumption.
The final assumption Davis makes, which underlines the hypothetical
character of the model, is that there is a wide distribution of wealth amongst
members of the society in which licenses are sold. This is to make certain that
anyone who does not get a license has lost it in a fair competition.
Now, with all these somewhat complex assumptions settled what does the
auction model imply with regard to the ranking of crimes? As mentioned, Davis’
contention is that the way crimes will be ranked according to prices at the auction
corresponds to the ranking reached by the instruction for crime-scaling in the seven-
step procedure. In other words, all we have to do to measure the unfair advantages
gained by different crimes is to rank crimes according to how much most people
fear to risk them. But are there reasons to believe that the auction model and the
seven-step procedure are in this respect equivalent? Davis presents two reasons in
favour of this crucial part of his view: “First, the quantity of licenses would have to
decrease as the seriousness of the crime licensed increased... Second, the demand
for licenses is likely to increase with the seriousness of the crime. (If that seems
unlikely given moral constraints on potential buyers, ask yourself whether you
75
would prefer to have a license to steal or a license to jaywalk.)” .
According to Davis’ first argument, the supply of different crimes should,
as we have seen, be determined by what would be socially tolerable. In his
discussion of poaching Davis rightly admits that it would be begging the question to
punish different kinds of poaching differently, that is, in a way that already reflects
a certain view on which some crimes are more serious than others. But is it not
equally question-begging to assume that there are differences in the supply of
licenses of different crimes? Scheid has criticized Davis’ model on exactly this point
by claiming that differences in the supply of different crimes imply that the prices of
the licenses will reflect a pre-auction notion of seriousness of different crimes rather
76
than the unfair advantages gained by the crimes. Davis, in his answer to Scheid,
has maintained that there should be such restrictions on the supply of different crime
licenses. We need not here dig deeper into who is right in this complicated
discussion, because even if we accept the restrictions on supply this does not
support the alleged correspondence between the results of the two methods. The
question is: what does it mean that the supply of licenses should be determined by
the amount of a certain crime the society is willing to tolerate? Davis claims that this
willingness depends on at least two factors. On the amount of income that licenses
will produce for society: the greater the better. And on how much people fear the
different crimes. Exactly how these factors should be balanced is not clear.
However, suppose that people’s fear would imply, as Davis himself suggests, that
the supply of robbery licenses is smaller than the supply of burglary licenses, or that
the supply of licences to harsh violent crimes is smaller than the supply to minor
violent crimes, does it then follow that a robbery license will fetch a higher price
than a burglary license or that the price of a licence to a harsher violent crime will
be higher than the price of the license to a minor violent crime? No, obviously not.
The prices will depend upon the demand on each of the different kind of license. If
there is a smaller demand on robbery licenses than on licenses to burglary then the
former licenses might fetch a lower price at the auction even though the supply of
these licenses is smaller. Similarly, the prices might, despite the differences in
supply, be higher on the licences to the less violent crime than on the licenses to the
more violent crime, if simply fewer people are interested in the latter sort of license.
Therefore, even if we accept the restrictions on supply it does not follow that the
ranking by prices corresponds to a ranking according to what most people fear.
The second argument Davis gives is concerned exactly with the question of
the demands on different licenses. Since the first argument, as we have just seen,
presupposes a certain assumption on the demand, it seems that much hangs on the
shoulders of this second argument. What Davis claims is that the demand will
increase with the seriousness of the crime and, to convince his readers, he asks
whether one would prefer a license to steal or a license to jaywalk, assuming that
the former would clearly be preferable. However, this argument has rightly been
77
criticized as conspicuously unpersuasive. Even if it is correct that the license to
steal would be preferred to the one to jaywalk there are certainly numerous cases
where there is not the same relation between seriousness and what would be
preferable. Considering whether one would prefer a license to illegal car parking or
to torture someone, to tax evasion or to incest, to jaywalking or to commit murder, I
certainly believe that that the former licenses in each pair would be preferable. Of
course, Davis’ claim is not that all people would pay more for the license to the
more serious crimes, but it certainly seems reasonable to expect that even if some
would bid on licenses to torture, incest, or murder the demand on these licenses
would be much smaller than the demand on some less serious crimes. But this is
sufficient to undermine Davis’ claim of the equivalence between the auction model
78
and the seven-step procedure.
As we have seen, it is not in itself clear what exactly the seven-step

procedure would imply with regard to crime ranking. What we have now seen is
that there is not even a reason to believe that the seven-step procedure and the
auction model are in this respect equivalent. But is that devastating for Davis’
theory? Could he not simply maintain that it is the auction model that measures
unfair advantages and then drop step 5 in the seven-step procedure? The answer is
that this would in fact be devastating. As we have seen, the auction model was not
meant as a “real possibility” but as a hypothetical model and, since we do not know
what the relative demand on the different crime licenses would be if we merely tried
to imagine what the model would imply, we would simply not be able to construct
the ranking of crimes. There would be no way to meet a challenge of relative
ranking. Thus, in the end, I believe that Davis, as the only adherent of the fairness
theory to have thoroughly considered the seriousness of crimes, does not have much
to contribute to the proportionalist discussion of how crimes should be scaled.
6. CONCLUSION
That there are problems related to the proportionalist view that crimes should be
ranked in terms of seriousness is not a point that has been left unnoticed by the early
79
critics of different versions of proportionalism. However, what the present chapter
has revealed, in my view, is that despite the significant increase of the interest in
proportionalism over the latest decades, proportionalists are still far from having
provided an adequate comparison or scaling of crimes in gravity. The harm-theorists
who regard harm and culpability and perhaps a prior criminal record as dimensions
determining seriousness have not provided a sufficient background for making
judgements on whether one crime scores more within one dimension than another
crime. And there is a genuine problem related to the question of how the seriousness
determining dimensions should be combined into a final computation of
seriousness. It is important to notice that, contrary to the impression one might get
by the often repeated claim that “one should not expect full precision”, these
problems are not epistemological. That is, it is not simply a matter of sometimes
being without the means to measure the precise degree of seriousness - or, as
Beccaria once put it, that “[t]he gravity of sin depends upon the inscrutable
wickedness of the heart. No finite being can know it without revelation. How then
80
can it furnish a standard for the punishment of crimes?” - though this, of course,
might also be a problem. Rather is it that the theoretical ground for judgements on
seriousness is, to a wide extent, missing.
Von Hirsch has, in one of his early writings, rejected a theory of
punishment which claims that one should simply combine diverse considerations
(rehabilitation, predictive restraint, deterrence, and desert), on the ground that, when
different objectives conflict, the theory would not “offer a principled way of
81
resolving the issue” . This objection is certainly reasonable. But, as we have
learned, it is basically the same kind of problem that proportionalists are faced with
in the comparison of crime gravity. And, as we have also seen, the possible attempts
to explain away the lacunas in the theory are not convincing. Furthermore, it was
clear that the attempt to provide a ranking of crimes within a fairness theoretic
framework was not successful. Davis’s model was, despite its promises, unable to
provide any real guidance. Together, these conclusions are, of course, relevant in a
philosophical discussion of proportionalism. Moreover, despite the fact that we have
not yet considered the scaling of punishments or the anchoring problem, they are
sufficient to establish that those legal systems which are based on proportionalism
cannot proclaim that their practice is justified, since it is theoretically unclear what
this would imply.
NOTES
1See, for instance, A. Ashworth, Principles of Criminal Law, Clarendon Press, Oxford, 1995, p. 35f.
2I. Primoratz, “On Retributivism and the Lex Talionis”, Rivista Internazionale di Filosofia del Diritto, vol.
61, 1984, p. 89.
3See, for instance, R. Sparks, H. Genn & D. J. Dodd, Surveying Victims, Wiley, 1977.
4See A. von Hirsch, K. A. Knapp & M. Tonry, The Sentencing Commission and Its Guidelines, Northeastern
University Press, Boston, 1987.
5For instance, Kleinig claims that the “disapproval which we naturally show towards wrongdoing is not
always appropriate or well-grounded ...”. J. Kleining, Punishment and Desert, Martinus Nijhoff, The Hague,
1973 p. 126. And von Hirsch declares that he does “not think ... that ratings of seriousness for sentencing can
simply be derived, without further analysis, from such surveys”. A. von Hirsch, Past or Future Crimes,
Rutgers University Press, New Jersey, 1985 p. 65. For critical comments on this kind of survey see, for
instance, A. Ashworth, Sentencing and Penal Policy, Weidenfeld and Nicolson, London, 1983, p. 198ff; or A.
Ashworth, Principles of Criminal Law, Clarendon Press, Oxford, 1995, p. 36f. In a more recent survey in
which Robinson and Darley compare “community views” on different aspects of crimes within the criminal
law, the authors claim that desert theorists might make use of surveys in the weak sense that if there is a wide
disagreement between what theorists and the community regard as just, then this might suggest a closer
scrutiny of the theoretical reasoning; P. H. Robinson and J. M. Darley, Justice, Liability & Blame, Westview
Press, USA, 1995, p.6. But even if this sounds plausible it obviously does not imply the lack of need for
theoretical considerations.
6A. von Hirsch & N. Jareborg, “Gauging Criminal Harm: A Living-Standard Analysis”, Oxford Journal of
Legal Studies, vol. 11 no. 1, 1991. For a summary of the main points in this article, see A. von Hirsch,
“Seriousness, Severity and the Living Standard”, in A. von Hirsch & A. Ashworth, Principled Sentencing,
Hart Publishing, Oxford, 1998.
7A. Ashworth, The Principles of Criminal Law, Clarendon Press, Oxford, 1995, p. 37.
8Von Hirsch and Jareborg mention that there is also another factor which has an impact on the size of a
premium, namely, the degree to which interest dimension conceptually overlaps. The thought is, for instance,
that humiliation and loss of privacy are more closely related than, say, humiliation and physical harm. The
latter combination of affected interests will therefore result in a larger premium than the former combination.
See A. von Hirsch and N. Jareborg, “Gauging Criminal Harm: A Living-Standard Analysis”, Oxford Journal
of Legal Studies, vol. 11, 1991 p. 32.
9See, for instance, D. E. Scheid, “Constructing a Theory of Punishment, Desert, and the Distribution of
Punishments”, The Canadian Journal of Law and Jurisprudence, vol. 10 no.2, 1997, p. 486; or A. Ashworth,
The Principles of Criminal Law, Clarendon Press, Oxford, 1995 chap. 11.
10A. Ashworth, “Sharpening the subjective element in criminal liability”, in A. Duff & N. Simmonds (eds.),
Philosophy and the Criminal Law, Franz Steiner Verlag, Wiesbaden, 1984, p. 79. See also his “The elasticity
of mens rea”, in C. F. H. Tapper (ed.), Crime, Proof and Punishment, Butterworth, London, 1991; or “Taking
the consequences”, in S. Shute, J. Gardner & J. Horder (eds.), Action and Value in Criminal Law, Clarendon
11A. von Hirsch and N. Jareborg, “Gauging Criminal Harm: A Living-Standard Analysis”, Oxford Journal of
Legal Studies, vol. 11, 1991 p. 30.
12D. N. Husak, “Is Drunk Driving a Serious Offence?”, Philosophy and Public Affairs, vol. 23, 1994, p. 66.
13See, for instance, J. Feinberg, Harm to Others, Oxford University Press, 1984, ch. 6. Or A. von Hirsch,
“Extending the Harm Principle: ‘Remote’ Harms and Fair Imputation”, in A. P. Simester and A. T. Smith
(eds.), Harm and Culpability, Clarendon Press, Oxford, 1996.
14A. von Hirsch and N. Jareborg, “Gauging Criminal Harm: A Living Standard Analysis”, Oxford Journal of
Legal Studies, vol. 11, 1991, p. 4.
15See, for instance, C. L. Ten, Crime, Guilt and Punishment, Clarendon Press, Oxford, 1987, p. 151-52;
or J. Klening, Punishment and Desert, Martinus Nijhoff, The Hague, 1973, pp. 120-23.
16R. Nozick, Philosophical Explanations, Harvard University Press, Cambridge, 1981, p. 363. Nozick
presents the formula as the product of harm and responsibility (H*R) where “responsibility” refers to the
degree to which a person flouts correct values.
17It is worth noticing that “culpability” is not always used unambiguously. Some claim that culpability is
a function of harm and mens rea (or responsibility), while others regard harm and culpability as two
components which both have to be present in order to determine the seriousness of a crime.
18For a short outline of other fault terms see, for instance, A. Ashworth, The Principles of Criminal Law,
Clarendon Press, Oxford, 1995 p. 194f.
19See, for instance, Hirsch, Past and Future Crimes, Rutgers University Press, New Jersey, 1985, p. 71;
or Gross & Ashworth, The English Sentencing System, Butterworths, London, 1981, p. 146-7. For a more
comprehensive discussion of the different mens rea terms, see A. Ashworth, The Principles of Criminal
Law, Clarendon Press, Oxford, 1995, chapter 5.
20A locus classicus is H. L. A. Hart, Punishment and Responsibility, Clarendon Press, Oxford, 1968,
chapter vi. For a defence of the opposite view, that negligence itself does not merit moral blame, see, for
instance, M. S. Moore, “Choice, Character, and Excuse”, Social Philosophy and Policy, vol. 7, 1990 p.
58.
21See, for instance, A. P. Simenster, “Why Distinguish Intention from Foresight?”, in A. P. Simenster &
A. T. H. Smith (eds.), Harm and Culpability, Clarendon Press, Oxford, 1996.
22A. Ashworth, Sentencing and Penal Policy, Weidenfeld and Nicolson, London, 1983, pp. 152-3. Under
the category ‘intention’, Ashworth distinguishes between: planned, deliberate, sudden, ‘spur of the
moment’, and impulse. Likewise, ‘recklessness’ covers: calculated risk, deliberate risk, sudden risk, ‘spur
of the moment’ risk, and a risk which could have been foreseen if thought about.
23See, for instance, S. H. Kadish, Blame and Punishment, Macmillan Publishing Company, New York,
1987, pp. 82-86; and M. S. Moore, “Choice, Character, and Excuse”, Social Philosophy and Policy, vol. 7
1990, pp. 30-31.
24For an outline and discussion of each of the theories, see, for instance, R. A. Duff, “Choice, Character,
and Criminal Liability”, Law and Philosophy, vol. 12 1993; M. S. Moore, “Choice, Character, and
Excuse”, Social Philosophy and Policy, vol. 7, 1990, reprinted in Placing Blame, Oxford University
Press, New York, 1997; N. Lacey, State Punishment, Routledge, Great Britain, 1988.
25H. L. A. Hart, Punishment and Responsibility, Oxford University Press, New York, 1968 p. 152.
26One possibility is to try to define capacity narrowly as an ability to recognize and foresee the relevant
emirical aspects of an action combined with a kind of rationality. A broader way of understanding the
term is to define it relative to what could be expected by a reasonable person possessing a proper degree
of virtues. See R. A. Duff, “Choice, Character, and Criminal Law”, Law and Philosophy, vol. 12, 1993 p.
358.
27See, for instance, M. S. Moore, “Choice, Character, and Excuse”, Social Philosophy and Policy, vol. 7,
1990 p. 40.
28See G. Fletcher, Rethinking Criminal Law, Boston, 1978; R. Brandt, Ethical Theory, Prentice-Hall,
Englewood Cliffs, 1959; M. Bayles, “Character, Purpose and Criminal Responsibility”, Law and
Philosophy, vol. 1, 1982; P. Arenella, “Character, Choice and Moral Agency: The Relevance of Character
to our Moral Culpability Judgments, Social Philosophy and Policy, vol. 7, 1990.
29R. Nozick, Philosophical Explanations, Harvard University Press, Cambridge, 1981 p. 383.
30P. Arenella, “Character, Choice and Moral Agency: The Relevance of Character to our Moral
Culpability Judgments”, Social Philosophy and Policy, vol. 7, 1990 p. 75f.
31See, for instance, Moore’s illuminating discussion in M. S. Moore, “Choice, Character, and Excuse”,
Social Philosophy and Policy, vol. 7, 1990 p. 40ff.
32See ibid. p. 41.
33 For a discussion of this assumption, see R. A. Duff, “Choice, Character, and Criminal
Liability”, Law and Philosophy, vol. 12, 1993 p. 371.
34G. Flecther, Rethinking Criminal Law, Little, Brown, Boston, 1978 p. 801.
35P. Arenella, “Character, Choice and Moral Agency: The Relevance of Character to our Moral
Culpability Judgments, Social Philosophy and Policy, vol. 7, 1990 p. 73.
36A. von Hirsch, Past and Future Crimes, Rutgers University Press, New Jersey, 1985, p. 74.
37A. von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993, p. 29; or “Seriouness, Severity
and the living Standard”, in Hirsch & Ashworth, Principled Sentencing, Hart Publishing, Oxford, 1998, p.
186.
38G. Fletcher, “The Recidivist Premium”, Criminal Justice Ethics, vol. 1 1982. R. Singer, Just Deserts,
Ballinger, Cambridge, 1979.
39I shall here talk of recidivism as a factor which affects the seriousness of a crime. If some would prefer
to say that seriousness is only affected by harm and culpability and that recidivism should be regarded as
a factor beyond seriousness affecting the appropriate punishment, then this way of speaking obviously
does not affect any of the theoretical problems which basically relate to the view.
40A. von Hirsch, Doing Justice, Hill & Wang, New York, 1976; “Desert and Previous Convictions in
Sentencing”, Minnesota Law Review, vol. 65, 1981; “Desert and Previous Convictions”, in A. von Hirsch
& A. Ashworth (eds.), Principled Sentencing, Hart Publishing, Oxford, 1998.
41Ibid. p. 193.
42A. Ashworth, Sentencing and Penal Policy, Weidenfeld and Nicolson, London, 1983, ch. 5.
43A. von Hirsch, “Desert and Previous Convictions”, in A. von Hirsch and A. Ashworth (eds.),
Principled Sentencing, Hart Publishing, Oxford, 1998, p. 195.
44Ibid. p. 195.
45Ibid. p. 195.
46See ibid. p. 196.
47A.von Hirsch, “Desert and Previous Convictions in Sentencing”, Minnesota Law Review, vol. 65,
1981, p. 601.
48Whether Hirsch regards the argument concerning human frailty as a reason for a discount
independently of the main argument concerning respect for the capacity to reflect on wrongdoing and to
show self-restraint, is not quite clear. However, after having mentioned the human fallibility which calls
for tolerance he claims that the “discount is also granted” (1998 p. 195) on the ground of this respect,
which seems to indicate that the two reasons are meant as separate arguments for a diminution of the
initial penal response.
49A. von Hirsch, “Desert and Previous Convictions in Sentencing”, Minnesota Law Review, vol 65,
1981, p. 601.
50A. M. Durham III, “Justice in Sentencing: The Role of Prior Record of Criminal Involvement”, The
Journal of Criminal Law & Criminology, vol. 78 no. 3, 1987, p. 633.
51A. von Hirsch, “Desert and Previous Convictions”, in A. von Hirsch and A. Ashworth (eds.),
Principled Sentencing, Hart Publishing, Oxford, 1998 p. 194.
52A. von Hirsch, “Desert and Previous Convictions in Sentencing”, Minnesota Law Review, vol. 65,
1981 p. 603.
53For a more thorough discussion see, for instance, J. Ryberg, “Recidivism, Multiple-Offending, and
Legal Justice”, Danish Yearbook of Philosophy, vol. 36, 2001.
54Von Hirsch himself admits that he has no ready answer to this question; see A. von Hirsch,“Desert and
Previous Convictions in Sentencing”, Minnesota Law Review, vol. 65, 1981 p. 616.
55Ibid. p. 617.
56Ibid. p. 616.
57See A. Ashworth, Sentencing and Penal Policy, Weidenfeld and Nicolson, London, 1983, ch. 6. Or N.
Jareborg, “Why Bulk Discounts in Multiple Offence Sentencing”, in A. Ashworth and M. Wasik (eds.),
Fundamentals of Sentencing Theory, Clarendon Press, Oxford, 1998.
98 JESPER
THE ETHICS RYBERG
58H. L. A. Hart, Punishment and Responsibility, Oxford University Press, New York, 1968, p. 162.
59Don E. Scheid, “Constructing a Theory of Punishment, Desert, and the Distribution of Punishments”,
The Canadian Journal of Law & Jurisprudence, vol. 10, no. 2, 1997, p. 484.
60Ibid. p. 485.
61C. L. Ten, Crime, Guilt, and Punishment, Clarendon Press, Oxford, 1987, p. 155.
62Section (2) above.
63I. Primoratz, “On retributivism and the lex talionis”, Rivista Internazionale di Filosofia del Diritto, vol.
61, 1984, p. 89.
64See, for instance, M . Davis, “Criminal Desert and Unfair Advantage”, Law and Philosophy, vol . 12,
1993, p. 138.
65Ibid. p. 139.
66M. Davis, “How to Make Punishment Fit the Crime”, Ethics, vol. 93, 1983, p. 739.
67The role the typing plays in Davis procedure is in my view not clear. It is simply unclear whether all
crimes are comparable in terms of being more, less or equally serious, or whether it is only crimes within
a group which are comparable. In some places, Davis seems to believe that it is only crimes within a type
that are in this sense comparable. However, as Dolinko has argued, this makes the anchoring of the crime
and punishments scales, prescribed in step 6 in the seven-step procedure, very arbitrary (D. Dolinko,
“Mismeasuring “Unfair Advantage”: A Response to Michael Davis”, Law and Philosophy, vol. 13, 1994,
p. 519, 522.). On the other hand, it is hard to see why the reason Davis gives for the typing, namely, that
this will give potential criminals a reason to choose the lesser crime, should be nothing more than a
recommendation to make the scale easily readable. Moreover, since all prices reached on crime licenses in
Davis’ auction model are comparable, and since Davis’ claim is that the rankings provided by the two
methods are equivalent, it seems to follow that also crimes belonging to different types must be
comparable in terms of seriousness.
68In the 1983 paper Davis suggests that the ranking should express what a rational person would prefer to
risk given a choice between different crimes.
69M. Davis, “Criminal Desert and Unfair Advantage”, Law and Philosophy, vol. 12, 1993, p. 154.
70Ibid. p. 142.
71Ibid. p. 150f.
72For instance, Dolinko believes that assumption c) is so unrealistic as to jeopardize the value of the
auction model even as a heuristic devise. D. Dolinko, “Measuring ‘Unfair Advantage’: A Response to
Michael Davis”, Law and Philosophy, vol. 13, 1994, p. 505.
73M. Davis, To Make the Punishment Fit the Crime, Westview Press, USA., 1992, P. 115.
74As mentioned, Davis believes that bidders on licenses to fail might be people who have not obtained a
license to succeed. However, there is perhaps another possible way Davis could respond to the argument;
this would be by holding that, even if poaching is impossible, there would still be some who would buy a
license to fail, namely, those people who already have a license to succeed. However, Davis’ own view is
that a license to succeed could be used to pardon failure. See his To Make Punishment Fit the Crime,
Westview Press, USA., 1992, p. 112.
75M. Davis, To Make the Punishment Fit the Crime, Westview Press, USA., 1992, p. 84.
76D. E. Scheid, “Davis and the Unfair-Advantage Theory of Punishment. A Critique”, Philosophical
Topics, vol. 18, 1990; and D. E. Scheid, “Davis, Unfair Advantage Theory, and Criminal Desert”, Law
and Philosophy, vol. 14, 1995.
77See, for instance, D. Dolinko, “Mismeasuring ‘Unfair Advantage’: A Response to Michael Davis”,
Law and Philosophy, vol. 13, 1994; or A. Ellis, “Punishment and the Principle of Fair Play”, Utilitas, vol.
9, 1997.
78At one place Davis suggests a third reason in favour of the equivalence, besides the two reasons
already considered. He believes that people who do not intend to use a license but who fear to become the
victim of a crime, might also bid at the auction, and that the prices therefore will also approximate a
ranking according to what people fear. However, Davis himself later rejected this proposal. See, M.
Davis, To make Punishment Fit the Crime, Westview Press, 1992, p. 240.
79See, for instance, S. I. Benn & R. S. Peters, Social Principles and the Democratic State, George Allen
& Unwin Lld., London, 1959, ch. 8.
80C. Beccaria, On Crimes and Punishment, in A. Manzoni (ed.), The Column of Infamy, Oxford
University Press, Oxford, 1964.
81A. von Hirsch, Doing Justice, Hill & Wang, New York, 1976, p. 75.
CHAPTER 3
THE SEVERITY OF PUNISHMENTS
In order to provide a full account of what proportionalism amounts to, and to unfold
the position in such a way that it is capable of functioning as a principle governing
punishment practice, it is obviously not sufficient to consider only the relative
ranking of crimes in gravity. Of equal importance is the question of what it means
that one punishment is more severe than another, and the challenge of providing
some sort of scaling of punishments in severity. This discussion, to which we shall
now turn, is from the outset complicated by the conjunction of two facts.
Firstly, it is the case that there exist many different ways in which a
criminal’s wrongdoing can be responded to in punitive measures. Much of the early
literature in the modern retributivist epoch has focused primarily on custodial
punishment. However, from the mid-80’s increasing attention has been directed to
other types of punishment. Motivated, for instance, by the contention that a
punishment system which offers only a relatively few punishment options will often
punish perpetrators either too severely or too leniently relative to the crime
committed, there has been a growing interest in intermediate sanctions as
constituting the tertium quid between prison or probation.1 That intermediate
sanctions, including for instance, home detention, community service, day fines,
electronic monitoring etc, have by proportionalists been recognized as alternative
punishments and not merely as alternatives to punishment, means that there are
great differences in the objective appearance between the possible punishments that
2
should be arrayed. Secondly, punishments within a certain type can obviously
differ very much in severity. The severity of imprisonment will usually vary with its
duration, a fine with the quantum of money, and so on with regard to other
punishment types. Together, these two facts imply that one cannot simply assume
that the scaling of punishment in severity follows the different types of sanction.
That is, for instance, that imprisonment is always more severe than alternative
punishments. It certainly makes sense to ask how a minor period of imprisonment
should be assessed in comparison to a large fine or a long period of probation under
onerous conditions. Answers to these questions presuppose a theory of how
punishment severity should be assessed.
As was the case with regard to the question of how crimes should be
compared in terms of seriousness, a number of researchers have approached the
question by adopting techniques to surveying popular perceptions of the severity of
various sanctions. In one of the first tentative explorations of this kind, Sebba and
his colleagues asked a number of respondents to provide scores for each of thirty-
six penalties, varying from a 10$ fine to the death penalty, in accordance with its
101
severity.3 A number of such surveys in more elaborated versions have subsequently

been carried out.4 As the motivation for this kind of research, Sebba specifically
claims that among other reasons the most compelling ground is found in the general
movement in penal theory towards a retributive model of punishment, and that the
prescription that punishment should fit the crime makes the development of an
empirically based scale of penalties highly pertinent. Notably, however, no
argument is given as to why this is the case. In more general terms, even though it is
the case that people have a certain perception of the severity of various punishments,
a perception which can be revealed through empirical surveys, it obviously does not
follow that such surveys establish how punishments ought to be scaled. What counts
as a plausible scaling of punishments is basically a normative question. Even if the
scaling which proportionalists would prescribe might in the end be developed
through some kind of opinion survey, a theory would still be required to explain
what is meant by punishment severity and why a scaling should be determined by
this sort of procedure. In this respect, the question on the scaling of punishments is
not different from the discussion on the ranking of crimes in gravity.
Though proportionalists are well aware of this fact, remarkably little has
been done to provide a sufficient theoretical foundation for the assessment of
punishment severity. Perhaps due to the belief that the answer is pretty
straightforward and that it therefore does not prompt any serious theoretical
challenges, only a few theorists have entered into a closer scrutiny of the question.
What I shall do in the following section is, firstly, to outline an important challenge
confronting the proportionalist idea of punishment severity and to consider how this
challenge might be met. In the subsequent section further problems which follow
from the notion of punishment severity will be discussed. As will be clear, the fact
that an answer as to what determines punishment severity may - in comparison with
the question of crime seriousness, be theoretically less complicated - in no way
means that the question does not confront proportionalists with serious challenges.
1. THE SENSIBILITY CHALLENGE
The comparison of the severity of various punishments and thus the construction of
a punishment scale does at first glance seem a theoretically manageable task. As
mentioned, the severity of punishments within a certain type often seems to vary
with an easily measurable dimension such as the duration of imprisonment or the
size of a fine. Moreover, there do not seem to be many different factors which
determine severity. As demonstrated in the last chapter, one of the problems in
ranking crimes is that there are several seriousness-determining dimensions, which
means that a theory is required to explain how these dimensions should be worked
together in order to reach a final judgement. The same problem apparently does not
exist when it comes to the assessment of punishment severity.
The obvious approach to the problem of comparing punishments is to
contend that what counts is the amount of suffering, hardship, deprivation or
inconvenience which a punishment inflicts on the punished. In more general terms it
might be said that punishment severity is determined by the negative impact on the
THE SEVERITY OF PUNISHMENTS 103
well-being of the punished. One year in prison is a more severe punishment than a
500$ fine if it, to a larger extent, affects a person’s life for the worse, that is, if it
causes more suffering, unpleasantness or deprivation. At the outset this answer
seems hard to reject. It corresponds well with how we evaluate many other
misfortunes and calamities that befall us during our lives. And it might even,
depending on which quality of life theory one adheres to, be held to provide a
theoretically attractive uni-dimensional answer to the severity question. There is,
however, one important snag, namely, the answer apparently undermines
proportionalism, at least in the way the principle is standardly interpreted.
The problem which the answer gives rise to is that objectively similar
punishments may well have subjectively variable effects on different persons or as
Bentham - who thoroughly considered the issue - has put it, that “a punishment
which is the same in name will not always either really produce, or even so much as
5
appear to others to produce, in two different persons the same degree of pain” .
Bentham refers to the disposition a person has to feel a quantity of pleasure or pain,
upon the application of a cause of a given force, as the degree of his or her
6
“sensibility”. That there might exist significant variations in sensibility and thus
large differences in the quantum of suffering caused by what is objectively the same
punishment is hard to dispute. In fact, such differences exist both at inter-personal
and intra-personal levels.
For instance, one person may endure much more suffering from spending a
year in prison than will another person. Differences in two persons’ physical and
psychological sensitivity obviously condition the amount of suffering which each
7
endures from spending a year under the same prison conditions. And, of course, the
same differences exist with regard to other sorts of punishment. Even in the case of
day fines, which are adjusted to some of the conditions of a criminal’s life, there
may still be differences in the impact they have and thus on the inconvenience they
cause. Furthermore, there may well be variations in sensibility within a single life.
An imprisonment may be much harder at the age of seventy than at the age of thirty.
Likewise, differences in many other respects may lead to large variations in the
suffering which a punishment will cause to a person at one time in his life rather
than at another. A person’s criminal record may also be relevant. For instance, the
first custodial punishment may be experienced as much more traumatic than later
but objectively similar punishments. In fact, there may even be changes in a
person’s sensibility during a punishment. Some inmates may adjust successfully to
prison life, but the opposite may also be the case. What at first starts out as a
bearable time in prison may - if a person loses his hope or instinct of self-
preservation - turn into a nightmare (even if the outer conditions are unchanged).
Thus, as the examples indicate, it seems reasonable to hold that one and the same
punishment may have a very different impact on people’s lives.
How does this conclusion affect the question of how punishments should
be ranked and the whole idea of proportionality in punishing? To contend, on the
one hand, that one should, in accordance with the proportionality principle, respond
to crimes of a certain degree of seriousness by always imposing the same
punishment - for instance, that an assault of a certain gravity should be responded to
104 JESPER
THE ETHICS RYBERG
by one year of imprisonment - while holding, on the other that punishment severity
should be determined by the suffering or discomfort that is experienced - which
might under the same objective treatment vary very much from one person to
another - seems incoherent. If A and B have each committed a crime of the same
gravity, then the imposition of the same (objective) punishment on the two may
imply that A is (in terms of severity) punished much harder than B. And,
alternatively, if we imagine that A has committed a crime which is more serious
than the one committed by B then the imposition of (objectively) different crimes on
the two, say, respectively two years and one year in prison, may nevertheless imply
that B is, due to his more sensible character, punished more severely. How can a
proportionalist respond to this challenge of sensibility?
Unfortunately, very few proportionalists have entered into a discussion of
this challenge. However, there are basically two ways of answering the problem: to
reject it or to accept and account for it. Either one can maintain that the
proportionality principle is in the end correct. If two perpetrators have committed
equally serious crimes they should receive objectively the same amount of
punishment and if they have committed crimes of different gravity they should be
punished objectively different. In this view, the principle is unaffected by the
sensibility challenge. However, one will thereby be committed to a purely objectivist
account of punishment ranking, that is, a ranking according to which what basically
matters in the assessment of severity is not something that is conditional on inter-
personally or intra-personally variable circumstances. Or, alternatively, one can hold
that a morally plausible account of punishment severity must at least partly rely on
something which is inter-personally or intra-personally variable, which means that
punishments which are the same in name may be more severe for one person than
for another. In this view the proportionality principle must be abandoned in favour
of a distribution principle according to which what basically matters is that crimes
are responded to with appropriately severe punishments even if these are objectively
very different. Let us consider the two approaches in turn.
Is it possible to maintain proportionality with regard to objective
punishments as the basic distribution principle? Though many theorists apparently
believe so, almost no one has felt prompted to explain more thoroughly why the
challenge of sensibility should in the end be considered insubstantial. However, a
8
hint is given in an article by von Hirsch, Wasik and Green. What the three authors
are concerned with is to offer a conception of how non-custodial penalties should be
arrayed. More precisely, they suggest a proportionately structured system of
punishments which is flexible, in the sense that it allows for the possibility of
interchangeability between different penalties. They consider as the two polar
extremes a model which does not permit substitution among punishments and a
model which, like the one that has been suggested by Robinson, permits full
9
substitution. Their own model is one in between which allows only for limited
substitution. The idea behind substitution models is the distinction between amount
and method of punishing, that is, that one sort of punishment may be substituted by
another as long as the substitute punishment is of equivalent severity. It is interesting
to note that, besides the fact that a substitution model makes it possible to let in
forward-looking considerations without deviating from proportionality, von Hirsch,

Wasik and Green also believe that a system which does not allow for substitution, at
least among some types of non-custodial punishments, will not be practically
workable. If this is correct, it further emphasizes the importance of providing
criteria for gauging punishment severity.
Faced with the crucial question of how this should be done, they answer by
merely outlining two different theories. According to the first, severity depends
upon the degree of pain or discomfort felt by those upon whom the punishment is
inflicted. According to the second, severity should be measured by the “normative
importance” of the personal interests which are compromised by a particular
punishment. Unfortunately, the authors do not comment on how these two
suggestions relate to each other, that is, to what extent they might provide different
answers and which of the two should be regarded as the more plausible. It is
obvious, however, that if one adopts something like the first conjecture one will
immediately be confronted with the challenge of sensibility. Degrees of pain and
discomfort are, as we have seen, sensibility-relative. The same treatment may in this
respect affect the callous person and the thin-skinned person very differently. But
what about the second approach? Is it possible to develop an interest account which
is not in the same way sensitive to individual differences?
The idea behind this suggestion is to provide an objectivist account of the
severity measurement. This has more recently been confirmed by von Hirsch who
now believes that it is the second approach that should be applied when punishments
10
are ranked and compared for possible substitution. More precisely, the suggestion
is that punishment should be measured by a living-standard analysis comparable to
the one adopted in the von Hirsch-Jareborg scheme for gauging seriousness of
crimes. Von Hirsch claims that, in comparison to when the method is applied for
gauging crime seriousness, a new taxonomy of interests will have to be developed.
Unfortunately, very little is said about what should count in this kind of analysis
except that “living-standard” refers to the means and capabilities that ordinarily
assist persons in achieving a good life. Thus, though we cannot, due to the lack of
clarification, assess the conjecture on its precise content it does, however, still make
sense to ask whether it is at all reasonable to expect that such a model can avoid the
challenge of sensibility. That von Hirsch believes so is clear. He explicitly rejects a
11
“subjectivist view”. But how can one possibly sustain the claim that an objectivist
approach is basically the one that should count? It is important to keep in mind that
the question we are asking is not what constitutes the best - that is, most precise -
method to gauging what determines severity but rather what it is that morally
matters in the comparison of punishments. If one believes that what counts is the
suffering, deprivation or discomfort which is inflicted on the punished but that the
most precise method we have for measuring this kind of impact is some sort of
living-standard analysis, then obviously the challenge of sensibility remains
unanswered. The challenge is not avoided simply by claiming that we do not have
very precise measuring methods. And obviously one cannot just redefine what
basically counts in some practically convenient but morally arbitrary way. Thus, the
question remains: how can one possibly defend a purely objective approach to
106 JESPER
THE ETHICS RYBERG
punishment severity, that is, an approach which is immune to the challenge of

sensibility? There seem to be two - and, I believe - only two ways in which this
question can be answered.
The first answer would be to contend that some objectivist approach
actually provides a correct picture of how someone’s life is affected for the worse
(or the better). That is, that what such an approach actually measures is the degree to
which someone’s life is deteriorated by being punished in a certain way. If this is
correct then obviously it is reasonable to rely on such an approach when the severity
of different punishments is estimated. However, this answer is clearly not plausible.
On a purely objectivist account, in the sense we are here using the word, what
determines degrees of severity is something which is independent of differences in
the sensibility of different persons. But this implies that the pains, frustrations, and
unpleasantness which a punishment may cause to the punished - and which are
certainly sensibility-relative features - have no effect on the degree to which one’s
12
life is worsened by being punished in a certain way. This claim is implausible.
Even the most objectivist-minded theories of well-being let in this kind of feeling.
13
Neither, it should be mentioned, is this the answer which von Hirsch has in mind.
The second answer would be to point out that even though something may
have an impact on a person’s well-being, this is not tantamount to holding that it is
of moral relevance. It sometimes seems to make sense to hold that there are certain
effects on a person’s well-being which are simply not morally valuable. Following
this line of thought, the claim must be that even though some of the inter-personally
variable experiences people have when they are punished clearly do affect their
well-being these experiences are not morally relevant and should therefore not
figure in the computation of punishment severity. However, neither is this answer
easy to accept. Even if it is correct that some effects on a person’s well-being (e.g.
the well-being gained from climbing Kilimanjaro or drinking fine bottles of
champagne) are not morally valuable, it is certainly far from obvious that effects on
well-being caused by a punishment belong to this category. To claim that the
suffering or inconvenience a criminal experiences should be what matters in
comparing the relative severity of punishments comports well, I believe, with how
we would usually morally compare the gravity of many other incidences that might
befall us.
Moreover, that suffering should figure at least as a part of what determines
severity seems to be exactly what we should expect from the theories of justice on
which the proportionality principle is claimed to be based. According to what was
called the simple desert theory, what wrongdoers deserve simply is to suffer. This
certainly also makes it most obvious to suggest that it is the amount of suffering that
counts in the comparison of punishment severity. Similarly, it seems reasonable to
expect that since what matters, from a fairness theoretical point of view, is that a
burden is imposed on the perpetrator in order to restore a fair balance of benefits
and burdens, the burden must be genuine for the individual criminal. This indicates
that suffering or inconvenience must be part of what determines a burden’s size and
consequently the severity of punishment. And if what matters to the expressionist is
that a means is used which makes it possible to reach the criminal, to offer what can
function as a penance for the person, or what will provide a prudential disincentive,
then it certainly seems reasonable to expect that it is the actual impact on the person
that matters which again seems to indicate that the suffering one experiences must
be a part of what counts when punishments are compared in severity. All in all, it
seems to require a strong argument to support the contention that suffering should
not be part of what determines the severity of different punishments.
The only argument which, to my knowledge, has been presented along
these lines is that some sort of objectivist account of punishment severity is needed
“as a safeguard against class justice”. If one accounts for differences in sensibility it
might follow, as has been put by von Hirsch, that “the middle-class person is put on
14
probation and the ghetto youth jailed for the same infraction ...” . This, it might be
held, would be unjust. Hence, an subjectivist impact approach is unacceptable.
It is understandable if one at first sight feels that there is something to this
argument, especially if one recalls how the least privileged members of society in a
historical perspective have been treated by the criminal justice systems. However, as
an objection against measuring punishment severity in a way that would imply that
one should account for differences in sensibility, the argument is - depending on
how it is interpreted - either unconvincing, unacceptable for the proportionalist, or
misplaced. What would certainly be morally dubious would be to treat members of
lower-classes in ways which would harm them more than the ways in which other
members of the society are treated, but obviously this is not what is implied if one
relies on a view which requires adjustments to differences in sensibility. The kind of
differential treatment which is implied is exactly the same as would be involved if
an older or psychologically vulnerable person is treated objectively less severely
than a young or psychologically less vulnerable person. Thus, it is hardly convincing
15
to hold that this kind of differential treatment is in general morally questionable.
Of course, it might be suggested that what is wrong is differential treatment
specifically in relation to lower-class members of society. For instance, if it appears
- contrary to what is the case - as if certain unprivileged are treated in a way that is
more harmful to them compared to the treatment of other groups then this might
have severe side-effects with regard to how this group is treated in other contexts. It
might, for instance, increase certain discriminatory tendencies. However, on this
interpretation the argument is hardly acceptable for the proportionalist: it would
certainly have rather extensive implications if this kind of forward-looking
consideration should be accounted for in the way criminals are punished for their
misdeeds. Perhaps the most plausible interpretation of the argument, and what von
Hirsch may have in mind, is that there is basically something unjust in punishing
those who are genuinely unprivileged in society or, put more generally, that the
application of the proportionalist ideal must to some extent presuppose justice in
society. But obviously this view, no matter how plausible it seems, misses its target:
it does not constitute an argument against what should count in the assessment of
punishment severity. It may be presented as an argument concerning how seriously
the crimes committed by the least privileged should be considered (e.g. whether
there should be a “deprivation excuse”) or whether it would be fair to punish them
for their misdeeds at all, that is, whether the preconditions for applying a certain
108 JESPER
THE ETHICS RYBERG
16
punishment system in the first place have been satisfied. But clearly it is not an
argument targeted at what constitutes a plausible account of punishment severity.
Thus, on closer scrutiny the argument is defective. And in the absence of any other
arguments supporting the purely objectivist account it seems most convincing to
maintain that suffering, unpleasantness, inconvenience and other sensibility-
dependant aspects of a punishment must be part of what determines its severity.
What this shows is that, even in the absence of a clarification of the precise
content of a purely objectivist account of punishment severity, the idea of such an
17
approach should be rejected. A morally plausible account of punishment severity
cannot avoid counting in features the strength of which to some extent will be
conditioned by the sensibility of the individual perpetrator. But this means that we
will have to give up proportionality in the sense in which the principle has typically
been presented, namely, as requiring that a criminal who has committed a crime that
is more serious than one committed by another criminal, should be punished with a
punishment which in name is more severe, and that criminals who have committed
equally serious crimes should be punished with objectively equally severe
punishments. How serious is this problem, which the sensibility challenge has
brought forward, for the proportionalist?
What the challenge obviously implies is not that one cannot maintain the
basic idea of proportionality between the seriousness of a crime and punishment
severity. Rather, what it shows is, as mentioned, that one will have to defend a
revised account of proportionality according to which there should be
proportionality between the gravity of a crime and the subjective punishment, that is,
the impact which a punishment will have on a criminal given his or her particular
degree of sensibility to punitive treatment. This means that if A has committed a
crime which is more serious than one committed by B then the right distribution of
punishment might, if A is assumed much more sensible than B, be one in which A
receives one year in prison while B receives one and a half years. And if A and B
have committed equally serious crimes the parity condition might be maintained by
imposing, on the two, prison terms of different duration. In this interpretation the
parity condition corresponds to what Ashworth calls “the principle of equal
18
impact” . This version of the proportionality principle makes the prescription of
proportionate punishment consistent with a plausible account of punishment
severity, whatever this more precisely consists in, and as we have seen it accords
well with the underlying theories of justice on which the principle is claimed to be
founded. Moreover, it seems to generally capture the idea of justice much better than a
version focusing on a mere objective account of punishment severity. It is the idea
of a day-fine system extrapolated to all sorts of punishing.
However, there is one obvious drawback to this account of proportionality:
a substantial loss of simplicity in application. In order to maintain proportionality,
that is, in order to ensure that a criminal receives the punishment which is warranted
by the crime, it will be required that one possesses a precise knowledge with regard
to the sensibility of the particular criminal. But the sensibility is obviously
determined by various factors to which there is no simple access for the punisher
and, even if some knowledge is available, the task of calibrating punishment severity
in a way which observes proportionality will indeed be a complicated task. As

something more than a mere curio it is worth recalling Bentham’s list of thirty-two
“circumstances influencing sensibility”:
1. Health. 2. Strength. 3. Hardiness. 4. Bodily imperfection. 5. Quantity and quality

of knowledge. 6. Strength of intellectual powers. 7. Firmness of mind. 8. Steadiness
of mind. 9. Bent of inclination. 10. Moral sensibility. 11. Moral biases. 12. Religious
sensibility. 13. Religious biases. 14. Sympathetic sensibility. 15. Sympathetic biases.
16. Antipathetic sensibility. 17. Antipathetic biases. 18. Insanity. 19. Habitual
occupations. 20. Pecuniary circumstances. 21. Connexions in the way of sympathy.
22. Connexions in the way of antisympathy. 23. Radical frame of body. 24. Radical
frame of mind. 25. Sex. 26. Age. 27. Rank. 28. Education. 29. Climate. 30. Lineage.
19
31. Government. 32. Religious profession.
Imposing the deserved punishment might in this perspective become a very delicate
matter. And obviously this is not merely a practical problem: there is a clear moral
side to it. If inaccuracies are unavoidable, or perhaps even what one should most of
the time expect, then it might with a not negligible frequency be the case that
perpetrators will receive punishments which are not those warranted by the gravity
of their criminal conduct. However, the extent to which this constitutes a problem
for the proportionalist is not a matter which will be pursued here. It will be
thoroughly discussed in a subsequent chapter. Thus, the conclusion which, on the
present ground, can be drawn is that the sensibility challenge constitutes a genuine
challenge which means that it cannot just in one way or another be ignored or tossed
aside; rather must it be accommodated for in the account of proportionality. In that
sense, the challenge can be seen as a way of clarifying what the proportionality
principle on closer inspection implies.
2. DELIMITATING PUNITIVE SUFFERING
The infliction of a punishment on someone can have many different effects in terms
of suffering, hardship or deprivation. As we have just seen, the effects may vary
from one person to another or from one time to another for one and the same person.
There is, however, another problem which relates to the different effects of
punishing, a problem which complicates our present discussion of punishment
severity even further, namely, that it may not be a simple matter to specify the
suffering or deprivation which follows from a certain punishment with regard to
how much of it should count in the final assessment of severity. There may be a
problem of delimitation.
At first sight, this might strike some as somewhat strange. After all, what
should count is precisely the suffering or deprivation caused by the imposed
punishment - no more and no less. However, a number of simple examples easily
demonstrate that it is not so obvious what this implies. A conviction and punishment
may have a number of long-term effects on the punished person’s life. To mention a
few: a prison term may, for instance, lead to the loss of job and the stigma if having
been imprisoned may imply that it is difficult to get a new job after release. The
punished may lose their wife or husband and other important social ties. And, more
simply, one may suffer from psychological after-effects from having been
imprisoned or from being labeled a criminal. Moreover, a punishment may have an
impact on the lives of other persons, most obviously on the close relatives of the
punished. Obviously it is not the case that such after- and side-effects always occur.
And the evidence on the matter may be both scrappy and equivocal. But what makes
such effects interesting is, firstly, they may have a quite serious impact in terms of
suffering and deprivation on the punished, and on other parties lives. And secondly,
even if a state carries out a number of initiatives designed to palliate such effects, we
cannot merely assume that they will be eliminated. How should a proportionalist
respond to such effects?
The reason for raising this question in the present context is not to consider
the question which, now and then, has been addressed by critics of retributivism,
namely, whether it is morally acceptable to punish someone if it produces these
effects. Rather, as indicated, the point is conceptual: to clarify what precisely
proportionality implies. The question is, should such effects figure in the assessment
of punishment severity? In discussing this, I shall limit the considerations to those
after- and side-effects which a punishment may have on the perpetrator himself and
leave out the effects on others. The relevance of the question we are faced with is
then pretty obvious. A theorist who has drawn attention to such effects in a moral
20
context is Walker, who refers to them as “incidental punishment” . Now, if Walker
is right, that is, if the after- and side-effects on the punished should be considered a
form of punishment then it follows that a criminal may end up receiving two
punishments; or perhaps rather, one punishment which, all things considered, is
much more severe than the punishment imposed on another criminal who perhaps
spends exactly the same length of time in prison or is allotted the same fine. And
notice, the explanation is now not that one person is psychologically or physically
more “sensible” than the other person but that the impacts on the outer conditions of
the person’s life under and after the penalty are very different. If this is correct then
obviously it has significant consequences for the proportionality principle. In order
to maintain the requirement of ordinal proportionality and the parity condition, one
will not only have to account for the differences in sensibility but also for the
difference in side-effects and possible long-term after-effects in terms of suffering
and deprivation. The alleged merit of the proportionality principle, namely, its
practical applicability is, therefore, not only challenged by the problem of sensibility
but also by the problems related to side-and after-effects.
Moreover, such effects might even result in the following problem. If the
after-effects on a criminal’s life will be much more severe than what is warranted by
the seriousness of the crime if the person is punished at all, then these effects cannot
simply be accounted for in the estimate of how severe the sanction should be: either
the criminal receives a punishment and then, due to after-effects, he will everything
considered be disproportionately punished, or he will not be punished, which
obviously is also a disproportionate (lack of) response. The point obviously is not to
claim that such situations will often occur but rather that they might occur and that it
will then not be difficult, but impossible, to mete out the appropriate punishment.
But should we accept that this modification, when it comes to the assessment of
punishment severity, is what proportionality implies? As the principle is usually
presented it is pretty clear that this is not something which proportionalists have
typically been willing to accept. But can this implication plausibly be rejected?
The problem which the proportionalist is confronted with is analogous to
the previously considered problem of remote harms in relation to crimes, namely,
that of specifying which of the harms triggered by a certain act should count in the
assessment of seriousness. Correspondingly, if all implications a punishment might
have to the punished with regard to suffering, hardship and deprivation should not
count in the computation of severity, some morally non-arbitrary way of
delimitating the relevant effects must be suggested. Or more accurately, we need to
know what should be regarded as the punishment, and what should be considered
only the non-punitive side- or after-effects.
The answer which would probably be suggested, and which perhaps
constitutes the main reason why the delimitation problem has so rarely been posed,
is that, for a very simple reason, it is not correct to hold that after-effects of the
outlined sort should be accounted for by the proportionalist. What the proportionalist
is concerned with is that a punishment should appropriately fit the crime, but such
effects simply fail to satisfy the necessary conditions for counting as punishments in
the first place. It follows from the concept of punishment that these effects are mere
side- or after-effects and not a part of the punishment and, even though such effects
may certainly be regarded as most unfortunate, and though initiatives should
perhaps be taken to palliate them, there is no genuine worry for the proportionalist
whose main purpose, after all, is not to minimize suffering, but to uphold justice.
But why should such effects not count as punishment? Asking this question
opens up the traditional discussion of the definition of punishment which is not
uncontroversial and which has sometimes been complicated by the fact that views
on the justification of punishment have been too narrowly incorporated in the
definitions. Luckily, however, there are obviously many aspects of the definition
question - such as whether a punishment must contain suffering, whether it must be
imposed upon an offender, and whether it must be for an offence against a legal rule
- which are irrelevant in the present context. However, there is also a condition
which has been underlined in the definitions which have been dominant since the
50’s - such as the Flew-Benn-Hart definitions - and which the proportionalist might
subscribe to, namely, as Hart puts it: that a punishment “must be intentionally
21
administered by human beings other than the offender” . Since we are not here
considering the possibility of self-punishment the latter condition is irrelevant. But
the former stipulation, that of intentionality, might seem to provide exactly what the
proportionalist would need. Since side-and after-effects are not intended by the
sentencer they cannot constitute or be part of a punishment. Therefore, there is no
22
genuine problem of delimitation.
This answer certainly has some immediate appeal. Moreover, it is hard to
think of any other obvious explanations of why after- and side-effects should be
disregarded. However, it is here important to make clear what the answer precisely
112 JESPER
THE ETHICS RYBERG
amounts to. What we are considering is not whether some sort of accident which
might happen to a criminal - such as breaking a leg or losing a wallet - can count as
a punishment for a certain misdeed. The idea of poena naturalis is not the subject.
What is under consideration are cases where a perpetrator is imprisoned, fined or
treated in another way which constitutes a standard instance of a punishment but
where this treatment also has an impact on the offender’s life in terms of side- and
after-effects. Thus, there is no doubt that what we are considering are intentional
acts of punishment. In order, to rule out side- and after-effects of a punitive act as
part of the punishment by reference to what was intended, what one will have to
hold is that it is only the suffering, hardship or deprivation which is intended that
should count as part of the punishment. Or put another way, the severity of a
punishment is determined by the intended amount of suffering, hardship or
deprivation. This, would rule out side- and after-effectual suffering on the
23
offender. No matter how severely a criminal suffers from psychological after-
effects of a punishment or how severely he or she will suffer from being stigmatized
in various ways these are merely unintended consequences which therefore should
not figure in the assessment of the severity of the punishment. However, on closer
scrutiny this suggestion is not plausible. Consider the following three - admittedly,
not all very likely - examples. (a) A person who should spend one year in prison
happens, due to a mere mistake, to be maintained for several months more. Should
we say that the person has only been punished to the extent of one year? (b) A
person manages to escape from prison when he has served only half his term;
another person, due to a mistake, never receives, and consequently does not pay, a
fine he should have paid for a certain misdeed. Should we say that both persons have
been punished equally severely as those who have served their sentence or who have
actually received and paid a corresponding fine? (c) Two persons are each allotted
one year in prison. The first person is placed in an old prison where he lives under
miserable conditions, the second serves his time in a modern well-functioning prison
under fine conditions. Should we say that they are equally severely punished?
The point of presenting these examples is not to appeal to our general
understanding of how we actually use the language, that is, whether an affirmative
or a negative answer will be most in accordance with our ordinary language usage of
the term “punishment”. An answer to this question would, I believe, turn out
equivocally and would be irrelevant. Rather the question is obviously whether the
answers should be in the affirmative or the negative. With regard to this question, it
seems that the answer must in all three cases be in the negative. The best explanation
for why one should object to the situations where a person unnoticed has managed to
avoid part of a prison term or a fine, or the case where a person has mistakenly
served much more time than he should, is - if one adopts a retributivist perspective
at all - that the persons have been punitively treated in ways that are not warranted
by the seriousness of their conduct. It would certainly seem unacceptable to hold
that as long as the intended punishments were appropriate there would in these cases
be no reason to object to the actual outcomes. However, if punishment severity
simply is intended severity then this is exactly how one would have to respond. And
if a person discovers that there are significant differences in the quality of prisons or,
to return to the previous discussion, that two persons have very different degrees of
sensibility to a certain form of punishment, then there certainly seem to be reasons

of retributive justice to make those who in the end mete out punishments acquainted
with these facts. (It is in this connection worth noticing that some of the few
proportionalists who have touched on the problem of sensibility have, as we have
seen, rejected the challenge by relying on objectivist accounts of impact but not by
maintaining that all there is to punishment severity is what is intended.)
What I am claiming, of course, is not that possible differences in sensibility
or in prison conditions would be irrelevant to the proportionalist sentencer, who
believes that it is intended severity that counts; if, for instance, there is an intention
to punish two persons equally severely and the sentencer knows that there are such
differences then they should be accounted for when the sentencer seeks to carry out
the intention. The problem rather is that, if the sentencer is misinformed or lacks
information on these matters, then there will not be reasons concerning justice to
object to the resulting punishments as long as what was intended did not violate
justice. It is not hard to think of examples which, to an even more extreme extent
than those already presented, underline the cruciality of this problem. (Consider the
sentencer who, due to an extreme degree of misinformation, believes that the death
penalty and a fine are equally severe and who in order to maintain the parity
condition allots to one of two perpetrators, who have committed equally severe
crimes, the death penalty and to the other the fine.) Thus, it seems most plausible to
hold that punishment severity cannot simply be a matter of what is intended and, as
mentioned, this even seems to accord with how some proportionalists themselves
consider the matter.
Some may well find this conclusion hard to swallow. Suppose that one
holds that it is the actual impact in form of suffering, hardship or deprivation which
a punishment has on the punished that determines the degree of severity. In that case
various effects which are triggered by the punishment but which are usually not
associated with the punishment might suddenly be considered part of it. If, for
instance, a person is accidentally hit by a falling tile just outside the prison after
release, and if it is assumed that this would not have occurred had the person not
been imprisoned, then it seems to follow that the suffering caused by the tile should
count in the computation of the severity of the imposed punishment. This, it might
be contended, seems absurd. After all, the falling tile was a genuine accident which
no one could possibly predict. The thought behind this view might be that
punishments cannot just happen to or befall a person but that they are treatments to
be understood within the context of responsibility, and that this must also be the case
when we consider the sufferings or deprivations which determine the degree of
24
severity. Since the sentencer was not - we can assume - in any way responsible for
the falling tile it should not be taken into consideration in the assessment of severity.
However, even if this view is correct it does not succeed in excluding
possible after- and side-effects as something which should figure in the assessment
of punishment severity. The point simply is that such effects are not analogous to the
falling tile. We might well have some general knowledge of how certain kinds of
punishment will affect people’s lives in after- and side-effectual ways. And we
might sometimes even have specific knowledge as to how a punishment will affect a
particular person’s life in terms of future suffering and deprivation. Since one cannot
plausibly hold that such effects fall outside the context of responsibility if they are
only caused knowingly, such effects cannot in this way be disregarded. Similarly, if
it was actually known that every time a person left a prison following release he or
she would be hit by a falling tile, it would certainly seem morally arbitrary to hold
that, since this happened just outside rather than inside the prison walls, this should
not be counted in when the severity of the punishment is estimated.
The upshot of these considerations therefore is that we have not found a
satisfactory answer which succeeds in excluding after- and side-effects in terms of
suffering and deprivation, from what should count when punishment severity is
determined. The argument that severity is simply intended severity was proved not
plausible, and the view that suffering or deprivation resulting from a certain
treatment must lie within the context of responsibility did not provide the sought-for
delimitation. Where then does this bring the proportionalist? Must he acknowledge
that the assessment of punishment severity is a complex matter, in the sense that it
might also involve respects to possible after- and side-effects? Before drawing this
conclusion there are some pretty obvious considerations that we will have to engage
in. Rather than considering the question of punishment severity released from more
basic views on justice, an obvious approach to the delimitation question would once
again be to confront the theories on the ground of which the proportionality
principle is held to be justified. The theories should, one might expect, provide us
with an answer as to what should basically count in the assessment of punishment
severity, and consequently with an answer to the delimitation problem. However, as
was the case with regard to the challenge of sensibility, it turns out that there is not
much to be gained from such scrutiny.
According to the simple desert theory, the suffering of the wrongdoer is
regarded as intrinsically valuable and the infliction of deserved suffering is exactly
the purpose of punishment. However, in this view there seems to be no reason to
hold that some part of the suffering which a punishment causes should not count in
the computation of severity. On the contrary; if the whole idea is that justice
requires that a misdeed should be responded to with a fitting amount of suffering,
then there would certainly be good reasons to count in all the suffering that follows
from punitive treatment. As we have seen earlier, Davis defends the theory by
appealing to the intuition that it seems altogether fitting if an irremediably wicked
25
person meets an unpleasant end in some natural disaster. But if the suffering
caused in this way can count as deserved suffering then there is clearly no ground
for disregarding the suffering which results from a traditional punishment in the
form of after- and side-effects.
The same is the case if we turn to the fairness theory. There seems to be no
reason in favour of the claim that such after- and side-effects cannot constitute a
genuine burden and therefore cannot count when the punishment system seeks to
restore the fair equilibrium of benefits and burdens which has been disturbed by the
unfair benefit gained by the criminal. In fact, Sadurski even accepts the idea of
poena naturalis, which certainly confirms that it cannot only be a limited part of
what befalls a person as a result of a genuine punishment which can serve the
function of burden. And Sher’s version of a the fairness view which, by measuring
26
burdens in terms of “usual moral wrongness” , was specifically designed to exclude
such occurrences as previous hardship in a criminal’s life, provides no reason for
why certain after- and side-effects of punishing should not be part of what
determines the usual moral wrongness of this sort of treatment (recall that, as we
have just seen, such effects cannot - at least not always - be held to be something
which was not knowingly caused by a punishment meted out by a sentencer and
something which therefore falls outside the context of responsibility).
Things might seem more promising if we finally turn to expressionism. As
we have seen, what is essential to the expressionist is the deliverance of an
appropriate condemnatory message to a wrongdoer. If there were reasons to believe
that the hardship which results from a punishment in the form of after-and side-
effects could not serve the function of mediating messages, or at least not messages
of the relevant sort, then there would be a ground for distinguishing between
sufferings which should count as part of the punishment and sufferings which
should be regarded as basically extraneous to it, no matter how unfortunate.
However, the problem is that there is no reason to believe that this is the case. Why
should the psychological after-effects of a punishment or certain sorts of stigmatic
after-effects not function as carriers of the message if we in the first place accept
that hard treatment can precisely serve this purpose? For the present, the
expressionist theory is not sufficiently theoretically fine-grained to answer this
question, if an answer can be provided at all. Moreover, neither do any of the more
special expressionist answers to the why-hard-treatment question provide any
assistance. It is hard to think of a reason why suffering caused by some kind of
after- or side-effects could not serve part of a possible penitential function. And,
with regard to the conjecture that hard treatment should provide the criminal with an
extra prudential disincentive to criminal conduct, it certainly seems reasonable to
hold that after- and side-effects may fulfil this purpose. In sum, it seems that neither
of the basic theories of justice provide us with reasons which succeed in excluding
suffering, hardship and deprivation resulting from after- or side-effects of a
punishment as something extraneous to the assessment of punishment severity.
The overall conclusion is therefore parallel to the one drawn in the previous
section. In the same way that there was no plausible ground for holding that
differences in suffering, hardship and deprivation, which result from different
persons’ varying degrees of sensibility to objectively identical punishments, should
not be accounted for when a sentencer metes out a punishment of a certain degree of
severity, there is no ground for claiming that the differences in suffering, hardship
and deprivation which result from punishments which are the same in name but
which, everything considered - due to after- and side-effects - affect persons very
differently (even it they possess the same degree of sensibility), should not figure in
the assessment of punishment severity. Once again, it is worth underlining that in
itself this conclusion is not a problem for the proportionalist. In fact, accounting for
such difference might even strike some as the only plausible interpretation of
proportionality. But the difficulty of assessing severity along these lines is obviously a
severe loss of simplicity in penal practice. However, is this conclusion premature? Is
there still a way in which the proportionality principle can operate with a plausible
116 JESPER
THE ETHICS RYBERG
account of punishment severity but at the same time avoid ending up in such
complexities of application?
3. RESORTING TO MERCY
A criticism which from time to time has been directed against the retributive theory
of punishment is that the goal of upholding justice will sometimes be tantamount to
27
an “inhumane doctrinaire rigorism” . In cases where a punishment will have terrible
consequences for the perpetrator or where it will place a significant burden on the
shoulders of the perpetrator’s family it will nevertheless follow from a retributive
concept of justice that the punishment should be imposed. A criticism of this sort
may perhaps with some force be presented against Kant or other theorists who
28
would subscribe to a principle like fiat justitia, et ruat mundus . However, it is
pretty obvious that this kind of rigorism is not what is typically defended by
retributivists. On the contrary. Several modern versions of retributivism present the
pursuit of justice as a paramount, but not absolute, duty which leaves room for such
concepts as mercy and pardon. In fact, some theorists even hold that the notion of
29
mercy gets a grip only on the ground of a retributivist view of punishment. An idea
of what mercy amounts to is given, for instance, by Hampton who, on the
background of her version of retributivism, stipulates that “mercy is the suspension
or mitigation of a punishment that would otherwise be deserved as retribution, and
30
which is granted out of pity and compassion for the wrongdoer” . More generally,
the view is often that mercy may be exercised on the ground of considerations of,
and compassion for, the plight of a particular offender. For instance, mercy might
perhaps be shown to a person who was severely injured in a car accident even
31
though it was the person’s own reckless driving that caused the accident.
Now, whether a retributivist theory must in some way or another allow for
merciful acting in order not to turn out as morally implausible and, in that case,
under which conditions mercy may then be granted, are not questions which we will
have to consider in the present context. However, if it is the case that actions of
mercy relate to the plight of a perpetrator then it might be suggested that the exercise
of mercy could operate in such way that the problems outlined in the previous
section, concerning the assessment of punishment severity, would be resolved. If a
punishment would have serious after- or side-effects on a perpetrator then this might
be the place where respects to mercy could supply the retributive view on justice in
such a way that the proportionality principle will not face any of the practical
problems sketched out above. More precisely, the argument might take the
following form: the most plausible conjecture as to what the measure of punishment
severity amounts to implies that at least part of the suffering and deprivation which
may be caused by a punishment in terms of after- and side-effects should be counted
in when the severity of a punishment is determined. However, in a case where there
will be such effects, justice should be set aside and mercy should be shown to the
perpetrator. Thus, there is no practical problem for the proportionalist with regard to
accounting for after- or side-effectual suffering since the proportionality requirement
is overruled in exactly those cases where such effects occur. In other words, once
the domain of cases where proportionate punishment should be imposed has been
properly set the problem of delimitation is resolved. A proportionalist could thus
avoid difficulties by resorting to a certain conception of mercy. The question is
whether this would constitute an acceptable way to get around the problem. That the
answer must be in the negative is, I believe, pretty clear.
The first thing that should be noticed is that, though it is generally
acknowledged that merciful actions must spring from considerations which relate to
the plight of a particular offender (or perhaps of other affected persons), the concept
of mercy which is required to fulfil the role suggested in the above argument will
significantly diverge from the content which has typically been given to the concept
by those who consider it important. First of all, it is obvious that mercy in the
present shape does not amount to what is sometimes referred to as mere “legal
mercy”, that is, an instrument which might be applied in particular cases where the
32
rigor juris would otherwise lead to undeserved punishments. Mercy cannot simply
be an instrument to calibrating justice but must itself possess a substantive moral
content. Neither can the exercise of mercy be something which works within the
framework of justice. It has been suggested that there might be a certain leeway with
regard to which punishments are just responses to a certain crime and that mercy
33
could be displayed by imposing the least severe one within this leeway. However,
the present concept of mercy must be one which in certain cases genuinely suspends
the requirement of justice. More noteworthy, neither does this idea of the concept go
well hand in hand with the often presented view that mercy should be considered
34
praiseworthy but not a matter of obligation. If mercy is a form of supererogation
then, unless there are reasons to believe that mercy will actually always be granted
in the case where a punishment has after- or side-effects, there will still be a
delimitation problem for the proportionalist: in cases where mercy is not shown but
where punishment will have such effects, the problem of counting in the extra
suffering or deprivation will still be pertinent. Thus, the present notion of mercy
does not correspond well with some of the traditional ideas of what mercy amounts
to. This, of course, does not show that mercy cannot plausibly play the role which
has been suggested. However, it does not, I believe, require much reflection to see
that such a conception of mercy would be quite unacceptable.
The first obvious problem is that a notion of mercy which would resolve
the problem which proportionalists might otherwise face with regard to accounting
for after- or side-effects in judgements of punishment severity might happen to let
respect to justice play only a marginal role. If those who are punished within the
criminal justice system should be punished in accordance with the requirement of
proportionality, then it will not be sufficient to display mercy by giving an offender
less than the full measure of deserved punishment. On the contrary, mercy must
35
fully exempt the offender from punishment. And more importantly, granting mercy
to offenders will not necessarily be something which occurs only exceptionally. If
all cases where a punishment would result in after- or side-effects of a sort which
would be relevant with regard to assessment of severity should be excluded from the
domain of cases where justice is exercised, then it might turn out, depending on the
118 JESPER
THE ETHICS RYBERG
frequency of such effects, that the imposition of proportionate punishment as a

reaction to criminal conduct would be the exception. Hardly an implication which
proportionalists would find attractive. This highlights the more fundamental
problem underlying the suggestion, namely, that the exercise of mercy along these
lines seems morally unacceptable. That an offender, who has committed a very
serious crime and who will, due to after- or side-effects, experience a little extra
suffering if he is punished, should be exempted from the punishment, while another
offender who will not experience such effects should be given the full measure of
the deserved punishment, is hardly something which a proportionalist will in the end
be willing to accept. Merely to contend that such a concept of mercy would resolve
a practical problem for the proportionalist is clearly not sufficient to carry the
weight of justifying such a practice.
What this indicates is not that the idea of mercy cannot play an important
role within the framework of a complete theory of retributive justice. As mentioned,
several proportionalists believe that there should be room left for mercy within their
theories; and a plethora of suggestions as to when mercy is properly exercised has
been presented. Although these suggestions all seem to agree that mercy must in
some way or another spring from considerations concerning the plight of the
offender it does, however, seem pretty obvious that a notion of mercy which a
proportionalist could resort to as a way of resolving the delimitation problem does
not seem a viable option.
4. CONCLUSION
The purpose of the previous discussion of punishment severity has not been to
suggest a precise and adequate account of the measure of severity. This is a task for
the proportionalist to solve. Rather what I have claimed is that it is hard to think of a
plausible account which does not consider the suffering, hardship, deprivation, or
discomfort which a punished person undergoes, as at least a part of what determines
the severity of the imposed punishment. It was in this connection worth underlining
the warning against inferring from the practical simplicity of something its moral
relevance. It is clearly not plausible to hold that, since it is much easier to measure
water in weight or density than it is to measure degrees of pain, what basically
matters in judging the effect of a harmful act on a person, are the tears which come
from his or her eyes. Similarly, we cannot just exclude suffering as relevant simply
because it complicates the measure of punishment severity. Furthermore, it is worth
underlining that several proportionalists, when touching on the question of severity,
seem to believe that suffering, hardship etc. should count. Neither was this in any
way opposed by confronting some of the more basic theories of justice underlying
the proportionality principle, nor does the literature offer any other arguments which
suffice to undermine the view. Thus, the claim is hard to get around. However, it
has important implications which have been pointed out in the foregoing.
Firstly, a punitive treatment which is one and the same in objective
appearance may, due to differences in physical and psychological sensibility, vary
very much in severity from one person to another. Secondly, a punishment which is
one and the same in name may, due to differences in after- or side-effects, on the
whole affect two persons very differently even if they are physically or
psychologically equally sensible. The point here is obviously not to defend a precise
distinction between differences in outer impact and in inner reactions but rather to
underline that there are several reasons as to why a certain punishment may all in all
affect people very differently. Tonry and Morris’ aphoristic claim that in relation to
36
imprisonment a “year often is not a year” nicely captures the point. And obviously
this fact cannot simply be tossed aside by reference to the law maxim de minimis
37
non curat lex . Such a maxim would itself require a moral justification. And,
furthermore, there is no reason to believe that the application of a standardized
account of severity will result only in trivial, that is marginal or rare, deviances from
proportionality. But this means that the often proclaimed simplicity of applying the
proportionality principle in actual penal practice, and thus one of the advantages
which has been emphasized in comparison to other punishment models such as the
utilitarian, are if not illusory then at least highly overestimated.
120 JESPER
THE ETHICS RYBERG
NOTES
1
See, for instance, N. Morris & M. Tonry, Between Prison and Probation, Oxfrod University Press,
United States of America, 1990; or J. M. Byrne, A. Lurigio, J. Petersilia (eds.), Smart Sentencing,
Sage Publications, United States of America, 1992.
2See, for instance, R. A. Duff, “Alternatives to Punishment - or Alternative Punishments?”, in W.
Cragg (ed.), Retributivism and Its Critics, Franz Steiner Verlag, Stuttgart, 1992.
3L. Sebba, “Some Explorations in the Scaling of Penalties”, Journal of Research in Crime and
Delinquency, vol. 15, 1978.
4M. L Ericson & J. P. Gibbs, “On the Perceived Severity of Legal Penalties”, The Journal of
Criminal Law and Criminology, vol. 70, 1979; D. Buchner, “Scale of Sentence Severity”, The
Journal of Criminal Law and Criminology, vol. 70, 1979; L. Sebba & G. Nathan, “Further
Exploration in the Scaling of Penalties”, British Journal of Criminology, vol. 24, 1984; R. E.
Harlow, J. M. Darley & P. H. Robinson, “The Severity of Intermediate Penal Sanctions: A
Psychophysical Scaling Approach for Obtaining Community Perceptions”, Journal of Quantitative
Criminology, vol. 11, 1995.
5J. Bentham, The Principles of Morals and Legislation, Prometheus Books, 1988, chap. 14, para.
14. (p. 182).
6For a modern discussion on sensibility see, for instance, N. Walker, Why Punish?, Oxford
University Press, Great Britain, P. 99ff.; or A. Ashworth, Sentencing and Penal Policy, Weidenfeld
and Nicolson, London, 1983, chp. 7.
7See, for instance, K. Adams, “Adjusting to Prison Life”, Crime and Justice, vol. 16, 1992.
8A. von Hirsch, M. Wasik & J. Green, “Punishment in the Community and the Principles of Desert”,
Rutgers Law Journal, vol. 20, no. 3, 1989. The same thoughts are presented in the ealier article: M.
Wasik & A. von Hirsch, “Non-custodial Penalties and the Principle of Desert”, Criminal Law
Review, 1988.
9See P. H. Robinson, “A sentencing system for the 21’st century?”, Texas Law Review, vol. LXVI,
1987; and P. H. Robinson, “Desert, crime control, disparity, and units of punishment”, in R. A.
Duff, S. Marshall, R. E. Dobash & R. P. Dobash (eds.), Penal Theory and Penal Practice,
Manchester University Press, Great Britain, 1994.
10A. von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993, p. 33ff; or A. von Hirsch,
“Seriousness, Severity and Living Standard”, in A. von Hirsch and A. Ashworth (eds.), Principled
Sentencing, Hart Publishing, Oxford, 1998.
11It should be mentioned that von Hirsch rightly rejects a version of subjectivism according to
which all there is to punishment severity is what someone judges when a certain punishment has
been described. Obviously this kind of subjectivism is implausible. But rejecting this does not show
anything with regard to the versions of subjectivism according to which it is the actual suffering or
deprivation caused by a punishment that determines its severity.
12Obviously, this is not the same as saying that objectivist living-standard analyses may not play a
legitimate role in parts of political decision-making.
13What he says is that the analysis ”does not focus on actual life-quality ... but on the means or
capabilities for achieving a certain quality of life”. A. von Hirsch, Censure and Sanctions,
Clarendon Press, Oxford, 1993, p. 30.
14A. von Hirsch , Doing Justice, Hill and Wang, Unites States of America, 1976, p. 90.
15It is in this connection worth noticing that von Hirsch and other proportionalists actually accept
the idea of a day-fine system.
16I shall, in chapter 5, consider more thoroughly the problem of just punishing in an unjust society.
17See also D. N. Husak, “Already Punished Enough”, Philosophical Topics, vol. 18, no. 1, 1990.
18A. Ashworth, Sentencing and Criminal Justice, Butterworth, London, 1995, chaps. 3 and 7.
19J. Bentham, The Principles of Morals and Legislation, Prometheus Books, United States of
America, pp. 44-45 (chap. 6 § 6).
20N. Walker, Why Punish?, Oxford University Press, Great Britain, 1991, p. 108.
21H. L. A. Hart, Punishment and Responsibility, Clarendon Press, Great Britain, 1968, p. 5.
22The same answer has been given to reject the idea that the possible effects a punishment might have on
other parties - the perpetrator’s family and friends etc - should in themselves be considered a punishment.
For instance, Primoratz contends that to talk of punishment in such cases “is confused, for the suffering of
the offender’s family is not part of his punishment but an unintended and regrettable consequence of it”.
I. Primoratz, Justifying Legal Punishment, Humanities Press International, United States of America,
1990, p. 109.
23In fact, this conjecture would even eliminate the challenge of sensibility. The fact that two persons
may, due to difference in sensibility, be very differently affected by objectively the same punishment,
simply does not imply that the one is punished more harshly than the other. If the intended suffering is the
same, then they will be equally severely punished independently of what they each actually have to go
through.
24See, for instance, J. Kleinig, Punishment and Desert, Martinus Nijhoff, The Hague, 1973, p. 17f.
25L. H. Davis, “They Deserve to Suffer”, Analysis, vol. 32, 1971-2.
26G. Sher, Desert, Princeton University Press, United States of America, 1987, p. 84.
27B. Blanshard, “Retribution Revisited”, in E. H. Madden, R. Handy & M. Farber (eds.), Philosophical
Perspectives on Punishment, Charles C. Thomas Publisher, United States of America, 1968, p. 78.
28”Let the world perish, so long that justice be done”.
29See, for instance, A. Smart, “Mercy”, Philosophy, vol. 43, 1968.
30J. Hampton, “The Retributive Idea”, in J. G. Murphy and J. Hampton, Forgiveness and mercy,
Cambridge University Press, United States of America, 1988, p. 158.
31For other examples, see A. Smart, “Mercy”, Philosophy, vol. 43, 1968.
32See, for instance, C. A. H. Johnson, “Entitled to Clemency: Mercy in the Criminal Law”, Law and
Philosophy, vol. 10, 1991.
33See S. Hestevold, “Disjunctive Desert”, American Philosophical Quarterly, vol. 20, 1983; or S.
Hestevold, “Justice to Mercy”, Philosophy and Phenomenological Research, vol. 46, 1985.
34For a discussion of the possible deontic implications of mercy see, for instance, G. Rainbolt, “Mercy:
In Defense of Caprice”, Nous, vol. 31 no. 2, 1997.
35As mentioned, retributivists often hold that mercy is a mitigation of punishment. But keep in mind, that
we are here considering mercy as a way of resolving the delimitation problem by ruling out all the
practically complicated cases in which after- or side-effects should otherwise be accounted for in the
computation of the proportionate punishment.
36N. Morris and M. Tonry, Between Prison and Probation, Oxford University Press, United States of
America, 1990, p. 94.
37”The law takes no account of trivialities”.
CHAPTER 4
THE ANCHOR PROBLEM
The ethical scrutiny in which contemporary proportionalists have been engaged

obviously has a clear practical side. By considering what constitutes a plausible
principled foundation for the distribution of punishments, the goal is to contribute to
the development of a sentencing system. Though reflections on this kind of
foundation, as former chapters clearly illustrate, open up a myriad of basic ethical
questions - a solution to which would in itself constitute a sufficient goal for the
purely philosophically minded - proportionalists have succeeded in keeping the
practical goal clearly in focus. In fact, alongside the philosophical investigations,
proportionalist discussions have now and then been accompanied by specific
suggestions as to how severely one should expect certain crimes to be punished on a
proportionalist ground and, more generally, how tough proportionalist penalty levels
would be in comparison to the currently existing levels. Even though there is not
complete agreement on these matters, there are indeed certain tendencies in the
recommendations.
Despite the fact that the growing interest in a just desert theory has, in some
countries, coincided with a toughening of sanction levels, and that the desert theory
1
has sometimes even been accused of leading to tougher punishments , there has been a
clear tendency, amongst proportionalists, to argue for a substantial reduction in
sanction levels. For instance, Singer suggests that confinement should be reserved
only for the most serious crimes and that, even then, its duration should be relatively
short. As he underlines, it is a misconception to think of the desert model as a
2
derivative of the “throw away the key” approach to punishment. Likewise, Murphy
claims that if the desert theory was followed consistently one would probably punish
3
less and in more decent ways than one actually does. And in Doing Justice von
Hirsch offers proportionalism as a means to restricting punishment, suggesting,
more precisely, that terms of imprisonment even for the most serious crimes should
seldom exceed five years; a position he has maintained while his theory has been
4
developed over the years. Thus, the general view is clearly that proportionalism
should not be considered a draconian theory.
The purpose of this chapter is not to compare proportionalist
recommendations and the actual sanction levels in different countries, but rather to
clarify how proportionalists in the first place believe that specific judgements as to
how different crimes should be punished can be reached. In other words, we now
have to consider how crimes and punishments, once they have been scaled in
accordance with proportionalist prescriptions, should be linked up. That a link is a
123
124 JESPER
THE ETHICS RYBERG
precondition for any practical punishment guidance is obvious. However, before

entering the discussion one might ask whether proportionalists need to engage in
further theoretical considerations in order to deliver this guidance. What we have
been concerned with in the previous chapter is the ordinal or comparative aspect of
proportionalism, that is, the claim that a crime which is more (or less) severe than
other crimes should be responded to by a punishment which is more (or less) severe
relative to other punishments, and that equally serious crimes should be equally
severely punished. Will this itself suffice as a theory of punishment distribution?
Several decades ago, Mundle apparently believed so. What he contended was that
“[i]n order to apply the principle of proportion, all that is necessary is that we
should be able (a) to compare different offences in respect of their relative moral
gravity, and (b) to compare different penalties in respect of their relative
unpleasantness - and surely we can make such comparisons in many, if not all,
5
cases” . However, there are good reasons to be sceptical with regard to this outlook,
that is, with regard to the claim that this is “all that is necessary”.
Firstly, purely ordinal proportionality would not provide any guidance as to
6
how a crime of a certain degree of seriousness should actually be punished. To this
it might perhaps be replied that, as long as ordinal proportionality constraints are not
infringed, nothing further is required from a theory of punishment distribution.
After all, it is not a necessary condition for the plausibility of an ethical theory that
it provides only one determinate answer as to how we should act. However, not
many would be willing to accept this answer. Surely, a more precise guidance on the
matter would constitute a theoretical virtue. Secondly, and more importantly, pure
ordinal proportionality would be compatible with very lenient and extremely tough
punishments. There would be nothing to object to in a system which would punish
the most serious crime with a minor fine or, alternatively, a system which would
punish a minor crime with many years in prison or even with the death penalty, as
7
long as ordinal proportionality was observed in both cases. Many would hesitate to
accept these implications. And, as the mentioned recommendations of reductions in
penalty levels indicate, proportionalists would also regard such implications as
unacceptable. Thus, even though we have initially defined proportionalism in terms
of ordinal proportionality in order not to bind up the principle with one particular
view on cardinal proportionality, the above arguments clearly show that further
theory is needed in order to reach a complete proportionalist theory of punishment
distribution.
In the following, it will be discussed what proportionalists in this respect
have had to offer. This will be done, firstly, by considering the significance of what
kind of scales crimes and punishments are held to be measurable on and, secondly,
by considering in turn the two main proportionalist suggestions as to how the scales
should be anchored. As I shall argue, at this vitally important point, proportionalists
face various problems.
THE ANCHOR PROBLEM 125
1. RATIO, INTERVAL, AND ORDINAL MATCHINGS
The answer to the question of how scales of crimes and punishments should be
linked, and what theoretical problems interscalar judgements give rise to, very much
depends on what kind of matching proportionalists advocate, that is, whether they
hold that the seriousness of crimes and the severity of punishments are measurable
on ratio, interval or ordinal scales. We can start by considering the sort of matching
which has traditionally been associated with proportionalist thinking, and then turn
to the matching to which most recent proportionalists subscribe.
The question of how a crime scale and a punishment scale should be
connected could be given a very simple answer if it was possible for each separate
crime, independently of prior comparative judgements, to mete out the fitting
punishment. In the same way as the size of a person’s foot provides sufficient
information to produce a shoe that fits, we can imagine that the relevant information
about a particular crime would in itself be sufficient to determine what would
constitute the appropriate punishment. A matching of this kind - which I shall
henceforth refer to as a “direct matching” - is provided by lex talionis. Though it has
been pointed out that the classical biblical formulation “eye for eye, tooth for tooth”
should, in the historical context, be understood not as a principle prescribing exact
retribution but rather as a way of restraining disproportionate cruel retaliation -
several eyes for an eye - it is nevertheless obvious that a literalistic interpretation of
“like for like” would constitute a simple way of linking a crime directly to a certain
8
punishment. However, as Kleinig has correctly pointed out, it is hard to find any
proportionalists who in a strict sense believe that what was done to a victim should
be done to the criminal. Not even the classical retributivists adhere to such a
principle - albeit they have sometimes been unfairly accused of doing so. Though
Kant comes very close to the view when he insists that a principle of retaliation “can
assign both the quality and quantity of a just penalty”, he nevertheless recommends
a number of punishments which violate the definiteness of lex talionis literally
9
understood. And Hegel explicitly says that he regards a strict interpretation of
retaliation as a way of exhibiting “the retributive character of punishment as an
10
absurdity” .
Indeed there are good reasons not to accept lex talionis in this sense. The
standard objection has been that lex talionis violates “ought implies can”. There are
a large number of cases in which it is simply impossible to reverse a crime against
the criminal himself. As Kleinig points out, it is hard to see what punishment one
should inflict on “a rapist, a blackmailer, a forger, a dope peddler, a multiple
murderer, a smuggler, or a toothless fiend who has knocked somebody’s else tooth
11
out” . And it is not hard to imagine many other examples of crimes which do not
readily admit of strict retaliation. However, what is much more important than this
traditional criticism is the fact that it is very hard to see how strict retaliation can be
ethically sustained at all. No matter whether proportionalism is based on the simple
desert theory, the fairness theory, a version of expressionism, or other accounts of
retributivism, there seems to be no reason for holding that a criminal should be
repaid by doing to him what materially amounts to the same as the crime committed.
126 JESPER
THE ETHICS RYBERG
Furthermore, this version of lex talionis might even contradict the

12
requirement of ordinal proportionality. Suppose that A has committed a crime
which is more severe than a crime committed by B. For instance, A has stolen a
large amount of money from someone while B has stolen only a very small amount.
Suppose further that A is in fact extremely rich while B is quite poor. Now,
requiring from A an amount corresponding to what he stole might cause him much
less hardship than would be the case with B. If the severity of a punishment is
determined on the ground of the degree of hardship which it causes then B will, in
this situation, be punished more severely than A. This contradicts the ordinal
proportionality constraint. But the fact that a literalistic interpretation of lex talionis
may in this way be in direct conflict with ordinal proportionality is itself sufficient
to exclude such a linking of crime and punishment as a viable option for the
proportionalist.
The tempting reaction to these objections, and the approach which has
actually been adopted both by classical and a number of modern retributivists, is to
defend lex talionis in a form which does not require that materially the same should
be done to the criminal as was done to the victim. This is what Hegel, on the
grounds of his theory that punishment annuls crime, had in mind when he said that
“[i]n crime ... the purely external specific character vanished all the more obviously,
and equality remains the fundamental regulator of the essential thing, to wit the
deserts of the criminal though not for the specific external form which the payment
of those deserts may take. It is only in respect of that form that there is a plain
inequality between theft and robbery on the one hand, and fines, imprisonment, &c.,
on the other. In respect of their ‘value’, however, i.e. in respect of their universal
13
property of being injuries, they are comparable” . By modifying lex talionis in this
way, as a principle which prescribes only that a punishment should equal or be
equivalent to a crime, one obviously avoids the most immediate objections
confronting strict interpretation. The problem of how one should punitively respond
to rape, blackmail, forgery and several other crimes, might well - by abandoning the
requirement of similarity in quality - be resolved.
Moreover, a principle requiring equivalence might be held not to suffer
from the same ethical arbitrariness as strict interpretation. For instance, Reiman
defends the principle on the ground of a Hegelian-inspired conception of “equality
of persons”14. In his view, there is affinity between the golden rule which mandates
that you should “[d]o unto others as you would have other do unto you” and lex
talionis which prescribes that you “[d]o unto others as they have done unto you”.
Both principles spring from the view that persons are basically equal. The claim is
obviously not that a punishment in a simple way restores the status quo ante - the
punishment does not leave the crime undone or heal the injury which the victim has
suffered - but rather that “it rectifies the indignity he has suffered, by restoring him
15
to equality with you” . (Noteworthily, Reiman even mentions the possibility of
responding punitively to a person, who has committed several murders, by bringing
him to within an inch of death and reviving him as many times as he has killed
(minus one) and then finally execute him. Though in the end he rejects this
possibility, this is, however, not because it conflicts with the version of lex talionis
that he defends, but rather because such an extreme action is excluded by other
moral considerations which, in his view, trump respects to justice.) Now, does lex
talionis interpreted as a principle prescribing equivalent punishment provide a
plausible ground for linking crimes and punishments?
There are several reasons for being sceptical about this suggestion. As was
considered in chapter 2, there is a problem of limiting the relevant amount of
suffering caused by an act. Reiman specifically mentions that the harm caused by an
offender, for which he should be paid back, is not limited to the harm done to the
immediate victim: it may as well include “the suffering of the victim’s relatives or
16
the fear produced in the general populace, and the like” . However, as we have
seen, what would usually be considered minor crimes may trick a causal chain of
events which in the end have fatal consequences. Thus, unless Reiman or other
adherents to a equivalence versions of lex talionis are willing to accept that all the
suffering that follows from an act should count in the assessment of the seriousness
of the conduct and consequently in the determination of the equivalent punishment,
the problem of limiting the relevant harm in a non-arbitrary way remains pertinent.
What is more important, however, is that a demand of equivalence faces a
problem concerning the complexity of judgements on the seriousness of crimes.
Certainly, only a few modern proportionalists would hold that all that should count
in the ranking of crimes is the amount of harm that is caused. As we have seen,
culpability is standardly considered a seriousness determining component as well.
But this complicates the concept of equivalence. Suppose that the seriousness of
crimes is measurable on a ratio scale. That is, that it is possible to make judgements
of the kind that one crime is three times as serious or half as serious as another. And
suppose further that the severity of punishments is measurable on a ratio scale. How
severe should a punishment then be in order to be equivalent to a crime of a certain
degree of seriousness? If lex talionis was understood as a principle prescribing that
we should literally treat like for like, then it would be - if we ignore the mentioned
practical problems - possible to determine exactly which punishment fits which
crime. But, if seriousness is determined both by harm and culpability, there no
longer seems to be a simple answer to how much on the scale of severity is
equivalent to how much on the scale of seriousness. That fact that one would be
able to claim that one crime is three times as serious as another and that it should
therefore be responded to by the infliction of a punishment which is three times as
severe, does not provide an answer as to how severely the crime should actually be
punished. Thus, the rejection of the strict and implausible interpretation of lex
talionis in favour of the equivalence interpretation is apparently tantamount to a loss
of a simple way of linking crime and punishment. Or, more precisely, either one
will have to interpret equivalence as a strict harm-for-harm equivalence which
would make clear what the principle implies but at the same time would exclude
considerations on culpability which retributivists usually regard as crucial, or one
can maintain that crime seriousness should also be determined by culpability, in
which case the concept of equivalence is left theoretically obscure. It is this latter
point which some of the very early opponents of retributivism had in mind in their
rejection of proportionate punishment. For instance, more than half a century ago,
Maclagan claimed that retributivism is impossible of application because it
128 JESPER
THE ETHICS RYBERG
presupposes a notion of equivalence which is itself impossible since ”[t]he two

things that are to be measured against each other are in their very nature
17
incommensurable” . The same criticism was put forward by Rashdall and Ferri
about a century ago and is sometimes still presented in modern debate. However,
though there certainly is something to this criticism when it is directed against the
idea of equivalent punishing, it is also clear that this does not suffice as a general
criticism against proportionalism.
As we shall see, most modern proportionalists do not defend the idea of
equivalence which the criticism is targeting. Moreover, the fact that the degrees of
seriousness of crimes and severity of punishments are not commensurable does not
justify the conclusion that the two scales cannot be linked. On the contrary, a theory
of anchor points may do the job. For instance, if it was possible in some way or
another to argue for what would be the fitting punishment for a particular crime,
then it would be possible, if one ex hypothesi assumes that seriousness and severity
can be measured on ratio scales, to reconstruct the full matching of the two scales.
That it is possible to identify this kind of anchor point, despite the fact that
seriousness of crimes and severity of punishments are not straightforwardly
commensurable, is exactly what modern proportionalists claim. Thus, rather than
simply rejecting proportionalism, the more proper provisional conclusion is that in
order to avoid problems of incommensurability proportionalists have to supply us
with a theory of anchor points which manages to identify the points at which the
two scales should be pegged together. And notice that the identification of one
anchor point was based on the presupposition of a ration matching. If this
presupposition is rejected things becomes more complicated.
As a matter of fact, the majority of recent proportionalists do not believe
that the seriousness of crimes can be measured on a ratio scale. Rather they defend a
position which accords well with the way in which we perhaps more commonly
make comparative judgements about different crimes. While most people would
probably hesitate to make ratio-judgements of the kind that, for instance, rape is
three times as serious as burglary or half as serious as murder, they would usually
allow for a stronger sort of comparison than the purely ordinal claim that one crime
is more (or less) serious than another. For instance, many would be willing to accept
that it makes sense to claim that, of three crimes ranked in ascending order of
seriousness, the second is only slightly more serious than the first while the third is
considerably more serious than the second. This would probably the case if the
crimes compared are the theft of 50$, embezzlement of 100$, and rape. In other
words, the claim is that judgements of comparative seriousness should account for
the spacing between various crimes on the scale. This is the view which is defended
by Kleinig, who characterizes the scaling he suggests as something in between an
18
ordinal and an interval scale. In the same way as we may be able to judge, by
observing chess players, that the superiority of experts over beginners is greater
than their inferiority to masters, we can, he suggests, make judgements concerning
the comparative differences in seriousness between different crimes. The same view
is presented in Ten’s defence of a “thick version of the proportionality principle”,
and in many other proportionalists’ accounts of how punishments should reflect
19
differences in the seriousness of crimes. Now, if we accept this position, namely
that crimes and punishments are respectively measurable on interval scales, then the
obvious question is how exactly should the two scales be related?
As mentioned, the standard answer is that certain anchor points are
required. However, as has correctly been observed, the anchoring of one
punishment to one crime is not sufficient to construct an interval matching of the
scales. The problem is that the differences in seriousness between different crimes
may be reflected in the punishments in many ways. Suppose, for reasons of
simplicity, that the crime scale contains only three crimes C1, C2, and C3 ranked in
ascending order, and that the relative interval between C1 and C2 is three times as
large as the interval between C2 and C3. Now, if we imagine that it was in some way
established that the appropriate punishment for C1 is 1 year in prison, then the
intervals would be reflected in the punishments both if C2 is punished by 1 year and
3 days in prison and C3 by 1 year and 4 days in prison, and if C2 is punished by 31
years in prison while C3 is responded to by 41 years of imprisonment. The relative
spacing of the crimes in all cases is the same (three to one).
Therefore, in order to provide guidance as to how different crimes should
be punished and perhaps to exclude punishments which are either too tough or too
lenient but which might be consistent with an interval matching linking the two
scales with only one anchor point, it seems that at least two anchor points are
required. If we imagine that it is in some way possible to argue both that the
appropriate punishment for C1 is 1 year in prison and that for C3 it is 5 years in
prison then, given the suggested intervals on the seriousness scale, it would follow
that C2 should be punished with 4 years of imprisonment. And in the same way it
would be possible to determine the punishments for all other crimes that might be
ranked on the scale. Thus, what we can conclude is that, in so far as proportionalists
hold that crimes should be ranked on an interval scale of seriousness and that this
ranking should be reflected in the way in which the crimes are punitively responded
to, some arguments which succeed in identifying at least two anchor points are
20
required in order to answer the question of the matching of the two scales.
What then if we turn to the final approach to the question of what kind of
scale crimes are measurable on, namely, the view that crimes can be compared in
seriousness only on a purely ordinal scale. In consideration of the problems which
we have seen that relate to the construction of a scale of crimes in gravity, this
approach might at first sight seem more attractive than the more ambitious interval
or ratio scalings. However, once the requirement with regard to the measurement of
seriousness is in this way modified, even more is required in order to provide a
plausible answer to the anchor problem. Now it seems that not even two anchor
points will be sufficient. While it is, as we have just seen, possible to construct the
whole matching of interval scales on the ground of two anchor points, this is not the
case if crimes and punishments are only ordinally scaled. If the crimes C1 to C6 are
ranked in ascending ordinal order, and if it is in some way possible to argue that C1
should be responded to by a minor fine and that the appropriate response to C3 is 30
days in prison, then the crimes C4 to C6 might be punished respectively with 31, 32
and 33 days in prison or with 30, 40 and 50 years in prison. Each possibility would
130 JESPER
THE ETHICS RYBERG
respect the ordinal ordering and be consistent with the two suggested anchor points.
And even if we alternatively imagine that the anchor points are the punishments for
the least and the most severe crimes, that is, for instance, that it has been argued that
the appropriate punishment of C1 is a minor fine and that the appropriate
punishment for committing C6 is 10 years in prison, then this would still not be
sufficient to provide a morally plausible matching of the scales. The suggestion, as
long as the ordinal order is reflected in the imposition of punishments, would be
consistent both with punishing the crimes C2 to C5 with minor fines, and with
responding to the crimes with punishments just below 10 years in prison. It is
exactly because this kind of bunching of punishments is regarded as implausible
that most proportionalists maintain that the matching of the scales must account for
21
the spacing between the crimes on the scale. What we can once again conclude is
that, in so far as one holds that seriousness is measurable only on an ordinal scale,
further moral considerations are required in order to answer the anchor problem.
And now it seems that even more is required than merely establishing the existence
22
of a few anchor points.
What we can conclude on the ground to the previous considerations is the
following. In order to answer the question of how the crime scale and the
punishment scale should once constructed be linked, there seems from the outset to
be two possibilities. One approach is to claim that for any particular crime it is
possible in a direct way to mete out the appropriate punishment. That is, once we
have obtained knowledge about the relevant aspects of the crime this is in itself
sufficient to determine how the person who has performed the wrongful act should
be punished. By making such calculations for all the different crimes we will end up
with a complete matching of the scales. This kind of direct linking would follow
from lex talionis. However, as we have seen, lex talionis, strictly interpreted, is
implausible. In fact, doing to a criminal what was done to the victim might even
violate the ordinal proportionality constraint. And the less rigid interpretation
prescribing equivalence between crime and punishment also failed to provide us
with plausible guidance. The alternative approach, therefore, is to give up the
thought of a direct linking in favour of some sort of indirect linkage through anchor
points. In the case of a linking of ratio scales which are not in a simple way
commensurable, this approach will require at least one anchor point. If the two
scales are interval scales it will require two anchor points. And, in the case of an
ordinal matching, even more will be required to link the scales in a morally
plausible way. Considering these three approaches, most proportionalists adhere to
the second. To require a matching as strong as a ration matching would, given the
outlined problems associated especially with the comparison of crimes in gravity,
seem like a dubious aspiration. And a mere ordinal matching would require much
more of the anchoring theory if we wish to avoid “bunching-problems”. Thus, it is
not surprising - and as indicated in accordance with our intuitive judgements - that
an interval matching is the one typically advocated. What an anchoring theory based
on this type of matching more precisely would amount to is the question to which
we shall now turn.
2. ANCHOR POINTS AND HUMAN DIGNITY
As mentioned, practically all proportionalists believe that it is not possible in any

direct way to determine the appropriate punishment for a given crime. The
alternative suggestion, as we have just seen, therefore, is that the crime- and
punishment scales should be linked by a sufficient number of anchor points. Such
an approach obviously raises the question concerning the ground on which anchor
points can be determined. A question which at first glance seems somewhat
intricate. After all, if it is not in general possible to link crimes and punishments in a
direct way, what reason could there possibly be for claiming that it is possible to
determine for one or a few crimes the appropriate punitive responses?
The suggestion which we shall consider in the following was made in an
early outline by Kleinig and later adopted by a number of punishment theorists. It
was suggested that the poles of the respective scales should serve as anchor points.
The idea is clearly formulated by Scheid in his more recent defence of the
suggestion in which he says that “[t]he main strategy for solving these problems is
to consider a scale that includes all crimes and a scale that includes the full range of
acceptable punishments, and then anchor the scales to each other at two points. We
anchor the most severe punishment to the most serious crime (say, life imprisonment
for murder) and the least severe punishment for the least serious crime (say, fine for
23
petty larceny), with the other crimes falling in between” . Given an interval matching,
two anchor points would, as we have seen, be sufficient to make the punishments for
all other crimes fall into place. The challenge facing this proposal therefore is to
establish that it actually makes sense to talk about upper and lower limits of
24
seriousness and severity.
At the outset the idea of limits on the seriousness of crimes is surprising.
With regard to the lower limit it might be claimed that this should be determined by
the point below which it no longer makes sense to criminalize an action. Where
exactly the limit should be can, of course, be discussed but it certainly does make
sense to talk about such a limit. It is more unclear what should determine the upper
limit. If it is agreed that harm is at least one of the aspects of a crime determining
it’s seriousness, then what does it mean to set an upper limit to seriousness? For any
possible crime it always seems possible to imagine another more harmful. Thus,
should the limit be set at one murder, two murders or twenty? Or perhaps at crimes
against humanity? When it comes to arguing where the limit should go, adherents of
the suggestion have often not been very clear on this. A possible answer might be to
rely on some sort of standardization. As we have seen in the discussion of crime
ranking, judgments of seriousness are often claimed to involve some degree of
standardization. Along the same lines it might be suggested that what should count
as the most serious crime must be one which perhaps not typically, but at least with
some not too small frequency, can be expected to occur. This will exclude the
twenty killings and would perhaps leave murder as the upper limit. In fact, this is
what Scheid suggests.
The most controversial aspect of this suggestion, however, does not
concern the limits of crime seriousness but rather the limits of severity of
132 JESPER
THE ETHICS RYBERG
punishment. Once again the lower limit is probably the least problematic. In so far
as it is possible to specify what constitutes the purpose of punishing someone, it
apparently makes sense to contend that the limits should be set at the point below
which this purpose can no longer be fulfilled. For instance, Scheid contends that the
minimum punishment “will be that penalty which is practicable to impose and that
25
will still count with people as a punishment” . But what about the upper limit? Can
we morally make sense of the concept of a maximum punishment? If that is the case,
then it seems that we should not expect this punishment to be determined purely on
the ground of desert. If it was possible to determine which punishment would be
deserved if someone has committed the most serious crime, then it seems that it
would just as well be possible to determine what would be the deserved punishment
26
for any other crime on the scale. But as we have seen, those who defend the
outlined approach to the anchor problem do not believe that crimes and punishments
can be linked directly. Rather, in exactly the same way as, for instance, the amount
of money distributed in the first and last bonuses in a company’s reward scheme
may be determined not on the ground of desert but on the ground of the amount of
money that is available, the view is that the maximum punishment must be
determined on the grounds of considerations extraneous to desert.
The basic idea, that there are certain limits to what kind of punishment it
can ever be justified to impose on someone, independently of their desert debts, has
been given a number of different formulations. A standard reason for the existence
of such limits is that there are punishments which violate or degrade human dignity.
Another reason often referred to is that there are some punishments which it is
wrong in principle to impose because these sanctions are basically inhumane. The
idea of such restrictions are also found in the much discussed Eighth Amendment of
the U.S. Constitution which enunciates that “Excessive bail shall not be required,
nor excessive fines imposed, nor cruel and unusual punishments inflicted”. In the
following, I shall focus primarily on the concept of human dignity. In explications of
some of the other concepts referred to in the discussion of a ban on certain
punishments, dignity often plays a basic justificatory role. For instance, in his
27
influential article “Cruel and Unusual Punishment” , Murphy explains
“dehumanizing” punishment as punishment which is inconsistent with human
dignity. And part of the interpretation of the prohibition of cruel and unusual
28
punishment has often been that punishment must accord with the “dignity of man” .
However, my point is not to argue against those who wish to hold that there is an
important distinction between inhumane treatment and treatment which degrades
dignity. Rather, since the concept of dignity often figures as a basic concept in the
discussion and, more importantly, since the questions to any such concept which
plays a justificatory role in relation to a moral restriction of punishment, are
basically the same, it will be sufficient to keep the main focus on the notion of
dignity.
Before entering the discussion, there is one further point worth making
concerning the relevance of such an analysis. In the previous chapter we considered
the ranking of punishments in severity. As we saw, many modern proportionalists
believe that such a scale might contain many sorts of punishment. For instance,
different kinds of intermediate sanctions. However, what was not discussed was
whether there are certain kinds of punishment which ought not to figure on the
proportionalist punishment scale. A traditional dispute, of course, is whether or not
one should accept the death penalty. But the question goes much further. What
about cutting off a perpetrator’s hand? Or punishments such as flogging or other
kinds of more or less brutal torture? For these kinds of punishment it need not be the
case that they exceed an upper limit in terms of severity. For instance, in terms of
harm it certainly need not be the case that several years in prison is a more severe
punishment than flogging. Nor does it seem reasonable to contend that different
sorts of torture cannot function as punishment, or that such treatment can never be
deserved. But several proportionalists would nevertheless be reluctant to include
these punishments on the scale, on the ground that such treatments are inconsistent
with human dignity. Thus, in this way a scrutiny into dignity might for some
proportionalists be doubly motivated. Not only does it help in resolving the anchor
problem, it also provides a ground for identifying unacceptable forms of treatment
which should be excluded from the punishment scale. However, I shall here stick to
the first issue.
That the concept of dignity is more generally regarded as being of moral
importance is witnessed by the fact that appeals to the concept are rather common.
There are references to dignity in the constitutions of various countries (e.g. Ireland,
Portugal, Spain, Germany and Sweden) and the Universal Declaration of Human
Rights of 1948 proclaims, in its Preamble, the “inherent dignity” of “all members of
the human family”. Over the last few decades references to the concept have
frequently been made in reports and declarations concerning the development and
use of modern biotechnology. However, as one might well expect, these works do
not contribute philosophical reflections on what dignity consists in and what
precisely it implies. What is more surprising is that such considerations are
sometimes ignored even in philosophical contexts, such as, for instance, in
bioethical discourse where references to dignity are numerous. As Kuhse has
pointed out, the vagueness of the term and the pathos it inspires seem to be part of
29
its attractiveness. The term, she holds, sometimes functions as a “conversation
stopper” which hinders rather than contributes to a rational resolution of ethical
problems. If this is correct, then obviously it is philosophically unsatisfactory.
In so far as references to dignity are to be of help in ethical discussions,
what is required is a clarification of, firstly, what exactly dignity consists in and,
secondly, what it implies with regard to how we should act, that is, which actions are
those that infringe human dignity. That we cannot at these points merely assume that
we are on firm philosophical ground is indicated by the fact that there are
disagreements on what dignity implies both amongst those who regard it is a
morally important concept in relation to punishment as well as in relation to other
areas of applied ethics. Strictly speaking, a third question also arises, namely, why
actions which in some way or another infringe someone’s dignity should be morally
prohibited. This is, of course, a crucial question; however, in the following I shall
for the sake of the argument simply assume that this is the case. This leaves as our
primary concern the clarificatory question of identifying which actions belong in
this category, that is, which punishments should be banned as a matter of principle.
134 JESPER
THE ETHICS RYBERG
One of those who have contributed to a preliminary clarification of the

30
concept of dignity is Meyer. As he points out in his analysis, a first overall
distinction in the way dignity is attributed to humans is between “acting with
dignity” and “having dignity”. The first sense is exemplified in the way that we
sometimes say of a person, he “speaks with dignity” or that he “carries himself with
dignity”. That it is not dignity in the adverbial sense of the term which is thought of
in relation to the question of which punishments should be banned is pretty obvious.
The second sense is clearly closer to what theorists usually have in mind. However,
as Meyer points out, the idea of “having dignity” is itself open to several
interpretations. On one interpretation, which perhaps appears a little old-fashioned,
a person has dignity by occupying a position of high rank in an established social
hierarchy. On another interpretation, “having dignity” refers to the notion of having
a “sense of dignity” that is understood as having a sense of some sort of self-worth.
Finally, in a third sense, “having dignity” expresses the idea of someone having
worth. This final interpretation diverges from the idea of social dignity by
conceiving dignity as something which everybody possesses. This is what is
expressed by talking about human dignity. Likewise, this conception of dignity is
not dependant on whether or not someone actually has a sense of dignity. If a person
lacks a sense of dignity this does not imply that he thereby fails to have dignity in
this third sense. When theorists refer to dignity in relation to punishment it is clearly
dignity in this final interpretation they have in mind. However, pointing at this sense
of the concept serves only as a first preliminary step in an analysis. Though, the
mentioned distinctions might remove some confusions in the discussion of dignity,
they do surely not bring us to a point at which the concept can in any way serve us
in the evaluation of actions. Thus, what can we make of this idea of possessing
human dignity?
Without doubt, the most influential proponent of the view that all human
beings have dignity is Kant. Though he did not compose a special treatise on the
subject but wrote only a few pages about it, it is nevertheless his idea which seems
to underlie much of the modern ethical discussion of the concept. Kant’s
Groundwork of the Metaphysic of Morals introduces dignity in relation to the
second formulation of his categorical imperative - sometimes referred to as “the
31
dignity principle” - according to which man is to be treated as an end in himself
and never as a means only. As such, man possesses in Kant’s sense a dignity which
he defines as “unconditional and uncomparable worth”. Dignity is contrasted with
“price” or conditional value. Anything with mere price has a value that “admits of
equivalents”, which means that it may be subject of calculated trade-offs. No matter
how great the value, in principle there is something else that could justify its
sacrifice. Dignity, on the other hand, makes an object irreplaceable. Furthermore,
Kant believes that only human beings can have such dignity, and that they possess
their dignity in virtue of their autonomy which, roughly speaking, consists in the
capacity for practical rationality. In short, dignity is basically grounded on the
human capacity for moral deliberation. Why precisely it is that autonomy in Kant’s
sense entails dignity is far from clear. However, in the present context we can leave
this question aside. What is more important is whether this notoriously influential
account of dignity - as a contrast to price - can provide any guidance in the question
of punishment.
Though it is in itself not indisputably clear what it means to treat someone
as an end or as a means only, it seems that the idea does have something to offer in
32
relation to punishment. For instance, punishing someone for purely preventive
reasons does not seem to accord with the principle. In his Rechtslehre Kant himself
says that “[j]uridical punishment can never be administered merely as a means for
promoting another good either with regard to the criminal himself or to civil society,
but must in all cases be imposed only because the individual on whom it is inflicted
has committed a crime. For one man ought never to be dealt with merely as a means
33
subservient to the purpose of another ...” . However, though the idea may have this
kind of implication with regard to what is morally acceptable, it is important to keep
in mind what the question which we are now considering is, namely, whether it is
possible to draw a line between those punishments which are wrong in principle and
those which may be inflicted on a perpetrator. However, whether someone is treated
as an end or merely as a means does not seem to depend on the level of severity
itself, but rather on the background against which a certain punishment is inflicted.
If a person is punished for purely preventive reasons then, as we have just seen, this
is to treat the person merely as a means. But if this is the case, then it seems that the
distinction between end and means or between having dignity and price cannot
deliver the kind of answer we are looking for. As mentioned, the suggestion is that
there are certain tough punishments which, despite the fact that they may be
deserved for a certain crime, should nevertheless not be inflicted because they
violate a person’s dignity. In other words, if the distinction between price and
dignity is interpreted strictly as a distinction between what may and what may never
be the subject of trade-offs, then this distinction is apparently incapable of drawing
the line between acceptable and unacceptable punishments in respects which go
34
beyond the question of desert. Something else or further must be understood by
dignity if the concept is to present us with an upper limit of punishment.
In fact, especially in his later writings, Kant makes clear that treating a
person as an end in himself requires more than simply refraining from acts that
would exploit the person as a mere means. For instance, in the Metaphysical
Elements of Virtue he indicates that indifference to a person is also forbidden, and
that positive assistance may be required. However, what this further aspect of the
concept of dignity amounts to is not made precise. Probably the best way to
approach it is by considering some modern Kantian-inspired interpretations of the
concept.
One of the few modern retributivists who have taken up this question
specifically in relation to punishment is Murphy. In the above-mentioned essay on
35
“Cruel and Unusual Punishments” , Murphy considers whether there are any
punishments which should be opposed in principle. As an extract from a Kantian
account of retribution, his conclusion is that a punishment should be banned in
principle if “(1) it represents a direct assault of the dignity of persons or (2) it is
36
radically disproportional to the seriousness ... of the conduct criminalized” . With
regard to (2), Murphy explains that disproportionate punishment is not addressed to
that for which a perpetrator is responsible and blameworthy, which means that it
thereby fails to show the proper respect for the person as an autonomous being.
Murphy’s view is here in agreement with what is often regarded as part of the
37
interpretation of cruel and unusual punishment. However, in the present context
(2) is not relevant. What Murphy has in mind is apparently not merely ordinal
proportionality but rather the view that a less serious crime should not be very
harshly punished. Thirty years imprisonment for a minor theft would be wrong in
principle (even if ordinal proportionality is observed). However, the point of
seeking to determine anchor points exactly is to find out what should, in the first
place, count as the proportionate punishment for different crimes. Thus, merely to
contend that punishment should not be disproportionate obviously does not help us
in resolving that question.
The part of the conclusion which is here of interest is (1), that a
punishment should never violate someone’s dignity. As an example of a punishment
which is morally prohibited by this principle, Murphy mentions “[s]ending painful
voltage through a man’s testicles to which electrodes have been attached, or boiling
him in oil, or eviscerating him, or gouging out his eyes ..38. In a comment on these
punishments, Murphy adds that he cannot “prove that it is wrong to treat people in
this way; for the wrongness of doing this is more obvious than any premises which
could be given to justify its being wrong. Anyone who did not see this could not be
made to understand anything else about morality”. If by “wrong” Murphy means not
only prima facie wrong but wrong in principle - which is what we are here
considering - then I believe this claim is too hasty. To hold that other retributivists
who actually accept torture (in fact, though it does not amount to exactly the same it
is worth recalling that Kant accepted castration as a punishment) or theorists in a
consequentialist camp who are willing to allow such acts if this is the only way to
avoid something even more terrible, are all morally blind, is in my view too easy.39
That there might well be something to it, if one cannot see that such treatments are
terrible for those who undergo them, is probably correct, but surely this is not
tantamount to claiming that they are wrong as a matter of principle. Be that as it
may, Murphy’s claim may nevertheless bring another question to mind, namely,
whether a strict definition of what dignity amounts to or implies is at all required in
order to draw the line between unacceptable and acceptable forms of punishment.
To take a standard example from the philosophy of language, it is obvious that we
are able to point at various instances of water when we face it, even if we do not
have any idea of how water could in any strict sense be defined. Similarly, a person
might very well know when another is lying to him even he has never
considered what definitionally characterizes a lie. Could we hope for the same when
considering dignity-violating punishments? The answer must surely be in the
negative. As already indicated, there are major disagreements as to which
punishments should be proscribed by reference to dignity. Some theorists, such as
Kant, believe that the death penalty does not fall into this category - he regarded it
as the only proper punishment for murder - while others hold that it should on this
ground be banned in principle. Similar disagreements exist with regard to torture.
Thus, in the same way as clarification would be required in cases where
disagreement or doubt exist as to whether something is really water or another

transparent sort of fluid or whether something should count as a lie, it seems that
clarification must also be required with regard to such an intricate matter as what
dignity implies. Murphy’s more intuitive approach to the question of dignity
infringements is not satisfactory.
It seems to me that a least two different clarifications of what dignity
implies can be extracted from the discussions. Besides the appeal to punishments
which are obviously wrong, Murphy mentions - in relation to the claim of a ban on
punishments which are inconsistent with dignity - that a just theory of punishment
must keep the special status of person at the centre of attention and that punishments
of the sort that should be banned treat the perpetrator as an animal rather than as a
human being. Corresponding claims are found in several places in Kant who makes
the basis of respect the “humanity” in each person, which leads him to regard
punishments such as drawing and quartering as forbidden. That perpetrators should
be treated as persons or human beings is also emphasized by several contemporary
thinkers. Though such formulations are in fact frequent it is not obvious what is
exactly meant. Unfortunately, Murphy does little to make this clearer. However, a
step further towards a clarification is taken in a recent article by Kleinig. According
to him, the “capacity to frame for oneself the choices one makes, the paths one
40
treads, and the goals one pursues is the foundation for human dignity” . Thus, as a
first definition we can say that dignity implies that such capacities should be
respected; or, oppositely put and with the focus on punishment, that a punishment
violates dignity if it deprives the punished of, or overwhelms, his capacity to
perform these kinds of basic operations.
However, though this rough definition captures what several of those who
have discussed the moral importance of dignity have in mind, there is also another
notion of dignity which can be extracted from the discussions. For instance, this is
clearly indicated in Spiegelberg’s illuminating analysis of the concept. In his view,
one of the best approaches to an exploration of what dignity means is to start from
concrete experiences of “indignities”. As examples, he mentions: being tortured and
forced to make confession; being packed into overcrowded prison cells; having to
stand in line for hours when there are other means for securing fair treatment. What
he thinks is revolting in these situations is not only the physical happenings but “the
attempt to break down the personalities of the victims and deprive them not only of
41
the respect of others but of self-respect” . Likewise, Gerstein claims, in his
discussion of cruel and unusual punishment, that treatment which undermines “a
42
man’s self-respect” cannot serve as a just punishment. Drawing from this second
account, we can therefore say that dignity concerns the self-respect of an individual,
which means that punishment which undermines a person’s self-respect belongs to
the category that should be banned in principle.
The two accounts of what dignity implies are clearly related, but not
identical. It is not impossible, though perhaps not very likely, that a person might
maintain self-respect even if he is deprived of some of his capacity to form and
pursue his own goals. And it is possible that a person’s self-respect is threatened by
a certain treatment even if it does respect his capacity to perform these operations.
138 JESPER
THE ETHICS RYBERG
However, it seems that there might well be a causal relation between the two
accounts, in the sense that someone who is not in the specified sense respected as a
person might well lose his self-respect. Thus, though perhaps interrelated, the two
accounts are not inseparable. The important thing here is, as the former quotations
indicate, that the two accounts seem to capture what is often regarded as important
with regard to the concept of human dignity. The crucial question, therefore, is
whether we have now reached the sort of guidance which is required in order to
identify a maximum punishment. I do not believe this to be the case. In fact there
are several reasons to be sceptical with regard to how far the two accounts, or any
related ones, can bring us towards a satisfactory solution of the problem.
A first important thing to note is that, though both of the two outlined
approaches are much more clear than, for instance, such a frequent claim as “one
should always treat other people as persons or human beings”, they are nevertheless
far from providing precise prescriptions of what respect of human dignity implies. It
is here worth recalling what kind of answer it is that the suggested solution to the
anchor problem presupposes, namely, that we are able to draw a clear line between
43
acceptable and unacceptable degrees of punishment on the severity scale. That is,
either between, for instance, torture followed by death and the death penalty; the
death penalty and torture and imprisonment; or between prison terms of different
duration. That the two accounts should be able to provide this kind of guidance is
far from clear.
Consider the first approach. What does it imply to hold that one should
respect each individual’s capacity to frame for himself the choices he makes and the
goals he pursues? That one should in no way affect the way a person frames for
himself choices and goals? And what kind of choices or goals are relevant? Murphy
objects to sending voltage through a man’s testicles, or boiling him in oil, on the
grounds that the person could not be expected to understand it while went on, have a
view about it, or conduct any other characteristically human activities during the
process. But what if the person were able to have a view about it before and after
the punishment, why would that not be sufficient (will it violate the dignity of a
person to force him to sleep for a number of hours?). And is the important thing not
to violate a person’s capacity to frame choices and goals, or should one also seek
not to hinder the person’s possibility of pursuing such goals? As an example of
deprivation of dignity, Kleinig mentions a prisoner who has to forgo the expression
of his political opinions or religious observances. But the fact that one is forced not
to express opinions or expose observances does not necessarily infringe a person’s
capacity to form views or opinions. On the other hand, if the important thing is not
to obstruct the possibility of pursuing wishes or goals, then how exactly does one
avoid the conclusion that punishment as such should be banned? After all,
punishments almost always conflict with the goals of a person or the possibility to
pursue them. Moreover, is the important thing that, in so far as there is an existing
person, we should respect his or her relevant capacities, or that one must not
eliminate the presupposition for such capacities, namely, existence? In short, does
44
respecting someone’s dignity exclude the death penalty?
The same sort of ambiguity exists if we turn to the second account. What
does it imply to hold that one should never punish a person in a way that
undermines or deprives him of his self-respect? That one must never affect another
person’s self-respect for the worse or, more radically, that what one should abstain
from is to cause a total break-down of another’s self-respect? Moreover, is it
unacceptable if the impact on someone’s self-respect is only momentary or is it only
punishments, which over a longer period or permanently deprive the person of self-
respect, that should be banned? And once again we might ask whether the important
thing is to avoid that an existing person is deprived of self-respect or whether
eliminating self-respect by ending a person’s life is also unacceptable; that is,
whether this notion of dignity excludes the death penalty?
Thus, it seems to me that both accounts lack the kind of precision which
the drawing of a sharp line between acceptable and unacceptable punishments
presupposes. Furthermore, it seems that if one actually did suggest the kind of
precision which is required in order to draw this line, then this might question the
plausibility of the idea that some punishments should be banned in principle.
Suppose it was made clear exactly what kind of impact on the way a person frames
choices for himself it is that threatens dignity. Or exactly what sort of impact on
self-respect it is that belongs to this category. Is it then still obvious why
punishments which violates dignity in the specified way should be banned? It seems
to me there might well be a tension between, on the one hand, such very precise
specifications of what dignity implies and, on the other, the hope that it remains
intuitively compelling that one particular punishment, but not one which is only
marginally different, should be banned on principle. However, I admit that this
claim is disputable.
A second related point which gives reason to doubt that references to
dignity can help solving the anchor problem is that such references seem more
suitable for excluding certain types of punishment rather than for drawing a
borderline between different degrees of a certain kind of punishment. Suppose, for
instance, that we accept that the death penalty and that torture are ruled out as
unacceptable punishments on the ground of references to dignity. Indeed, this is
what most contemporary theorists suggest. What then of imprisonment? As already
indicated, most proportionalists accept imprisonment - at least if certain minimal
conditions are satisfied - as an acceptable form of punishment. But where should the
upper limit be set? It is hard to see any reason based on dignity for claiming that
there is a relevant limit between 5 or 6 years in prison, or between 30 and 31
45
years. However, if it is correct that such a limit cannot be drawn, then it follows
that the maximum prison term will constitute the maximum acceptable punishment.
But what is the maximal prison term? To contend that this is a life term will surely
not help; at least not if it is understood literally as imprisonment until death. This
would imply that it would be perhaps 20 years for one person and 40 years for
another. Thus, obviously this will not help with regard to setting the maximum
anchor point on the ground of which all other punishments could be determined.
Should the limit then be 50 years or, perhaps, 100? It is hard to think of a non-
arbitrary way in which this can be answered.
140 JESPER
THE ETHICS RYBERG
A third problem facing a solution to the anchor problem based on the two -
or related - accounts of dignity concerns differences in the way punishments affect
different persons. Suppose we accept the two outlined accounts of dignity as
concerning either a person’s capacity to form choices or goals for himself or a
person’s self-respect. In that case, it is far from obvious that a certain punishment
can be said either to violate or not to violate dignity. There might will be great
differences in the way one and the same punishment affects different people. A
punishment might in the relevant way hinder one person’s capacity to form choices
or goals but not another person’s. And when it comes to what will in the relevant
way affect someone’s self-respect, then it seems that this is clearly a matter on
which there may be great differences from one person to another. More generally,
even if we can give a general account of what it means to violate dignity, there
might nevertheless be differences between persons as to when a certain action
constitutes an instance of a dignity violation. That there might be such differences is
not in itself a theoretical problem. If one accepts the idea of dignity as placing side-
constraints on our actions, then it simply implies that there are certain actions which
should not be done to one person but which might be done to another. However,
though this is not in itself problematic, it is a problem when the purpose is to locate
one maximum punishment on the ground of which crime scale and punishment scale
could be anchored. What adherents to the suggestion have in mind is surely not that
the maximum punishment should vary from one person to another. If we imagine
that for one perpetrator A the maximum punishment would be set higher than the
maximum punishment for another perpetrator B and that there should be an interval
matching between the two scales, then this would imply that for every possible
crime committed by A he would be punished more severely than B. This would not
only violate the parity condition but might also contradict what in the previous
chapter was referred to as the principle of equal impact. In short, this proposal is
inconsistent with ordinal proportionality. How, then, can this implication be
avoided?
As we have seen in several other situations, one might try to avoid the
problem by applying some sort of standardization. That is, one might determine the
maximum punishment by considering what would typically be the level at which
people’s dignity is violated. However, how exactly such a standardization should be
made is not clear. Moreover, one might wonder whether this would be at all
46
consistent with the idea that violations of dignity are wrong in principle. Another
possible reply would be to reject violations of dignity, in the above-mentioned way,
being an individual matter. One might claim that if a certain punishment violates
dignity, then it does so for all human beings. Obviously, if this were the case it
would solve the problem. However, this answer takes us back to the point from
where we started, namely, of clarifying what we mean when we talk about human
dignity. The two accounts outlined above, which I believe are some of the most
precise suggestions that have been presented, certainly do not seem to exclude
47
interpersonal variations in impact. Moreover, in my view, it is not easy to think of
a plausible notion of dignity which does exclude such variations.
To sum up, I believe that the two suggested accounts of what dignity
implies face several problems. They are imprecise, which is a problem when the
task is to draw a precise borderline. They do not seem well suited to drawing a line
between different degrees of a certain form of punishment, which means that - even
though references to dignity might assist when it comes to the exclusion of certain
types of punishment, which is what writers on cruel and unusual punishment have
typically been concerned with - it is far from clear that genuine guidance is
provided when it comes to the identification of a maximum punishment limit.
Finally, there might be individual differences with regard to when someone’s
dignity is violated, which constitutes a further problem when it comes to the
identification of a general upper limit of punishment. Some theorists have
emphasized that it is not only the dignity of the person who is punished that matters
but also the dignity of the punisher. For instance, Scheid says that “one human
being should never impose on another suffering that is beyond certain limits: part of
48
my honor and dignity as a human being is that I will not torture” no matter what a
criminal has done. However, adding this extra perspective to the considerations does
not contribute anything with regard to the problems that have been sketched. On the
contrary, the problem remains intact independently of whom is considered the
possible victim of dignity infringements.
In sum, I believe that even if the arguments I have presented are not knock-
down arguments which manage to establish the impossibility of basing a view on
maximum punishment on considerations of dignity, they nevertheless provide
reasons for being sceptical with regard to the whole idea of anchoring scales on the
ground of dignity.
There is, though, still a way in which the problems I have pointed at could
be avoided. The procedure in the previous discussion has been, firstly, to consider in
general what human dignity implies and then, secondly, to investigate how far this
brings us with regard to identifying the maximum punishment. However, there is an
approach which in a very simple way succeeds in getting around this procedure,
namely, to claim that the sort of punishments which we regard as really terrible are
those which violate dignity, or at least those that should be banned on principle. A
view of this sort has been defended by Davis. Davis does not speak about human
dignity but about “inhumane punishment” as something which can never be
justified. The important thing according to him, however, is that a penalty is
inhumane if its use on anyone “would normally shock us”. Thus, as Davis admits
49
“[e]verything then depends on what shocks us” ; where “us” refers to a social
consensus. This account, therefore, is relativist in the sense that what should count
as inhumane punishment might well change from one society to another. Does this
simple proposal provide the answer which proportionalists have sought? One might
ask whether the mere fact that a punishment shocks us is a satisfactory reason for
holding that it can never be justified. But no matter what the answer is to this
question, there is another objection which shows that even this easy way of
identifying a maximum punishment will still not bring us the final step towards a
plausible anchoring of crimes and punishments. In fact, this final objection holds
independently of whether one adopts a view like the one suggested by Davis, or
142 JESPER
THE ETHICS RYBERG
whether one believes in more traditional accounts of dignity as providing a base for
50
the ban of certain punishments.
Suppose that we have in some way or another succeeded in identifying the
maximum punishment which it could be justified to impose on a criminal. Now,
according to the suggested solution of the anchor problem, we should then peg the
most severe punishment to the most serious crime. But why? Why is it that a
criminal’s most serious wrongdoing should be pegged to the most severe
punishment? Why not peg it to the punishment just below the most severe on the
scale? Certainly it does not follow from the mere fact, that a certain punishment is
the most severe it could ever be justified to impose, that the most serious crime
should be responded to with exactly this punishment. The demand of ordinal
proportionality could obviously just as well be maintained if the crime were coupled
to a less severe punishment. The question can, of course, be repeated with regard to
the lower anchor point. As mentioned, the suggestion was that the least offensive
crime should be pegged to the least possible punishment. But why not peg it to one
a little more severe? Obviously, whether we do the one or the other will, if we are
considering an interval matching, affect the level of all the punishments assigned to
crimes on the crime scale. I believe that this objection shows that, even if we
disregard the problems which we have just considered concerning the possibility of
identifying a maximum punishment, the suggested solution to the anchor problem is
still not sufficient if what one is aiming for is a principled way of relating
51
punishments to crimes.
3. DESERT, PREVENTION, AND PARSIMONY
A proportionalist theory of punishment which does not accomplish more than

determining certain upper and lower limits of punishment which should, in
principle, never be transgressed, is not of much assistance with regard to how
punishments should be distributed. As the final objection in the previous section
showed, restrictions concerning maximum and minimum punishments leave it an
open question as to how different crimes should actually be punished. For instance,
a person who has committed a serious crime might receive fifty years imprisonment
or alternatively a fine, as long as the punishment limits and the ordinal
proportionality constraint are observed. Similarly, a theory which required nothing
further would also be consistent both with responding to minor crimes with several
years of imprisonment and with a fine. Such a leeway in the distributions of
punishment is hard to accept. It would leave us without a more precise guidance as
to how particular perpetrators should actually be punished and, more importantly, it
seems implausible to hold that we should be morally indifferent in the choice
between these alternatives. This indicates, at best, that something more is required
of a satisfactory solution to the anchor problem. Strangely, proportionalists have at
this point been remarkably silent. There is, though, one noteworthy exception. Von
Hirsch has suggested an approach which succeeds in answering the outlined
indeterminateness problem and which therefore deserves a closer inspection. In
52
short, the suggestion combines considerations of desert, prevention and parsimony.
As the first part of the suggestion, von Hirsch holds that there are
constraints against inflating punishment pro rata above a certain upper level.
Draconian punishments which prescribe that even the least serious crimes should be
responded to with several years of imprisonment will - even if ordinal
proportionality requirements are scrupulously observed - be morally unacceptable.
A crime which does not cause much harm is not very reprehensible. To punish a
person who has committed such a crime, in a way which has significant impact on
his or her life, is to trivialize that person’s interests: it is to convey the message that
the interests of the punished are of little importance. Desert sets an upper bound of
cardinal proportionality, that is, the limits for how severely different crimes may be
responded to. However, von Hirsch points out that he does not believe it to be a
sharp limit but rather a shaded area which grows progressively darker as the
punishments become more drastic. So much for the upper constraints on cardinal
proportionality.
Does desert also dictate lower limits to the extent to which a punishment
scale may be deflated? As we have already seen in the outline of von Hirsch’s
expressionism in chapter 1, he does not believe this to be the case. The important
thing is that perpetrators are censured for their misdeeds but, since blame and
censure can be expressed in other ways than through hard treatment, desert does not
imply a lower limit. The reason that von Hirsch, nevertheless, maintains that there is
a lower limit of punishment, concerns his view on prevention. As long as there is
need for prevention, this will be what determines the constraint on the deflation of
punishment levels. However, these outer constraints leave much leeway for the
actual anchoring of the scales. In Past and Future Crimes von Hirsch suggested that
respects to prison capacity might help in fixing the anchor points; a suggestion he
has in his more recent writings rightly redrawn. To determine the punishment level
partly on the ground of the capacity of the penal system is to turn things upside
down. The justificatory direction should be the other way round: prison capacity
should be determined (at least partly) on the ground of how perpetrators ought to be
53
punished and not vice versa. But what should then determine the anchoring?
The key to the answer lies in von Hirsch’s interpretation of the concept of
prevention. The important thing is that what he has in mind is not a traditional
“optimizing” view of prevention. Thus, the thought is not that the ordered crime
scale should be connected to the punishment scale in such a way that an optimal
crime prevention is gained. The reason is that an optimizing conception of
prevention might - given certain empirical circumstances - lead to a significant
increase in the punishment level. And in that case, a person who receives a
punishment might - according to von Hirsch - rightly object that his “vital interests
are being sacrificed to the social good” and that he is being used as “a tool for
54
promoting the most socially efficient system of prevention” . This von Hirsch
regards as unacceptable. On his revised account, prevention should be perceived in
the light of the fact that people are liable to temptations and that the infliction of
hard treatment provide them with a reason for compliance. But this should only be
an extra pragmatic reason, that is, a reason which is supplementary to the fact that
144 JESPER
THE ETHICS RYBERG
the person is being censured for his wrongdoing. In short, according to von Hirsch,
prevention is perceived as a supplementary prudential incentive for compliance.
Now, with this concept of prevention settled, we are almost home. What
von Hirsch suggests is that one should adopt a decremental strategy. A punishment
system should start by ordering the crime scale, from the least to the most serious,
and by indicating a punishment level which as far as possible corresponds to the
current level in the society. Punishments at this level will express the proper
censure. However, censure expressed through punishment is a convention. The final
step which then induces the decremental procedure is to adopt the principle of
parsimony, according to which state-inflicted suffering should be kept to the
minimum necessary to achieve the purpose of the intervention. Given this principle,
what should be done is to change the censure-expressing convention by a pro rata
reduction of the penalties. But this still leaves the question of where the progressive
diminution of the punishment level should stop. The answer obviously is: at the
point at where the floor suggested by the preventive considerations has been
reached. Thus, in sum, what von Hirsch’s somewhat complex answer to the question
of cardinal proportionality amounts to is that desert sets upper shaded limits to the
punishment level; but that the punishment level should ideally be determined by the
lowest acceptable level, that is, by the minimal point at which the system still
succeeds in providing a prudential incentive for compliance.
In comparison to the theory discussed in the previous section, this
suggestion certainly has its attractions. It does not appeal to concepts like dignity or
humane treatment and, more importantly, by adopting the principle of parsimony,
which indeed is a plausible moral principle, it manages to get around the problem of
only setting outer limits. Thus, it apparently succeeds in prescribing directions for a
genuine anchoring of the scales. However, despite these apparent advantages, it
nevertheless strikes me that von Hirsch’s theory on closer scrutiny faces several
challenges. Challenges which basically leave the anchor problem unsolved.
Consider first the question of the upper limits of the punishment level. As
mentioned, von Hirsch believes that draconian punishment trivializes the punished
person’s vital interests: it conveys the messages that these interests are of not much
importance. What does this mean? It sounds as if von Hirsch is claiming that a
punishment trivializes interests if it is too tough. But when is that? The example that
is given is imprisonment for the least serious crimes. But what is an interest
trivializing punishment for the more serious crimes? Does the death penalty or 50
years in prison for a serious crime
55
exceed the shaded upper bound (does it fall in the
dark or the less dark zone)? This is not clear. It is simply not so obvious when a
punishment constitutes an instance of trivialization.
Moreover, it strikes me that the reference to trivializations of the interests
of the punished comes very close to presupposing a view on cardinal proportionality
rather than establishing it. Von Hirsch’s point is not that certain forms of treatment
per se trivialize interests, for instance, that a term of two years in prison in itself
constitutes a trivialization. Rather, this depends on the seriousness of the crime one
has committed. But in that case, one is very close to presupposing a view on
whether a punishment is proportionate or disproportionate as a part of what makes it
a trivialization of interests. In order to avoid a petitio principii, the defining
characteristics of when something is an interest trivialization must obviously be

independent of references to (dis)proportionality. But can such characteristics be
given? Is it not precisely the fact that a certain punishment is disproportionately
severe, relative to the crime, that turns it into a trivialization of interests? Suppose
one was asked whether imprisonment for a minor theft would trivialize the
perpetrator’s interests, would not the reason for an affirmative answer be that this
was a too tough response for this sort of crime? If a punishment conveys the
message that a perpetrator’s interests are not of much importance, is it not precisely
because the punishment is disproportionately harsh? This seems to me to be the
most reasonable answer (if one accepts at all a retrospectively oriented view on
punishment). But obviously it is not a satisfactory answer if we are searching for a
justification of which punishments are proportionate. What von Hirsch needs to do
is to provide an independent ground for determining when interests are being
trivialized; however, since no such ground is presented we are, as far as I can see,
still left with the core problem that the suggestion does not provide any justification
for when the punishment level should be regarded as unacceptably high.
As already indicated, it is, however, not the upper bound of cardinal
proportionality which plays the most important role in von Hirsch’s theory. The
crucial part of the suggestion is the lower limit which - in a world like the present
one in which there is a need for the above mentioned sort of prevention - ideally
indicates the point at which the scales should be anchored. When do we, through the
application of a decremental strategy, reach this floor? The answer, in my view, is
notoriously unclear. Suppose, for reasons of ease in exposition, that the crime scale
contains only the crimes from C1 to C5 ordered from the least to the most serious.
Suppose further, for purely illustrative reasons, that the prudential disincentive to
committing C1 is 3 units of hard treatment. Should the decremental procedure,
starting from the current punishment level, then proceed until C1 is punished with 3
units of hard treatment? Apparently, it is something like this von Hirsch has in
mind; however, the question is how the disincentive to committing the other crimes
C2 to C5 should affect the anchoring. What if, by pegging C1 to a punishment which
inflicts 3 units of hard treatment, not one of the other crimes has reached the level of
what would count as the minimal workable disincentive to committing each of these
crimes? Suppose that the disincentive to committing each of the crimes C1, C2, C3,
C4 and C5 is respectively 3, 6, 7, 14, and 15 units of hard treatment, and that the
punishment level when we start the decremental procedure is, for each crime, above
this level. When should we then stop deflating the scale?
At this point there is a difference depending on what kind of matching of
56
the scales one is defending. Suppose again, as most do - including von Hirsch -
that one adheres to an interval matching. That is, that the differences in seriousness
between different crimes should be reflected in the differences in severity between
punishments. In that case, the differences in seriousness between C1 to C5 need not
at all coincide with the differences between the disincentives to committing each of
the crimes. What should one then do? Should the deflation stop at the point where
C1 is met with a punishment imposing 3 units of hard treatment? Why not the point
at which one of the other crimes reaches the related disincentive? This is not clear.
146 JESPER
THE ETHICS RYBERG
Suppose, therefore, alternatively that one adheres to nothing stronger than

an ordinal matching between the crime and punishment scales. In that case, if C1 is
pegged to a 3 units punishment then each of the other crimes could as well be
pegged to punishment inflicting exactly the hard treatment which each of the
disincentives require. Ordinality would then be observed. And the problem that the
differences between the incentives do not correspond to the differences in
seriousness between the crimes would be avoided. However, this is only because in
the example it is assumed that a more serious crime requires a larger disincentive.
However, there is no reason to believe that this is the case. In fact, there are reasons
to believe that it is not always so. As we have seen, von Hirsch holds that
disincentives are required because people are liable to temptations. But there is no
reason to think that a crime which is more serious than another is also one which is
more tempting. As we have already discussed in some of the previous chapters, tax
evasion might well - to most people - be seen as more tempting than a violent crime.
But that certainly leaves an open question as to how the scales should be anchored if
the disincentives related to the crimes C1, C2, C3, C4, and C5 are, for instance, 3, 6,
2, 13 and 8. And this will be a problem even if the matching is only ordinal. Why
then stop the decremental procedure at the point at which C1 is responded to with a
punishment inflicting 3 units of hard treatment, and not at the point where C3 is met
with a 2 units punishment, or C5 with one inflicting 8 units? It is hard to think of a
non-arbitrary answer. Perhaps a suggestion would be that the scales should be
anchored in such way that all the crimes come closest to their respective
disincentive floors. But how exactly this should be estimated and whether it is at all
57
practically possible is far from obvious.
By presenting these examples concerning crimes and disincentives the
point is obviously not to suggest that one can actually make such very precise
estimates on the size of hard treatments, but rather to illustrate some problems that
seem to exist even if such a degree of precision was in fact achievable. However,
there is a way one could get around the problem. In the examples, it is assumed that
the disincentives are related to each particular crime. This means that, when a
perpetrator is punished, he or she receives a hard treatment which provides him or
her with a supplementary prudential reason for not repeating that particular crime
for which he or she is punished. However, at this point it might be responded that
the incentives should not be determined in relation to a particular crime but rather to
compliance in general. That is, the criminal should simply receive an incentive not
to break the law. However, this proposal also faces problems.
The first question is how the size of this disincentive should be determined.
If the disincentives are different for different crimes, then how should one then
determine an incentive for compliance in general? Should it be determined on the
58
ground of the most tempting crimes? If not, will the hard treatment not simply fail
to function as a disincentive to committing these crimes? The second question
concerns the anchoring itself. Suppose that, for some reason or another, this general
disincentive is provided by a 3 unit hard-treatment punishment. Should C1 then be
pegged to a 3 units punishment? If that is so, then what about the other crimes? If
one defends an interval matching between the scales then, as we have seen in section
(1), it is not sufficient with only one anchor point. Another anchor point will be
required as well in order to determine how the different crimes should actually be
punished, which means that the anchor problem remains unsolved. And if the
matching is only an ordinal one, should one then press all punishments as close to
the 3 units limit as is possible within the constraints of ordinal proportionality? In
such a case one would probably end up with a matching at which there is only a
marginal difference in punishments of the different crimes, which would mean that
even the least and the most serious crime on the scale would be punished in almost
the same way. It is certainly not obvious that adherents to an ordinal matching
would be willing to accept this conclusion. As we have seen, von Hirsch himself
defends some sort of interval matching which, as indicated, leaves the actual
connection between the scales an open question. Thus, to hold that the incentive is
an incentive to compliance in general does not provide an easy way out of the
problems. In my view, the claim that the anchoring should in the end be determined
by deflating the punishment level to the point set by von Hirsch’s revised
conception of prevention does not, therefore, succeed in providing a theoretically
clear answer to the anchor problem.
There is one further point worth making which gives further support to this
conclusion. Von Hirsch claims that - in contrast to an anchoring based on the
optimizing view of prevention - a responsible person would accept a punishment
59
based on his account of prevention. It is not clear to me what this means. But even
if we ignore this detail, another more important problem remains. The prudential
disincentive is supposed to provide the punished with a supplementary reason for
compliance. This is needed because people are fallible, that is, liable to temptations.
However, the temptation to commit crimes might well vary from one person to
another. It is an indisputable fact that some persons are much more tempted to cheat
in tax or commit other sorts of crimes, than are others. But in that case, how should
the size of the disincentive be determined? Von Hirsch cannot respond that it should
be individually determined. This would imply that the floor set by prevention would
vary from one perpetrator to the other, which would entail that the whole matching
of crimes and punishments would be different for different persons, which would
violate the ordinal proportionality constraint. Two persons might commit the same
crime but be punished differently. But if this is not a viable option, then how should
one respond to such interpersonal differences?
The only possible answer once again seems to be to adopt some kind of
standardization, which means that the size of a disincentive is determined as the
amount of hard treatment that would typically provide a person with this
supplementary prudential reason for compliance. However, as earlier pointed out in
relation to standardizations, it obviously has to be shown that this can in fact be
60
justified; more precisely, that it is not inconsistent with other parts of the theory.
Moreover, if that is how von Hirsch would respond, then it seems to me that he
might well be hit by the same argument which led him to the rejection of optimizing
prevention in the first place. Suppose that a person is less tempted to commit crimes
than the typical person. Could this person then not argue that it is unacceptable that
he should be punished harder (than he would be if the anchoring was based on what
148 JESPER
THE ETHICS RYBERG
would constitute a sufficient prudential disincentive for him) just because other
people are more tempted than he is and therefore require larger incentives to
compliance? The problem, as mentioned, is that this kind of objection is very close
to von Hirsch’s argument against optimizing prevention. In short, there seems to be
a tension between, on the one hand, providing an answer as to how the disincentive
should be determined when different persons are very differently tempted to commit
crimes and, on the other, maintaining invulnerability to the arguments which led to
the rejection of optimizing prevention. However, even if this tension could be
resolved, we would still be left with the more substantial problems outlined above.
Therefore, the conclusion, in my view, is that the suggestion does not succeed in
providing a plausible solution to the anchor problem.
4. CONCLUSION
To hold that ordinal proportionality is all that should be required from a theory
concerning the distribution of punishments is - as we have seen in the beginning of
this chapter - not plausible. It would not provide any guidance as to how different
crimes should be punished; it would be consistent with arbitrary jumps in the way
different crimes are punished; and it would allow for extremely harsh punishment of
perpetrators who have committed only minor crimes. All in all, implications which
it indeed are morally hard to swallow. Once this is realized it is clear that
proportionalists are faced with a genuine challenge with regard to how the scales of
punishments and crimes should be connected. To contend that each crime on the
scale could be directly linked to an appropriate punishment by adopting a literalistic
interpretation of lex talionis or the more modest principle of harm-for-harm
equivalence was not acceptable. And neither is this what typical recent
proportionalists would suggest. The question therefore was what indirect methods
there might be for connecting the scales. The first suggestion to be considered - to
peg the least serious crime to the minimum punishment and the most serious crime
to a maximum punishment determined on the ground of dignity - faced several
problems. As we have seen, it was unclear how the anchor points, especially the
upper points, should be determined and, even if this could in fact be determined, it
was left an open question why the scales should be pegged at these points rather
than at others within the limits. The second suggestion, presented by von Hirsch, did
at first glance avoid the problem of arbitrariness in the anchoring by incorporating
the principle of parsimony. However, as I have argued, this suggestion still left the
anchoring of the scales unsettled.
Though I do therefore not believe that proportionalists have succeeded in
answering the anchor problem, it would of course be premature to conclude that the
problem cannot be solved in ways different from the ones considered. One
possibility would be to hold that the problems facing the different solutions are the
result of an over-rationalization of the distribution of punishments. That is, a result
of the fact that one requires reasons sustaining the anchoring. If one simply
connected the two scales in a way that one would find appropriate, without requiring
any further rationales, then problems of the sort facing the suggestions which have
been considered, would be avoided. Perhaps, it is something like this that Primoratz
has in mind when he says that, when it comes to the construction and coordination
of the two scales “[n]ot very much is needed in order to carry it out with reasonable
61
success: it is enough to turn to one’s moral conviction and apply common sense” .
However, this approach strikes me as unacceptable. Firstly, there seems to be some
arbitrariness in the role reasons play if one requires, as proportionalists usually do, a
justification for the ordinal proportionality claim, that a more serious crime should
be punished more severely, but not for cardinal proportionality considerations on
how perpetrators should be punished. Secondly, and more importantly, it is very
hard to accept that one should not require any reference to reasons when it comes to
such vital questions as whether a perpetrator should be executed, tortured,
imprisoned or fined for a misdeed. Thus, proportionalist theories like those
considered in the previous sections are at the outset on the right track in the search
for a principled ground for the imposition of punishments.
Another possible approach to the anchor problem, which does not consist
in the rejection of a justification, would be to hold that the scales should be linked
on the ground of traditional preventive reasons. On this proposal the ordered crime
scale should be connected to the punishment scale in the way that will lead to an
optimal crime prevention. However, for many proportionalists this would certainly
be considered a too radical step in the direction of a utilitarian view on punishment.
This might, as proportionalists have typically complained, at least in principle imply
62
that crimes should be very toughly punished. From at proportionalist perspective it
is not an attractive option to withdraw respects to justice from cardinal
proportionality considerations.
The conclusion to be drawn on the background of the considerations in this
section is that proportionalists have not succeeded in answering the anchor problem.
Moreover, it is not easy to think of a way in which the problem could be solved
within a proportionalist framework. Several proportionalist have from time to time
emphasized that one should not expect mathematical precision when it comes to the
distribution of punishments. However, the problems which I have pointed out as
facing the suggested solutions do not consist merely in the lack of this kind of
precision. Rather are they theoretical problems, in the sense, that they constitute
problems that would exist even if mathematical precision was in fact fully
obtainable. Moreover, the objections cannot be held merely to stick at trifles. If a
theory cannot tells us whether a crime, say a burglary, should be punished with a
fine or with several years of imprisonment, then the theory is not - at least, so I have
suggested - morally satisfactory. And it is this sort of leeway that the considered
theories allow for.
150 JESPER
THE ETHICS RYBERG
NOTES
1See, for instance, the discussion in D. J. Rotman, “More of the Same: American Criminal Justice
Policies in the 1990s”, in T. G. Blomberg and S. Cohen (eds.), Punishment and Social Control, Aldine De
Gruyter, New York, 1995. Or B. Hudson, Justice Through Punishment, London, 1987, chap. 3. But see
also A. von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993, chap. 10.
2R. G. Singer, Just Deserts, Ballinger Publishing Company, United States of America, 1979, p. 44.
3J. G. Murphy, Retribution, Justice, and Therapy, Dordrecht, 1979, p. 230.
4See A. von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993, chap. 10.
5C. W. K. Mundle, “Punishment and Desert”, in H. B. Acton, The Philosophy of Punishment, St Martin’s
Press, Great Britain, p. 73. Admittedly, Mundle’s claim is ambigious. It might also be interpreted as the
view that proportionality only requires (a) and (b), but that a complete theory of punishment distribution
is not provided by the proportionality principle itself. However, Mundle says nothing about how
proportionality should be applied in order to reach a complete distribution theory.
6Mundle is, of course, aware of this but does not regard it as a problem.
7See, for instance, Kleinig’s convincing examples in Punishment and Desert, Martinus Nijhoff, The
Hague, 1973, p. 118.
8For historical remarks on lex talionis, see, for instance, I. Primoratz, “On some arguments against the
retributive theory of punishment”, Rivista internazionale di filosofia del diritto”, 1979, vol. 56.
9See J. Kleinig, Punishment and Desert, Martinus Nijhoff, The Hague, 1973, p. 121f.
10F. Hegel, Philosophy of Right, translated by T. M. Knox, Oxford University Press, United States of
America, 1967, p. 72.
12That is, if proportionality is understood in the “impact” interpretation considered in the previous
chapter.
13F. Hegel, Philosophy of Right, translated by T. M. Knox, Oxford University Press, United States of
America, 1967, p. 72.
14J. H. Reiman, “Justice, Civilization, and the Death Penalty: Answering van den Haag”, in J. Simmons
et al. (eds.), Punishment, Princeton University Press, United States of America, 1995, p. 280f. See also E.
van den Haag, “The Lex Talionis Before and After Criminal Law”, Criminal Justice Ethics, vol. 11 (1),
1992.
15Ibid. p. 281.
16Ibid. p. 278f.
17W. G. Maclagan, “Punishment and Retribution”, Philosophy, vol. 14, 1939, p. 290.
19C. L. Ten, Crime , Guilt, and Punishment, Clarendon Press, Oxford, 1987, p. 158. See also A. von
Hirsch, “Proportionality in the Philosophy of Punishment”, in M. Tonry (ed.), Crime and Justice, vol. 16,
1992; or D. E. Scheid, “Constructing a Theory of Punishment, Desert, and the Distribution of
Punishments”, The Canadian Journal of Law and Jurisprudence, vol. 10, no. 2, 1997.
20See J. Kleinig, Punishment and Desert, Martinus Nijhoff, The Hague, 1973, p, 118.
21See also Scheid discussion in “Constructing a Theory of Punishment, Desert, and the Distribution of
Punishments”, The Cancadian Journal of Law and Jurisprudence, vol. 10, no.2, 1997, section IX and X.
22Davis has defended a certain version of an ordinal matching. As we have seen in chapter 2, he suggests his
seven-step method as an easy way of reaching a fitting penalty. I have already argued that there are
reasons to be sceptical with regard to the claim that there is a structural isomorphism between the seven-
step method and the hypothetical auction model. But even if we ignore this problem and focus only on the
seven-step procedure, there are questions which are left unanswered. As earlier outlined, Davis’ idea is
that all crimes should be ranked within types, which should then be combined into a scale. The same
should be done with regard to the penalties. Finally, the greatest penalty should be connected with the
greatest crime, and the least penalty with the least crime, and so on with regard to all other crimes and
punishments. However, apparently Davis does not even defend full ordinality. What he claims is that the final
scale may resemble a New York subway map. And it seems as if Davis does not hold that a crime of one type
on the scale is comparable in seriousness to a crime of another type (except when the lines cross). If this is
what Davis means - that is, if the scaling is only partially ordinal - then it leads into several problems. For
instance, as Dolinko has argued, it becomes unclear what is meant by the greatest penalty being assigned to
the greatest crime; and even though Davis claims that, when the scales are linked, the lines connecting them
should never cross, this would nevertheless open up a number of very different ways of linking the scales.
What this indicates is that partial ordinality will require an even more supplementing theory in order to
provide a plausible matching of the scales. See M. Davis, To Make the Punishment Fit the Crime, Westview
Press, United States of America, 1992, p. 77ff; and D. Dolinko, “Mismeasuring ‘Unfair Advantage’: A
Response to Michael Davis”, Law and Philosophy, vol. 13, 1994, p. 516ff.
23D. E. Scheid, “Constructing a Theory of Punishment, Desert, and the Distribution of Punishments”, The
Canadian Journal of Law and Jurisprudence, vol. 10, no. 2, 1997, p. 494. It should be mentioned that, in
contrast to Kleinig who believes that this procedure gives the full answer to the anchor problem, Scheid
regards it as only part of the solution. The other limit of the punishment scale is set by the “no-crime-control
principle”, according to which a given magnitude of the scale must provide more overall crime control than
any lesser magnitude. See p. 497-8.
24Note that, if one believes that a strict interval scale can be developed then one might use the fact that no
crime merits no punishment as a lower anchor point. However, this option is no longer open if one, as
Kleining, does not advocate a strict interval scale but rather a scale somewhere between and ordinal and an
interval scale (where the intervals themselves are not comparable in ratio terms).
25Ibid. p. 494.
26The appropriate punishment for the most serious crime could, of course, be determined on the ground of
desert if it was held that there is an upper limit to what can ever be deserved and that the most serious crime
should be responded to by exactly this punishment. However, to my knowledge, no one has presented the
arguments that would be required to sustain this position.
27The article can be found in J. G. Murphy, Retribution, Justice and Therapy, Reidel, Dordrecht, 1979. A
shorter version occurs in M. A. Stewart (ed.), Law, Morality and Rights, Reidel, Dordrecht, 1979.
28See R. A. Belliotti, “Gloom and Doom: Executing the Eighth Amendment”, The International Journal of
Applied Philosophy, vol. 3, 1986-7, p. 43. Kleinig quotes Chief Justice Warren’s claim that “the basic concept
underlying the Eighth Amendment is nothing less than the dignity of man”, “The Hardness of Hard
Treatment”, in A. Ashworth and M. Wasik (eds.), Fundamentals of Sentencing Theory, Clarendon Press,
Oxford, 1998, p. 284. For another discussion of interpretations of the Eighth Amendment see, for instance, M.
R. Gibbs, “Eighth Amendment - Narrow Proportionality Requirement Preserves Deference to Legislative
Judgment”, The Journal of Criminal Law and Criminology, vol. 82, no. 4, 1992.
29H. Kuhse, “Is there a Tension Between Autonomy and Dignity?”, in P. Kemp, J. Rendtorff & N. M.
Johansen (eds.), Bioethics and Biolaw (vol. 2), Rhodos, Copenhagen, 2000, p. 63.
30See, for instance, M. J. Meyer, “Dignity, Death and Modern Virtue”, American Philosophical Quarterly,
vol. 32 no.1, 1995. See also his “Dignity, Rights, and Self-Control”, Ethics, vol. 99, 1989. Another analysis
can be found in A. Kolnai, “Dignity”, Philosophy, vol. 51, 1976.
31See T. E. Hill, Dignity and Practical Reason in Kant’s Moral Philosophy, Cornell University Press, United
States of America, 1992, p. 203.
32For an excellent discussion of how this version of the categorical imperative should be interpreted, see,
ibid., chapter 10.
33Quoted from A. von Hirsch, “Proportionality in the Philosophy of Punishment”, in M. Tonry, Crime and
Justice, vol. 16, The University of Chicago Press, United States of America, 1992, p.60.
34Of course, one might say that dignity does set a limit, namely, the limit set by what the perpetrator
deserves. But if there is such a limit for each crime, then there will be a direct way of linking crimes and
punishment, and this is not the suggestion we are here considering. And, obviously, this proposal will still
leave the question of what the deserved punishment for a given crime then is.
152 JESPER
THE ETHICS RYBERG
35In J. G Murphy, Retribution, Justice and Therapy, Reidel, Dordrecht, 1979.

36Ibid. p. 235.
37See, for instance, R. S. Gerstein, “Capital Punishment - “Cruel and Unusual”?: A Retributivist
Response”, Ethics, vol 84, 1974; or R. A. Belliotti, “Gloom and Doom: Executing the Eighth
Amendment”, The International Journal of Applied Philosophy, vol. 3, 1986/7; or M. R. Gibbs, “Eighth
Amendment - Narrow Proportionality Requirement Preserves Deference to Legislative Judgment”, The
Journal of Criminal Law and Criminology, vol. 82, no. 4, 1992.
38J. G. Murphy, Cruel and Unusual Punishments” in Retribution, Justice and Therapy, Reidel, Dordrecht,
1979, p. 233.
39For a proportionalist defending torture as a legitimate punishment, see S. Kershnar, “Objections to the
Systematic Imposition fo Punitive Torture”, International Journal of Applied Philosophy, vol. 13, 1999.
40J. Kleinig, “The Hardness of Hard Treatment”, in A. Ashworth & M. Wasik (eds.), Fundamentals of
Sentencing Theory, Clarendon Press, Oxford, 1998, p. 287.
41H. Spiegelberg, “Human Dignity: A Challenge to Contemporary Philosophy”, in R. Gotesky & E.
Laszlo (eds.), Human Dignity. This Century and the Next, Gordon and Breach, United States of America,
1970, p. 60.
42R. S. Gerstein, “Capital Punishment - ‘Cruel and Unusual’?: A Retributivist Response”, Ethics, vol. 85,
1975, p. 78.
43To suggest that we might at least be left with some sort of grey-zone between acceptable and
unacceptable punishments will not do: this would still leave it an open question what we should regard as
the most severe but still acceptable punishment. That is, we would not have succeeded in identifying the
sought-for upper anchor point.
44See J. Cottingham’s excellent discussion in “Punishment and Respect for Persons”, in M. A. Stewart, Law,
Morality and Rights, Reidel, Dordrecht, 1979.
45I very much agree with von Hirsch and Narayan when they say that violations of a person’s dignity are
a matter of degree. See A von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993, p. 84.
46Standardizations might well imply that the identified maximum punishment would infringe some
people’s dignity, more precisely, those whose “dignity-level” is below average. This problem could be
avoided by setting the punishment limit at a point at which even the person who is most sensitive when it
comes to dignity violations would not have his dignity infringed by the maximum punishment. But which
level is that?
47To contend that a certain punishment simply does affect people’s capacity or their self-respect in
exactly the same way, in my view makes it a mystery as to what we are talking about when we refer to
such a capacity or to self-respect.
48D. E. Scheid, “Constructing a Theory of Punishment, Desert, and the Distribution of Punishments”,
The Canadian Journal of Law & Jurisprudence, vol. 10, no. 2, 1997, p. 498.
49M. Davis, “The Death Penalty, Civilization, and Inhumaneness”, Social Theory and Practice, vol. 16,
1990, p. 250.
50In fact, the objection can be directed against any theory which attempts to accomplish an anchoring
merely through the identification of punishment limits.
51See A. von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993, p. 39; or D. E. Scheid, “Davis,
Unfair Advantage Theory and Criminal Desert”, Law and Philosophy, vol. 14, 1995, p. 408.
52Von Hirsch’s considerations on cardinal proportionality can be found in “Proportionality in the Philosophy
of Punishment”, in M. Tonry (ed.), Crime and Justice, vol. 16, 1992, University of Chicago Press, United
States of America; or “Proportionality in the Philosophy of Punishment: From ‘Why Punish?’ to ‘How
Much?’”, Israel Law Review, vol. 25, 1991. However, his suggestion is only fully developed in Censure and
Sanctions, Clarendon Press, Oxford, 1993.
53 Furthermore, as Braithwaite and Pettit pointed out, we should be troubled “if, within a federal
structure, punishments in State X are twice as severe as in State Y, simply because State X has a lot of
spare cells”. J. Braithwaite and P. Pettit, Not Just Deserts, Clarendon Press, United States of America,
1990, p. 150.
55Moreover, it is not obvious what precisely the talk of a grey-zone implies with regard to whether it is
right or wrong to impose a punishment which lies in this zone.
56Von Hirsch contends that ordinal proportionality includes a “spacing requirement”; see
“Proportionality in the Philosophy of Punishment”, in M. Tonry (ed.), Crime and Justice, vol. 16, 1992,
University of Chicago Press, United States of America, p. 52.
57One problem with the suggestion that the pro rata deflation should stop at the place where the
punishments come closest to the respective disincentives is that this might well be at a point where some
crimes will be punished with an amount of hard treatment which is not sufficient to constitute a
supplementary prudential reason to compliance. However, if a punishment does not fulfil this purpose for
some of the crimes, then it seems that parsimony implies that these crimes should not at all be punished
(as we saw in chapter 1, von Hirsch believes that censure can be conveyed in other ways than through
hard treatment). Another possibility, therefore, is to claim that the deflation should continue until the
point at which a further deflation implies that for at least one crime one will reach below the floor set by
what would be a sufficient disincentive. The problem with the proposal is that, if there are some crimes
which are not very serious but which require a relatively much harder treatment to provide a sufficient
disincentive compared to other more serious crimes, then the deflation might stop at a very early stage. In
fact, it will no longer be obvious that the strategy should be decremental rather than incremental. For both
suggestions it is obvious that adjusting an interval matching to the disincentive floor will be an extremely
complex matter.
58As far as I can see, this suggestion would conflict with the principle of parsimony.
59If it is not meant as a mere truism, that is, if responsibility is not simply defined in such a way so that it
includes a person’s acceptance of the punishment von Hirsch is suggesting, then it does not strike me as
obvious that (responsible) persons will always regard their punishments as acceptable. Is a
consequentialist-minded perpetrator who believes that nothing is gained by imposing on him a
punishment of a certain degree of severity not a responsible person? Or what about a person who feels
that he should be punished even more severely than von Hirsch’s scheme would prescribe - should he be
characterized as an irresponsible person?
60This kind of standardization might, for instance, imply that a person is punished with a degree of hard
treatment which - for the typical person - is sufficient to provide the disincentive but which is not
sufficient for him. But if the hard treatment does not fulfil its purpose, would the principle of parsimony
not then imply that this person should not be punished at all? After all, the infliction of unnecessary
suffering is exactly what this principle proscribes (recall that appropriate condemnatory messages may be
conveyed in ways that do not involve hard treatment). Thus, it is obvious that it is necessary to show that
standardisation coheres with what else is prescribed by the theory.
61I. Primoratz, “On some arguments against the retributive theory of punishment”, Rivista Internazionale
di Filosofia del Diritto, vol. 56, 1979, p. 56.
62See, for instance, A. von Hirsch, Censure and Sanctions, Clarendon Press, United States of America,
1993, p. 41. In fact, it would even be consistent with punishment of the innocent (because the claim that
innocents should not be punished is a cardinal - not an ordinal - requirement).
CHAPTER 5
PROPORTIONALISM AND PENAL PRACTICE
The philosophical discussion of an ethical theory usually consists in maneuvers such

as conceptual clarifications, examinations of deductive implications, considerations
on whether the theory is arbitrary or well-sustained and, in broader terms, on
whether the theory provides a coherent solution to the problem it aims at solving.
The moral investigation of punishment has not - nor in the way it has been carried
out in the previous chapters - constituted an exception to this traditional picture of
philosophical work. However, one thing is whether a theory is theoretically or
formally correct, another is whether it is materially correct in the sense that it is
applicable to the actual world. If a theory suffers from theoretical defects - such as,
e.g. conceptual ambiguity or obscurity - then this may well lead to problems of
application. However, even if the theory is formally correct, this does not imply that
it is practically applicable. In his much celebrated article “Marxism and
Retributivism”, Murphy has directed attention to this fact. As he shows, this insight
constituted the core in Marx’s criticism of the classical retributivists. In Marx’s
view, Kant and especially Hegel failed when it came to “the union of theory and
1
practice”.
What I shall do in the present chapter is to pick up the thread from Marx
and Murphy by considering the question of the applicability of the proportionality
principle in the actual social world in which we live. Naturally, the question of
practical application gives rise to many questions - many more than I have space to
2
consider here. However, I shall focus on two overall questions of which the latter
especially seems to confront proportionalism with a genuine challenge. A challenge
which exists even if it is assumed that the problems that have been pointed out in the
previous chapters could all be satisfactorily answered.
The first question concerns what might be called “practical consistence”. A
general assumption in the ongoing dispute between competing theories of
punishment has been that these theories are, if not correct, then at least free from any
kind of internal contradictions. However, this assumption has been questioned by
some of the modern critics of retributivism. It has been suggested that retributivist
prescriptions cannot consistently be applied when we take the step from the ideal
spheres of philosophical theory to the imperfect and fallible reality of penal practice.
In the following, I shall argue that this challenge of self-defeatingness can
apparently be presented even more forcefully against the proportionality constraint
than the critics have usually believed. However, I shall also show that the challenge
in the end does not constitute a genuine problem for the proportionalist. That I,
nevertheless, find the challenge worth considering is, firstly, due to the two facts that
155
I believe that several misunderstandings have encumbered the retributivist replies to

the challenge and that the challenge, helps clarifying in which way proportionalism
manages to provide consistent action-guidingness despite the fallibility of any real-
life penal system. And secondly, that even though the challenge does not itself
constitute a problem it nevertheless raises a severe priority-problem and,
furthermore, it increases the force of other problems to which we shall later turn.
The second problem of application I shall discuss concerns the question of
how retributivism is affected by considerations of social justice. It will be argued
that the question raises problems for proportionalism when the principle is based on
an expressionist view of punishment, but especially when it is advocated on the
ground of a fairness theoretic foundation.
1. THE CHALLENGE OF SELF-DEFEATINGNESS
An argument which is often presented in the more popular debate on the legitimacy
of the death penalty is that this type of punishment should be rejected because any
system which imposes it on assumed criminals will now and then execute
individuals who are not guilty of the crimes for which they are convicted. The crux
of this argument, namely, that punishment mistakes do occasionally, if only rarely,
occur has not only been suggested as an objection to the death penalty in general but
has also been directed more narrowly against such a penalty as part of a retributive
penal system. In an article assessing the moral bases for capital punishment,
Lempert argues that, though the death penalty may be the punishment which is
deserved for the most serious crimes, it nevertheless can never in practice be
3
justified as part of a retributive system of punishment. Since retributivism forbids
the execution of innocents and since any punishment system practising capital
punishment will, no matter how carefully it is designed, from time to time perform
mistaken executions, it seems to follow that no such system can be retributively
justified. The point is not that the death penalty violates human dignity or any other
possible cardinal proportionality constraints (though, as we have seen, many
retributivists believe this to be the case) but rather that, even if capital punishment
may in some cases be the appropriate punishment, the inevitable mistakes in real life
punishment do themselves violate the retributive principles.
A noteworthy thing about this argument is that, in so far as it can stand a
closer scrutiny, the point to which it directs attention is hardly of relevance only in
the case of capital punishment. Apparently it can easily be extrapolated to other
sorts of punishment. As we have seen in chapter 1, the retributivist injunction does
not only concern the execution of the innocent but punishment of the innocent. And
obviously punishment mistakes will occasionally occur in all sorts of punishment.
To contend - as Lempert surprisingly does - that the argument cannot be stretched to
cover non-capital sorts of punishment because when it comes to these punishments
there is always the possibility that the innocent’s blamelessness is proved and that
the person might then, for instance, be released from prison and be properly
compensated for the time he has served (which obviously is not a possibility in the
4
case of capital punishment) will not do. If it is correct that punishment mistakes do
PROPORTIONALISM AND PENAL PRACTICE 157
violate the basic claims of retributivism, then the important thing is not whether
mistakes may sometimes be revealed and followed by appropriate procedures of
compensation but rather that they will sometimes take place in any feasible system
of punishment. There will be cases where someone’s innocence is not proved and
this is so independently of whether we are talking capital punishment,
imprisonment, fines or other types of punishment.
What the argument aims at establishing is that no matter whether
retributivism seems to provide a theoretically plausible rationale for the punishment
of criminals it nevertheless suffers from a serious flaw, namely, that it cannot
5
consistently serve as a guide for punishment practice. Once we take the step from
theory to practice the retributivist position turns out to be self-defeating. If this is
correct, it is obviously of vital importance with regard to an ethical evaluation of
retributivism. Many would agree that an ethical theory’s capacity when it comes to
action-guidingness constitutes a very important point in the assessment of the
theory. Thus, the argument deserves attention. What I shall argue is that the
argument does not succeed in establishing that proportionalism suffers from genuine
inconsistencies but that it nevertheless is important because it helps in clarifying
how the principle can provide action-guidingness and because it prompts a closely
related, but poorly discussed, problem. Moreover, as we shall see in the following
sections, the argument indirectly has important implications with regard to the step
from theory to practice and does challenge certain versions of the retributive view
on punishment distribution. However, in order to show this, it is important to first
present the argument as forcefully as possible and to clarify the precise content of
the apparent challenge.
Not many participants in the modern discussion of just desert have devoted
attention to the suggestion that punishment mistakes may constitute a problem for
retributivism. However, those who have commented on the question have - as has
Lempert - focused narrowly on the inadvertent punishment of the innocent. Perhaps
this is due to the fact that retributivists themselves have generally done much to
emphasize that punishment theories, such as the utilitarian, which under certain
conditions allow for the punishment of innocents, are highly unacceptable.
However, what matters for a proportionalist is not only that individuals are not
punished if they have not committed crimes, but that each individual receives a
punishment that is proportionate to the crime. Obviously this is not the case if an
innocent is punished, but neither is this the case if someone who is not innocent
receives a punishment which does not appropriately fit the crime he or she has
committed. Punishment of the innocent, therefore, may be regarded as only one
instance of a much broader problem, namely, inadvertent disproportionate
punishment. Now, given the way that proportionalism has been defined, namely, as
at least requiring ordinal proportionality, this widening of the focus is not without
importance. While punishment of the genuinely innocent is something which may
be expected to occur only relatively rarely in a well-functioning sentencing system,
cases of disproportionate punishment may occur with a much different frequency.
To see this, we shall shortly have to return to the questions of what it is that
determines the seriousness of crimes and the severity of punishments.
158 JESPER
THE ETHICS RYBERG
Let us start by recapitulating the discussion of crime seriousness. As we

have earlier seen, the standard view on the matter is that this should be determined
on the ground of harm and culpability. Consider firstly the harm dimension. What
matters in this respect is the harm which a particular crime causes. However,
obviously the harm may differ significantly between two instances of materially the
same crime. Take, for instance, homicide. The questions of whether a person’s death
can be said to harm the person and, if so, how much harm it causes are traditional
philosophical problems. According to one of the most influential modern theories
on the matter, death harms a person who leads a valuable life because it prevents
that person from leading the future valuable life he would have had had he not
6
died. In this view, the amount of harm which a person is caused if he is killed,
therefore, is dependant on the prospect of that person’s future life. But his is a
function both of the degree to which the person’s future life would have been worth
leading and of how much future life the person has been deprived. Obviously, these
matters will differ radically from one person to another. Thus, if this theory is
correct it is reasonable to expect major differences in the harm caused by what is
usually considered to be in the most serious category of crime.
However, even if we leave this admittedly more controversial issue aside
and turn to crimes which in a more straightforward way harm the victims, it seems
reasonable to expect great variations in the harm that is caused. A violent crime may
cause much more harm to the thin-skinned than the callous. And since harms are
generally very much dependant on the psychological and social vulnerability of the
victims, the same variances will also exist in relation to many (perhaps all) other
sorts of crime. And this is so even if the focus is put only on the harm caused to the
direct victim of crime. Once we widen the perspective to include side-effects - e.g.
the impact on a victim’s relatives - and more remote harms, the difference in the
harm caused by two instances of the same type of crime, as well as the difficulties in
estimating the harm in individual cases, become even greater.
The upshot of these remarks is pretty simple. No matter how much is done
by a criminal justice system to determine the harm which, in each individual case, is
caused by a crime, it is unavoidable that all one will in most cases be able to
produce are rough estimations with a pretty large margin of error. Not only are such
estimations very difficult to make but, due to reasons of feasibility, there are strict
limits as to how much time and how many resources a criminal justice system can
use on producing them. The best one can do will in most cases be to rely on some
sort of standardization. As we have seen, this point was explicitly admitted by von
7
Hirsch and Jareborg in their living standard method to gauging harm. That the harm
which is ascribed to a crime in many cases does not correspond to the actual harm
that was caused is thus something which one might very often expect to be the case.
The sources to mistaken judgements of crime seriousness are no less when
it comes to culpability. Given the quadruple mens rea distinction between intention,
knowledge, recklessness and negligence and the problem there is in making
judgements on others’ mental attitudes, it is inevitable that mistakes will occur.
Obviously this risk is enhanced if it is admitted that the way in which the criminal’s
mind was focused on bringing about the fatal result allows for even more fine-
grained mens rea distinctions, such as different kinds of intention or recklessness

each representing different degrees of culpability. The risk obviously also exists
with regard to judgements of responsibility and hence with regard to the degree to
which a criminal should be excused for his misdeed. According to the choice
theoretic conception of responsibility, the degree to which a person may be excused
for a misdeed is, as we have seen, dependant on the extent to which the person was
able to choose freely, which is a question of both the extent to which the person
possessed a sufficient choosing capacity and of the person’s chance of exercising
this capacity. However, both with regard to the physical and mental equipment of an
actor and the circumstances affecting the actor’s opportunity to exercise his
choosing capacity, it seems reasonable to expect a sliding scale of intensity. This is
why excuses may well play an extenuating rather than a full exculpatory role. But
even if it is assumed that it is absolutely clear what it means that a person in some
way or another is deprived of his choosing capacity, it is obvious that judgements on
such complicated matters as mental and circumstantial factors are highly fallible.
This point is not changed markedly if the approach to excuses is grounded on the
character theory of responsibility. Even if is made clear what it means that a person
acted out of character, it seems that precise judgements on the degree to which this
was so in individual cases are indeed fallible as well. Thus, with regard to
culpability, it is reasonable to believe that this is a determinant of crime seriousness
which, in real life, will often be encumbered with mistakes.
Considering factors determining the seriousness of criminal acting, it is
finally worth noticing that, in so far as one believes that respects to prior criminal
record should also count, this is a factor which might itself constitute a source of
mistakes. This is the case independently of whether one adheres to a theory of
8
recidivist premiums or non-recidivist discounts. Suppose, for example, that it is
held that one of the reasons for showing less disapproval for the first than for
subsequent misdeeds is the acknowledgement of human frailty. As I argued earlier,
it is far from clear what this exactly means, but if it is assumed that this is clarified
then it is certainly a possibility that frailty turns out to be something which may vary
significantly in degree from one person to another and which, furthermore, it will be
quite difficult to make precise judgements on in individual cases; this means that the
size of the recidivist discounts or premiums might well be mistakenly determined. A
similar problem relates to some of the other reasons which have been pointed at as
justifying the significance of criminal record, such as people’s capacity to reflect on
wrongs and to restrain themselves accordingly. Obviously, proportionalists
defending the moral relevance of criminal record will usually rely on standardized
judgements but, as in the case of estimations of harm, this is not what basically
matters but rather something which is required of a workable punishment system.
But even if we ignore considerations on what determines the size of recidivist
premiums or non-recidivist discounts, the fact that prior criminal record is accounted
for may itself constitute a further source of punishment mistakes. If a person has
been wrongly convicted and punished then this does not only itself constitute an
instance of disproportionate treatment but it will, if the mistake is not subsequently
revealed, manifest itself in all (or at least a number of) the punishments which the
person may henceforth receive. A single mistake may thus result in various
unjustified aggravations of punishments during someone’s criminal career. So much

for the mistakes which might follow from the judgements of crime seriousness.
The picture does not change if we turn to the question of punishment
severity. As we have seen earlier, the claim that the severity of a punishment should
be determined on the ground of the suffering or deprivation which is inflicted on the
9
punished, confronts proportionalists with the challenge of sensibility. Even though
two persons receive what is objectively the same punishment the impact on the two
may vary significantly. A person’s sensibility is a function of various things. Merely
to claim that a punishment will affect the hard-boiled more than it affects the thin-
skinned is clearly an oversimplified way of presenting the matter (recall Bentham’s
thirty-two circumstances influencing sensibility). Thus, even if the estimates of the
seriousness of the different crimes which two criminals have committed are correct
it will be a highly fallible matter to mete out the appropriate punishments.
In sum, ensuring that perpetrators receive appropriate punishment for the
crime they have committed is a matter which will in practice be extremely
complicated, that is, it is something which contains various sources of mistakes. Two
persons who have committed equally serious crimes may be punished very differently,
and a person who has committed a crime less serious than the one committed by
another may nevertheless in the end happen to receive a more severe punishment. And
this may well be the case even if one defends nothing stronger than an ordinal
matching of the two scales. However, obviously the risk of mistakes will be even
greater if one defends the interval matching which most proportionalists favour. All in
all, in my view, it is reasonable to conclude that in contrast to inadvertent punishment
of the genuine innocent, which may happen only now and then, the imposition of
inadvertent disproportionate punishment is indeed the most likely result of any real life
punishment system no matter how carefully it is designed. That a person receives the
punishment he deserves, that is, the one which is precisely proportionate to the crime
he has committed, will be the exception, not the rule.
To object, as some proportionalists have done, that one should not expect
absolute or mathematical precision in the allotment of punishments to criminals is
not sufficient to avoid the conclusion. If what is meant is that in practice it will most
of the time not be possible to make precise judgements of crime seriousness and
punishment severity then this is exactly the point of the conclusion. On the other
hand, if this is supposed to mean that we should simply not take the proportionality
requirement so seriously, that is, that it does not matter if one perpetrator is
punished slightly more severely than another who has committed a slightly more
serious crime, or that all that matters is that two perpetrators who have committed
equally serious crimes are not punished too differently, then this is simply to give up
proportionality in favour of another punishment distribution principle which of
course will be confronted with the challenge of explaining what the modifications
precisely amount to and how they are justified. Moreover, it is worth noticing that,
given the sketched sources to mistakes, there is no reason to expect that if the
punishments which are allotted to criminals turn out to be disproportionate they will
be so only to a marginal extent. For instance, if a person is judged slightly more
culpable for a crime than he actually was and if the sort of crime he has committed
is, due to standardizations, considered more harmful than it actually was, then the
seriousness of the crime may be significantly overestimated. If we further assume
that the criminal is a person who, due to his psychological sensitivity, will suffer
more from a certain punishment than the average person, then there might in the end
be a significant discrepancy between how the criminal ought to be punished and the
severity of the punishment he actually receives. That is, the criminal might well be
punished in a way which would be considered unjust even from the perspective of a
do-nottake-it-so-seriously version of a punishment distribution principle. However,
we can leave further considerations on possible modifications of proportionalism to
the following chapter. For the present, all that is important is the conclusion that
disproportionate punishment is what a sentencer most of the time will impose on
criminals in real-life penal practice.
With this conclusion in hand, it is now possible to rephrase the initially
presented argument in the following way. Let A denote a sentencer (whether it is the
state or a person acting on behalf of the state is not important here) and B denote a
person who stands to be punished. Given a standard interpretation of
proportionalism as a constraint proscribing disproportionate punitive responses to
criminal conduct, the following dilemma rises. If A either abstains from punishing B
or punishes B in accordance with the prescriptions of a non-proportionalist theory of
punishment, it seems that A will almost certainly be acting wrongly. Of course, it
might happen that the response actually turns out to be exactly what would be fitting
according to proportionalism but this is certainly most unlikely. On the other hand,
if A seeks to punish B in proportion to the committed crime then, due to the outlined
reasons, the punishment will most likely be disproportionate. Hence, no matter what
A does the expectable result is that A will be violating proportionalism. Let us refer
to this conclusion as “the challenge of practical self-defeatingness”. The
challenge presents the core of Lempert’s narrow death-penalty argument in a fully
extrapolated form, suggesting that once we take the step from theory to practice the
proportionality principle must be rejected on the ground of its own prescriptions.
Now, this is obviously a serious challenge which immediately gives rise to two
questions: 1) is the challenge genuine, that is, does it rest on a correct representation
of the content of proportionalism? And 2) does the challenge on closer scrutiny
constitute a problem with regard to applying proportionality as the governing
principle of penal practice? As mentioned, it is not many proportionalists who have
commented on these questions, but those who have apparently believe that 2) can be
ignored by answering 1) in the negative. The two answers which I shall consider
take this form. However, as I shall argue, a more promising way in which a
proportionalist might try to meet the challenge seems to be simply to reject 2).
The first possible answer that easily comes to mind is to contend that the
challenge of self-defeatingness ignores the fact that the proportionalist constraint
should be formulated in terms of mental attitudes. For instance, commenting on the
inadvertent-punishment-of-the-innocent problem, Moore claims that the “probable
punishment of the innocent by any real-world punishment scheme is not much of a
worry .. for deontological versions of retributivism” because “agent-relative norms
bind us absolutely only with respect to evils we either intend or knowingly visit on
specified individuals”10. Consider firstly the former disjunct, i.e. the reference to
what is intended. A standard answer to the sort of moral dilemmas which have have
been the object of much discussion on modern deontology - such as, for instance,
when a person is faced with a situation in which innocents will die whatever he
does, but where it is in his power to decide by his action who shall die - has been to
advocate a distinction between what is intended and what is not intended, holding,
for instance, that what is impermissible is the intended killing of an innocent, not
necessarily the merely foreseen death caused by one’s action. The view is presented
in a general form by Nagel who says that “... to violate deontological constraints
one must maltreat someone else intentionally. The maltreatment must be something
that one does or chooses ... rather than something one’s actions merely cause or fail
11
to prevent but that one doesn’t aim at” . In accordance with this suggestion, it
might therefore be held that what the proportionalist constraint prohibits is intended
disproportionate punishment. It is important to notice what this implies. The claim
obviously is not merely that it is impermissible for A to intend to punish B if the
punishment turns out to be disproportionate. If A punishes B then this is exactly
what will most likely be the case, which means that proportionalism would still be
self-defeating. Hence, what the constraint prohibits is that A intends
disproportionality in punishing B.12 Now, given this exposition of what
proportionalism implies, that challenge is apparently resolved. If A punishes B and
it happens that the punishment does not fully correspond to the one B should have
received then this does not, as long as the disproportion was not intended by A,
violate proportionalism. Hence, inadvertent disportiotionality does not constitute a
problem.
However, though references to what is intended may perhaps in other
contexts constitute an appropriate way of formulating deontological constraints they
do not succeed in the present. The problem is that proportionalism by following this
path avoids the challenge of self-defeatingness only at the cost of becoming much
too permissive. If what proportionalism amounts to is a constraint against intending
disproportionality in punishing, then, if A decides not to punish B, this will - as long
as what explains A’s omitting the punishment is not his belief that his will be
tantamount to disproportionate treatment - be consistent with proportionalism. In
fact, proportionalists may even have to accept the different sorts of utilitarian
punishing which, as we have earlier seen, they have traditionally used much power
in rejecting. Consider the standard example of scapegoat punishing which was
discussed in chapter 1. Suppose that A, in order to prevent certain riots, punishes B,
who has committed theft, of a rape in which he is innocent. A thus decides to
impose a punishment of a certain degree of severity on B, but does A also intend
disproportionality by imposing this punishment on B? As far as I can see, this is not
the case. Following a standard notion of intention, according to which one’s
intentions in acting are defined by which of one’s beliefs about consequences
explain one’s acting in that way, there may be two reasons to hold that A did not
intend disproportionality, even though the inflicted punishment was in fact be
13
disproportionate. If “when A punished B he did not intend to punish
disproportionately” means “it is not the case that A’s punishing B is explained by
A’s expecting the punishment to be disproportionate”, then the claim is equivalent to
either (a) A did not expect the punishment to be disproportionate or (b) A expected
the punishment to be disproportionate but this does not explain his punishing B in
that way. How could (a) be the case? If A does not have any idea of what
proportionalism amounts to or of what the proportionate punishment for theft is,
then A may not expect the punishment of B to be disproportionate. This might, for
instance, be the case if proportionalists have not succeeded in providing a solution
to the anchor problem. But even if we disregard (a), it seems reasonable to hold that
(b) may well be correct. If what matters for A is that he manages to prevent certain
riots then, even if he expects the punishment imposed on B to be disproportionate,
this is certainly not what explains A’s punishing B in that particular way. Thus,
what has apparently happened is that proportionalism has now become consistent
with scapegoat punishing. Likewise, it seems that neither deterrent, incapacitative
nor rehabilitative punishing will be ruled out on the ground of a constraint against
intended disproportionality.
Put more generally: since all that matters for the utilitarian is the
maximization of well-being, whether or not the utilitarian punisher expects a certain
punishment to be disproportionate, is not what explains his meting out punishments
in the way he does. (There is, of course, a possible exception: if we, for instance,
imagine that a terrorist has proclaimed that he will cause an enormous catastrophe
unless a particular person is disproportionately punished then obviously it will
matter for the utilitarian that the punishment which is imposed on the person
actually is disproportionate and the utilitarian will have to adjust the punishment to
assure that it is not proportionate. Thus, this might be a case where the belief that a
punishment is disproportionate is what explains the utilitarian’s acting in a certain
way. But obviously it is not only this sort of instance of utilitarian punishment
which proportionalists will usually object to.) Though the question of intentional
acting is certainly a controversial issue, it seems to me that proportionalists
themselves would not be satisfied with formulating proportionalism merely as a
constraint against intended disproportionate punishing; such a principle simply fails
to capture the “positive” aspect of proportionalism, namely, that what matters is that
14
people actually receive a fitting punitive response.
What happens then if we turn instead to the second disjunct of Moore’s
proposal, namely, that of formulating deontological constraints in terms of
knowledge rather than in terms of intentions? In such a view, what the constraint
prohibits is that one knowingly imposes disproportionate punishment on someone.
That is, if A punishes B in a way which A knows does not appropriately fit the
seriousness of the crime B has committed then the constraint will be violated, and
this will be so even if the disproportion was not intended by A. Perhaps this idea of
doing something knowingly comes closer to our ordinary language usage of what we
regard as intended than does the above suggested somewhat technical interpretation
of intention in terms of desirability-characteristics. For instance, in ordinary
language we are often prepared to say that someone intended something such as, for
instance, a given harm, when he realized that his act would result in the harm and he
deliberately performed the act anyway. Moreover, some philosophers have adopted
this way of speaking. For instance, Sidgwick proposes including “under the term
15
‘intention’ all the consequences of an act that are foreseen as certain or probable” .
Be that as it may, I shall here go on talking about knowledge in order to distinguish
the present idea from the notion of an intention considered above. The interesting
question is whether this exposition of proportionalism provides a more plausible
answer to the challenge.
In one respect this sort of constraint seems more promising than the
previous conjecture: it succeeds, at least to a larger extent, in ruling out the
utilitarian punishing which proportionalists regard as highly unacceptable. If A
punishes B for reasons of scapegoating, deterrence or incapacitation or whatever
other utilitarian grounds there might be for punishing someone, and if A expects the
punishment to be disproportionate then it will be wrong to impose the punishment
even though the prospects of disproportionality is not part of what explains A’s
acting as he does, that is, even though it is not intended. Obviously, there might still
be the problem that A may not know whether the punishment diverges from what
proportionalism would imply. However, even if we leave this problem aside, the
constraint faces another more important problem, namely, that it does not succeed in
answering the challenge of self-defeatingness. Since disproportionality is exactly
what A should expect from imposing a punishment on B, even when A seeks to
punish proportionately, the punishment will in practice be ruled out on the ground
of the constraint. What A knows/expects in punishing B is precisely what the
constraint prohibits.
To object to this claim by pointing out that what the constraint forbids is
only that A knowingly punishes B disproportionately and that in penal practice all
we know is that disproportionality is what we should expect from punishing
someone, that is, that A cannot genuinely be said to know the punishment to be out
of proportion, is hardly an acceptable way out of the problem. If what is meant by
“knowing” something is, in the way the term is used in this context, something
stronger than what we usually mean when we estimate that something can
reasonably be expected to follow from our actions, then it is correct that A’s
punishing of B is not prohibited by that constraint. But the problem obviously is that
in practical life such a constraint fails to prohibit anything. If A is a hardcore
utilitarian who intends to punish B with the purpose of using B as a scapegoat in
order to prevent a more terrible outcome then, even though it is certainly reasonable
to expect the punishment to be very much out of proportion, this will nevertheless
not be forbidden by the constraint because A does not, in a strong sense, know that
the punishment is disproportionate: it might turn out that B actually gets exactly
what he deserves if he has earlier committed the very crime for which is is now
punished but was not, at that time, convicted for. Thus, this interpretation of the
constraint is, from a proportionalist perspective, much too permissive. If, on the
other hand, what is meant by “knowing” something is that we can reasonably expect
it to follow from a certain act, then proportionalism remains self-defeating.
Obviously these considerations on when it is reasonable to talk about
knowing something are not very precise. However, I believe that they are sufficient
to indicate that the answer to the challenge does not lie in a specification of what it
means to know or expect something. If this is correct, then the conclusion is that
neither of the suggested ways of interpreting proportionalism is satisfactory. While a
constraint against intended disproportionality succeeded in answering the challenge

of self-defeatingness only at the cost of becoming much too permissive from a
proportionalist perspective, the constraint against knowingly imposing
disproportionate punishment was more acceptable in terms of permissiveness but
failed to meet the challenge. Thus, if it is correct that the challenge arises from a
misrepresentation of the content of proportionalism then we are still left with the
16
question of how precisely the proportionalist constraint should be put.
However, perhaps the claim that it is wrong to punish people in a way that
is not proportionate to the gravity of their criminal conduct does not need any
revisions. As mentioned, the few theorists who have commented on versions of the
challenge of self-defeatingness seem to believe that, once the proper content of the
constraint is acknowledged, the challenge is avoided. However, it is reasonable to
ask whether the challenge constitutes a problem in the first place. It is well-known
that a theory may be characterized as self-defeating in a number of ways not all of
which render the theory implausible. Thus, to what extent does the challenge
constitute a problem?
The core of the challenge is that, once proportionalism is adopted as the
principle governing penal practice, it will (most likely) violate its own prescriptions.
Or more precisely, that if one, on the one hand, seeks to punish perpetrators in
accordance with this principle the result will nevertheless be that the principle is
violated. If on the other hand, one does not seek to punish proportionately the
principle will be violated as well. Now, obviously this does not amount to a genuine
contradiction: we are not faced with a theory which prescribes that one should do X
and not do X. Neither does this amount to a standard version of a performative
inconsistence. Rather what we have is a theory which implies that no matter whether
you attempt to do X or do not attempt to do X you will (most likely) be acting
wrongly. However, the fact that an ideal ethical theory implies that once we take the
step from theory to practice we will, due to the construction of the world, be acting
wrongly no matter what we do, does not constitute a logical problem. On the other
hand, a theory which says that whatever you do in real life acting you will be acting
wrongly may suffer from another problem, namely, a lack of genuine action-
guidingness. Surely action-guidingness is one parameter which many would regard
as important in the assessment of a moral theory. Singer nicely makes the point
when he says that “ethics is not an ideal system that is noble in theory but no good
in practice. The reverse of this is closer to the truth: an ethical judgement that is no
good in practice must suffer from a theoretical defect as well, for the whole point of
17
ethical judgments is to guide practice” .
However, at this point it is important to keep in mind what proportionalism
implies. The principle obviously does not imply that in whatever way a criminal is
punished that punishment will be wrong. On the contrary, what it implies is that
there is one punishment which is right for a certain crime, namely, the one which is
proportionate to the gravity of the misdeed. The problem is that we will most likely
not end up inflicting this very punishment on a criminal even if we try. But in that
case there is a simple answer to the question of action-guidingness: act in such a
way that it is most reasonable to expect that you will not be acting wrongly, that is,
punish a criminal in such a way that it is most reasonable to expect the punishment
to be proportionate to the crime. The point is not that we will thereby be acting
rightly - on the contrary, most of the inflicted punishments will turn out not to be
perfectly fitting - but that in attempting to punish proportionately we will enhance
the chance of punishing rightly.
Thus, following this idea, the answer to the challenge of self-defeatingness
is that on closer scrutiny the challenge does not constitute a problem for the
proportionalist. What the challenge does is that it directs attention to the fact that
most punishments will, even if imposed by a penal system pursuing proportionality,
be wrong. But this does not prove the principle logically problematic. Nor does it
remove that action-guidingness of the principle. Therefore, in my view this
constitutes the best answer to the challenge. It leaves the proportionalist in a
position where he can simply reject the claim that the challenge constitutes a logical
problem for his position. However, this is not the end of the discussion. An answer
along these lines gives rise to a further question of clear practical significance.
If proportionate punishing is not easily achieved, how far should a society
go in the attempt to ensure that justice is done, that is, that the punishments that are
imposed are in fact proportionate? The more money that is used in investigations
and the more manpower that is engaged in the work of clarifying the nature and
consequences of a crime and of potential punishments in each individual case, the
larger is the likelihood that one will be able to mete out precisely the appropriate
punishment. This apparently prompts a priority question. If the answer is not that
one should do all that can possibly be done within the limits of the society (which is
certainly quite a lot and much more than is done in any real-life penal systems), then
proportionalists will have to explain how - without violating their own device and
without opening the door for forward-looking considerations in a way that
undermines the whole idea of proportionality - it can be justified to compromise
18
with justice out of regard to economic and other circumstances. Rather than
establishing that punishment mistakes themselves constitute a logical problem, the
interesting problem raised by the likelihood of such mistakes, as I see it, concerns
the question of justified priorities. To my knowledge this is not a problem to which
proportionalists have presented a convincing answer (or, for that sake, even been
engaged in). This being said, the challenge of self-defeatingness also prompts other
questions which, as we shall shortly see, have an impact on proportionalist thinking.
2. JUSTICE IN AN UNJUST SOCIETY
The question to which we shall now turn is one that naturally arises in relation to
considerations in the step from theory to practice, namely, how does justice in
punishing relate to justice in a society? Or more accurately, what implications does
injustice in a society have for the possibility of pursuing just punishment? For
instance, should the fact that a theft was committed by a poor person or that a
burglary was committed by a homeless person have any effect on the way these
perpetrators should be punished? The relevance of such questions follows pretty
obviously from the fact that there exist pockets of deprivation and misery even in
those countries which today do most when it comes to social insurance. And,
moreover, that poverty and other forms of deprivation are often - no matter what the
precise causal mechanisms - strongly correlated with criminal behaviour.
What role the question of “just punishment in an unjust society” plays for a
19
retributivist is not an issue which admits of a straightforward answer. On the
contrary, the possible answers may fall into three overall categories. (1) The first
possibility is to hold that considerations of justice in a society are irrelevant in the
sense that they do not at all affect the question of just punishment distribution.
However, this is clearly not the route which modern retributivists usually seem to
follow. (2) Another possibility is to hold that the question is something which may
have implications within a retributivist punishment scheme, that is, something
which the retributivist should take into account in the actual distribution of
punishment. For the lack of any better term let us refer to this as the view that social
injustice is of “internal relevance”. (3) The final possibility would be to contend that
the question of injustice concerns the preconditions for a justified application of
retributivism, that is, that it is not something which merely can be accounted for
once retributivism is put into practice, but rather something which concerns the very
presuppositions for a legitimate application of retributivism in the first place. We
can refer to this as the view that social justice has a “preconditional” or “external
relevance”. Which of the latter two answers is correct from a retributivist point of
view is not something that can be answered in general terms. That approaches to the
matter may differ significantly is not surprising. Whether social justice is of internal
or external relevance must be determined by the precise content of the retributivist
theory. And, as we know by now, retributivism is the heading of theories of very
different shape. In the following, I shall focus primarily on how general justice
considerations affect the fairness theoretic version of retributivism and its
application of the proportionality principle. However, firstly, a few comments will
be made on the implications such considerations have for the other major
retributivist position, namely, expressionism. What I shall conclude is that, for the
latter theory, injustice may lead to new theoretical challenges while, for the former
theory, it constitutes a serious problem. In both cases, it is clear that a discussion of
the issue is highly pertinent in relation to the step from penal theory to penal
practice.
Let us firstly turn to expressionism. As we have seen earlier, there are, once
we dig below the overall idea that punishment should be perceived as a language in
which appropriate condemnatory messages are communicated, wide disagreements
between expressionists with regard to what constitutes the main aim of punishing
and how precisely such a practice is morally justified. These disagreements also
exist when it comes to the question of punishment in a unjust society. Expressionists
have held the question to be of both internal and external relevance.
The first approach is adopted, though not very thoroughly discussed, by
von Hirsch as part of his account of expressionism. Von Hirsch acknowledges that
social deprivation is something which may play a role in relation to punishment
distribution. More precisely, his view is that a socially deprived offender may
deserve a reduced punishment and that this is warranted on the ground of reduced
culpability. Even though von Hirsch rejects the view that social injustice in any way
168 JESPER
THE ETHICS RYBERG
undercuts the legitimacy of applying the proportionality principle, he nevertheless

believes that the granting of such penal reductions will create a host of practical and
theoretical difficulties and that the question therefore remains a perplexing issue.
However, he also adds that, in so far as such mitigations encounter theoretical and
practical obstacles, one “solution” remains, namely, “to keep such matters in mind,
20
when setting a scale’s anchoring points” . This comment strikes me as puzzling.
Suppose that two persons B1 and B2 have each committed a crime and that B2’s
crime is more serious than the one committed by B1 but that B2 is deprived, for
instance, by poverty and living in a grim social environment. Now, the question
which arises is not only how severely B2 should be punished but also how he should
be punished relative to B1. And the latter surely is a comparative question, that is,
something which concerns the ordinal aspect of proportionality. It is therefore
difficult to see how a solution can be provided by cardinal proportionalist
considerations on where the scales should be anchored.
Be that as it may, a more interesting question is how precisely social
injustice may affect a perpetrator’s culpability. At this point I do not find von Hirsch
perfectly clear. However, there seem to be two possibilities. The first would be to
hold that social deprivation is a further factor influencing culpability beyond the
other determinants which have been outlined in chapter 2. That is, that even when
such issues as mens rea and responsibility have been fully addressed there may still
be more to culpability, namely, considerations concerning the degree to which the
offender has been socially disadvantaged. If this is the view which a retributivist
wishes to defend then obviously it is necessary to make clear what precisely it is
that affects the degree of culpability, that is, whether it is the deprivation itself or
something caused by the deprivation. Moreover, this will add further fuel to what
we have previously referred to as the “challenge of absolute comparison”: it will
have to be clarified how social deprivation, in comparison to mens rea and
responsibility, contributes to the overall assessment of a perpetrator’s culpability.
Thus, if this approach is what von Hirsch has in mind, then he his certainly right
when he claims that social deprivation is an issue which encounters serious
theoretical problems for the proportionalist.
The second possibility, which is theoretically more attractive, would be to
contend that social deprivation has an impact on the standard determinants of
culpability. An obvious suggestion would be that social deprivation plays a role in
so far as it affects the deprived person’s responsibility. For instance, on the ground
of a choice theoretic conception of responsibility, it might be held that social
deprivation, at least in certain forms, does not provide a person with a fair
opportunity to exercise his capacities for doing what the law requires and that the
person consequently should not be judged fully responsible for his violation. Thus,
according to this suggestion, respects to social deprivation may perhaps raise
practical problems but no further theoretical challenges: social deprivation may be
handled within an existing theory of excuses. This, I believe, was the view which
Hart had in mind when he said that one should “incorporate as a further excusing
21
condition the pressure of gross forms of economic necessity” . Other more recent
22
theorists have advocated corresponding proposals. Be that as it may, as mentioned
not all expressionists have held social deprivation to be of internal relevance.
A more radical view of the consequences social injustice may have with
regard to the realization of an expressionist penal system is defended by Duff.
Against the background of his theory of communicative penitential punishment, he
holds that social injustice is of external relevance, that is, it is something which
affects the preconditions for the application of the theory. One of the preconditions
that must be presupposed, but whose existence in reality must be regarded as
doubtful, concerns the answerability of a defendant. What this implies is that a
person who has not had his fair share in the rights and goods of citizenship but who
has been injustly disadvantaged may not be bound by the demands and requirements
of the law and, therefore, may not legitimately be regarded as answerable for his
crime. Moreover, if the collective treatment of a person has excluded him from
citizenship then it does not seem reasonable to hold him answerable to the political
community (no matter whether the person was bound by the law). According to
Duff, it is doubtful whether these preconditions of answerability are satisfied for
many of those who appear before the criminal courts. But how does this affect the
application of his theory?
Duff has in several writings expressed his pessimism when it comes to the
step from theory to practice. He has suggested that his theory may constitute an
ideal which cannot be transferred to practice until the necessary preconditions are
satisfied (and that, meanwhile, an alternative may be some version of a deterrence
23
theory) . However, more recently he has said that he finds himself veering back
and forth between “on the one hand, a wholly pessimistic view that under present -
and foreseeable - conditions criminal punishment simply cannot be or become such
a mode of moral communication; and the slightly more optimistic view that, once
we grasp the fact that ‘the criminal justice system’ is less a monolithic and unitary
institution than a set of diverse and partly autonomous sub-systems and practices,
we will also see that there may be room, in some contexts, for at least modest efforts
24
at a communicative penalty” . Thus, despite the touch of optimism in the latter of
the views, there still remains an element of openness with regard to the question of
what should be done in actual penal practice.
The differences between the different versions of expressionism mean that
it is not justified to draw any general conclusion except the one that injustice does
play a role and that there may still be more to be said on the matter. As we shall now
see, the implications are more far-reaching when we turn to the traditional
retributivist rival to the expressionism, namely, the fairness theory. I shall therefore
consider these implications at some length.
That considerations concerning injustice in the society may play a role for
the fairness theoretic approach to punishment is not from the outset something that
is unexpected. As we have seen, the theory of punishment is usually presented as
part of a more general view on justice and some of its advocates have specifically
objected to the idea of a compartmentalization of justice into various, mutually
25
independent, spheres. How precisely the question of social deprivation, as one
possible aspect of a discussion of injustice, effects the theory cannot be answered in
170 JESPER
THE ETHICS RYBERG
general terms. As in the case of expressionism it depends, as we shall see, on the

more precise content that is given to the theory. However, it seems clear that the
theory in general does rest on a number of preconditions concerning justice. As we
have previously seen, the fairness theory begins with a conception of society as a
cooperative endeavour secured by coercion. More narrowly, punishment as a
practice is justified because it is necessary for the maintenance of a social order: a
fair balance between benefits and burdens. A criminal, on the one hand, takes
advantage of the opportunities which the legal order affords but, on the other, gains
an unfair advantage by not contributing to the preservation of that order. Things are
not even until this advantage is erased, that is, until the equilibrium of benefits and
burdens is restored which is exactly what an appropriate punishment does. The idea
of an initial equilibrium or social order is clearly expressed, for instance, by Dagger
who says that “punishment is justified only when there is a just balance of benefits
26
and burdens to begin with - when the social order is just, or reasonably so” .
Now, what does this initial order more precisely consist of? An obvious
answer would be that what fairness theorists are presuming is an equal distribution
of benefits and burdens. That is, that members of the society must gain equal
benefits and carry equal burdens. This is a view which several exponents of the
theory advocate. For instance, Sadurski explicitly talks of an “equal distribution of
27
... benefits among the members of the community” and Murphy talks of the
28
fairness of bearing “comparable” burdens. Likewise, Morris, in his original
modern formulation of the theory underlined the idea that fairness dictates that
29
“benefits and burdens are equally distributed” . Moreover, there seem to be good
reasons in favour of this view. Conversely, to hold that all that matters is that
members of the society have a share in benefits and burdens but that it does not
matter whether these benefits and burdens are of equal size seems to contradict the
idea of fairness and, furthermore, would undermine the possibility of delivering a
fairness theoretic foundation for proportionalism. Suppose that A and B do not
enjoy the same degree of benefits from the law. In that case, it would not seem fair
to hold that they should bear the same degree of burdens. But in that case, the
degree of burdens which A and B could each rightly be expected to bear may vary
from one to the other, which means that they may gain different degrees of unfair
benefits from committing the same crimes - so that, finally this means that the
legitimate degree of punishment for this crime may be different depending upon
who perpetrates it. Conversely, if it is assumed that A and B bear different burdens
but perhaps gain equal benefits then once again this can hardly be held to be fair.
And even if it was in fact regarded as fair or at least acceptable that some bear only
a relatively less heavy burden than others, then there would no longer be reason to
mete out punishment: if the unequal distribution of burdens is not a problem then
neither, it would seem, does it constitute a problem if this distribution is reached
through criminal conduct. There would be no initial distribution that ought to be
restored. Therefore, let us assume that the initial just balance of benefits and
30
burdens which fairness theorists have in mind is in fact an equal distribution.
However, what the theory ideally presupposes is one thing, what however
should we expect once we take the step from theory to practice? In other words, can
we in reality reasonably expect there to exist the sort of social order which the
fairness theory presumes? There is no straightforward answer to this question. On
the contrary, what we will have to do in order to provide an answer is obviously to
address the controversial question of what precisely benefits and burdens consist in.
As mentioned earlier, this is a question to which different fairness theoretic answers
have been suggested. Thus, what I shall do now is, firstly, to consider the question
by examining different conjectures with regard to what a benefit amounts to and,
secondly, turn to the same question with regard to burdens.
What kind of benefit is it that one gains from engaging in a cooperative
enterprise such as the rule of law? The obvious answer is that one gains from the
cooperation of other members of the society, that is, and more precisely, one gains
from the self-restraint of others. Once others restrain their actions in accordance
with the law, it becomes possible to live without others’ harmful interferences.
Dagger puts it clearly when he says that under the rule of law everyone “is free to
31
act, to enjoy his or her right, with a security that would otherwise be impossible” .
This, in his view, is the benefit everyone shares. Along the same lines, Sher
underlines that one benefits from other persons restraining themselves from
wrongful activities. As examples he mentions that one may benefit by not being
32
physically assaulted or by not being defrauded. However, if this is what is
understood by the claim that “one benefits from the rule of law”, what then does this
imply with regard to the presumption of an initial equal distribution of benefits?
If one considers particular laws then it is hard to see why one should
expect that the presumption is, or perhaps even could be, satisfied. The degree to
which one benefits from others restraining themselves obviously depends on the
degree to which it is reasonable to assume that one would have suffered from
another’s activities if these persons had not restrained themselves by submitting to
the law. If a person were in a position in which he could not become the victim of
physical assault then he would not benefit from a law against such actions. Now, in
real life none of us are above the risk of becoming victims of assault. However, it is
equally obvious that the risk that we actually become victims of assault may vary
much from one person to another. One is much more likely to be physically
assaulted if one is a prostitute living and working in a bad neighbourhood than if
one is an old lady living in a good neighbourhood almost never leaving her
apartment. The same picture counts for many other particular laws. The risk of
being wronged if the law did not exist may vary depending on one’s circumstances
(e.g. one’s address, job, gender, age, income etc.) which means that the law will not
provide the same degree of “security” to each and, consequently, 33
that the actual
benefit one gains from the law will vary from person to person. In fact, there are
some laws from which certain persons do not benefit at all. For instance, a law
against embezzlement protects only those persons who are in a position from which
34
they can be embezzled. Thus, in so far as such a law should not simply be
eliminated from the total set of laws - which no one to my knowledge has suggested
- it follows that benefits from particular laws cannot be assumed to be equally
distributed. Neither can an equal distribution be obtained by simple changes in the
existing order. But is it possible to get around these objections?
172 JESPER
THE ETHICS RYBERG
An approach which some adherents of the fairness theory have defended is

to reject the idea that benefits relate to particular laws. This is what Sher holds when
he suggests that, even if someone has not benefited from a particular law, such as a
law against embezzlement, the person may still have benefited from other persons
35
restraining themselves from many other sorts of wrongful activities. And Dagger
explicitly underlines that “reciprocity does not mean that everyone must benefit
from and feel the burden of each and every law”. On the contrary, he believes that
“it is the system of laws - laws as a cooperative practice - from which each must
36
benefit ...” . However, to suggest that this appeal to generalized benefits answers
the distributive problem that rises from relating benefits to particular laws strikes me
as an unconvincing maneuver. It is simply hard to see why the benefit gained from
the whole system of laws should turn out to be the same benefit for everyone. It
seems reasonable to assume that the total benefit a person enjoys from the system of
laws must in some way or another be a function of the benefit gained from
particular laws. However, since the latter benefit may, as we have just seen, vary
from person to person there is no reason to believe that everyone will share the same
total benefit. Suppose that A and B are participants in a cooperative practice,
namely, a law system which consists of only two laws: a law against embezzlement
and a law against physical assault. Suppose, further, that A is in a position from
which he can be embezzled but that this is not the case for B, and that A lives under
conditions which would make it much more likely that he, in comparison to B,
would have been the victim of assault had it not been prohibited by law. In such a
case, it would seem that A as opposed to B benefits from the law against
embezzlement and furthermore that A benefits more from the law against assault
than does B. But if this is so, it certainly does not make much sense to hold that A
and B are equally advantaged as long as we consider the general benefit gained by
the system of laws. Now, by containing only two laws this example of a law system
is obviously extremely simplistic, but there is certainly no reason to believe that the
point that is illustrated would be changed by that fact that criminal law systems in
37
practice contain a much greater number of laws.
Thus, in sum, it seems that no matter whether one considers benefits from
particular laws or from law systems there is no reason to believe that there will be
an equal distribution of benefits to begin with. And, what is perhaps more
important, neither does it seem that such an initial order can easily be realized by
changes in the existing order. What precisely this means for the fairness theory is a
question to which I shall return later. However, first we have to consider whether
the idea of an equal distribution of burdens constitutes a more plausible initial
assumption. This question - as the discussion in chapter 1 witnessed - is complicated
by the fact that there is disagreement with regard to what precisely the burden that
the law imposes consists of. As far as I can see, three different proposals can be
extracted from the literature.
The first and most obvious suggestion as to how obedience to the law
constitutes a burden is to contend that the law-abiding person is one who abstains
from performing certain acts which he would otherwise - had the law not existed -
have performed. This idea was advocated by Morris who held that the “burden
38
consists in the exercise of self-restraint by individuals over inclinations ....” . The
criminal, thus, is one who fails to exercise self-restraint by giving in to his
inclinations and who, thereby, renounces the burden which others have assumed.
We have earlier seen that this suggestion leads to problems with regard to the
question of the relative ranking of crimes in seriousness. However, it also seems
difficult to reconcile it with the assumption of an initial equal distribution of
burdens. It simply does not seem plausible to hold that everyone has the same
inclinations. An inclination is very much determined by circumstances and personal
preferences and is consequently something which may well differ from person to
person. There is no reason to believe that the inclinations A has to perform a certain
criminal act are at all shared by B. And even if A and B both have inclinations to
perform a certain criminal act there is - given the natural assumption that the size of
a burden is determined by the strength of an inclination - no reason to believe that
they are inclined to the same extent; which means that they may nevertheless
undertake different burdens by obeying the law. Moreover, as in the case of
benefits, it is once again clear that the claim that one should not focus on the
burdens from obeying single laws but rather on the burden from the system of laws
39
(which e.g. is what Davis regards as “a plausible answer” to the fact that not
everyone has to restrain himself from committing each and every possible crime)
has nothing to offer if one admits that the size of the overall burden in some way is
a function of the burdens related to particular laws. All that may follow from this
step is perhaps that everyone bears a burden (A may restrain himself from
committing some crimes while B restrains himself from committing others), but it
does not even tentatively follow that these burdens are equally large. Thus, on the
first interpretation an equal burden-distribution seems very far from reality.
The second proposal that has been defended and which apparently
succeeds in avoiding some of the problems which the first suggestion gave rise to is,
roughly, to hold that a burden consists in a limitation of freedom. This view has
been advocated by Sadurski who underlines that what he has in mind has nothing to
do with the actual inclinations persons have for acting in one way or another.
Sadurski admits that there are people who do not consider a duty to refrain from
murder, assault or other criminal acts as an actual inconvenience. But in his view
this does not change the fact that the law limits people’s freedom. As he puts it “the
point is that those restraints that are prerequisites for the effectiveness of rights can
be presented reasonably as burdens upon a person’s life since they cut off a range of
40
options which would be otherwise available to him” . That the range of options is
reduced is burdensome independently of whether the law-abiding person
experiences himself as being in any way constrained or deprived. Thus, the mere
fact that the prohibitions resulting from the criminal law reduce the number of
possible options, compared to a hypothetical situation in which there was no
criminal law, is all that is needed to draw the conclusion that the law imposes a
burden on those who submit to it. But if that is the case then it seems reasonable to
hold that we might well expect an initial equal distribution of burdens. As long as
the law counts for everyone, it seems that all members of the society bear the same
burdens since the law formally cuts off the same options for all. Sadurski’s
suggestion in this respect seems more attractive than the first proposal.
Nevertheless, I believe there are reasons to question its plausibility.
Firstly, it is clearly not sufficient that a certain interpretation of what a
burden consists of makes it reasonable to assume that the preconditions concerning
a just social order are satisfied. The interpretation will also have to be ethically
plausible. One question that might be raised is whether it is plausible to hold that a
mere reduction in the range of possible options is in itself burdensome. Is it really
correct that I am in an interesting sense worse off if I may choose only from among
apples and pears rather than from apples, pears and bananas, even if I do not like
bananas? Dworkin has, in an influential article, defended the view that the question
41
“Is more choice better than less?” may well be answered in the negative. However,
I shall not here engage in this discussion. But rather turn to a closely related
question which is left open even if Dworkin’s position is mistaken, namely, whether
the mere formal burden of reduced options is what we should focus on rather than
some sort of experienced burden. Suppose that A and B have hitherto had the
possibility of choosing between three options but that option number three is now
suddenly closed. Suppose, further, that A does not care about this, because he would
never have dreamt of choosing option three while, on the other hand, it strikes B
hardly because he has always found option three the preferable one. Now, in one
sense it might be held that, due to the mere fact that an option has been closed, both
A and B are worse off. However, in another sense, it might be held that B - but not
A - is worse off because he experiences the closed option as very burdensome. In
order for Sadurski’s proposal to be plausible, it will have to be argued that it is only
a burden in the former formal sense and not in the latter experienced sense that
should count when considering the burdens imposed by the criminal law. However,
it is far from clear why this should be so. It is certainly not self-evident - in fact, my
intuition points in the opposite direction - and, to my knowledge, no convincing
argument to this effect has been produced.
Secondly, if one holds that the burdens imposed by the criminal law consist
in the fact that the law-abiding person will have to observe self-restraint in the sense
that has to do with the fact that certain options are prohibited, but not in the sense
that relates this burden to a person’s inclinations to break the law then, on the one
hand, one may have succeeded in avoiding the above outlined problems which an
inclination-based account of burdens leads into but, on the other, one would be left
with the problem of providing an alternative account of what determines the size of
a burden. To claim that the legal prohibition of an act constitutes a burden
independently of what kind of act we are talking about, that is, that each option that
is cut off constitutes a burden but that there is no variation in the size of burdens,
would be unacceptable. This would imply that the unfair benefit a criminal gains by
renouncing a burden would be the same no matter whether we are considering acts
of minor theft or serious assault. Since this is surely not what Sadurski has in mind,
and since references to inclinations are ruled out, we need another answer as to what
makes one burden more burdensome than another. Sadurski’s answer seems
somewhat puzzling. He holds that “if in the generally accepted hierarchy of values
freedom from assault on our life is more precious than freedom from invasions of
our property, then the benefit of non-restraining oneself with regard to the former is
42
higher than with regard to the latter” . That is, the burden which a criminal
renounces is larger relative to another burden if “the benefits he removes from his
43
victim are more precious to the victim” relative to another benefit.
However, this connection between burdens and benefits strikes me as
muddled. Why do I bear a larger burden, that is, perform a higher degree of “self-
restraint” if the person who benefits from my self-restraint regards the benefits as
more valuable? I must admit that I fail to see how this connection can leave any
substantial idea of what it means to bear a burden. But a precise conception of what
a burden is and what determines the size of burdens is required if the theory is to be
lifted from the ground in the first place, and in particular if it is to provide a
rationale for proportionality in punishing. Be that as it may, there is another problem
which seriously challenges the idea that such burdens can be expected to be equally
distributed. However, since this is a problem which confronts all other accounts of
burdens as well, I shall postpone discussion of it and turn firstly to a final account of
what constitutes a burden.
This third account is presented by Sher, whose version of the fairness
theory has already been considered earlier at some length. Sher’s account has clear
affiliations with Sadurski’s in that it starts out by rejecting the idea that the burdens
of the law-abiding (and consequently the unfair benefits gained by the criminal)
have anything to do with the exercise of self-restraint over inclinations or impulses.
What Sher contends is that a burden is a “moral restraint”, and he thereby also
believes one is able to provide a measure of the size of burden: one bears a larger
burden, he suggests, if one submits to a moral prohibition of greater seriousness.
The more serious a moral prohibition is, the larger the burden carried by the law-
44
abiding, and the more one gets away with if one violates the prohibition. The
affiliation with Sadurski’s account, in my view, makes Sher’s account of a burden
vulnerable to the same kind of objection, namely, that it is far from clear that we are
actually presented with any substantial conception of a burden, namely, something
we can recognize as being genuinely burdensome. What is meant by bearing a
burden of moral restraint? As we know, this cannot mean that one exercises self-
restraint by not giving in to one’s inclinations. This would imply that a person who
has no inclinations to perform an illegal act would not bear a burden. We would be
faced with the problems facing the first account of a burden outlined above. Neither
can it mean that one simply has an obligation not to act in a certain way. It is not
clear why this is burdensome and, furthermore, it would undermine the crucial
assumption that a criminal gets an unfair benefit by renouncing the burden: a
criminal is certainly not released from the obligation simply because he violates it
(on the contrary, it is precisely because he is under an obligation that his violation is
wrong). Thus, on this account the law-abiding as well as the criminal would carry a
burden and the whole theory would fall apart. Thus, in my view, it is hard to see
45
what should be understood by the idea of a burden of moral restraint.
Now, what do the previous considerations on burdens show? The first thing
we saw was that, in the case of the first and apparently most obvious interpretation,
there were good reasons to doubt that there will ever be the initial equal distribution
of burdens which the theory presupposes. With regard to the other two
interpretations, I have suggested that, by distancing themselves from the first

interpretation, it becomes much less obvious that what we are talking about is
something that can be identified as a genuine burden. However, if we leave this
objection aside, is it not the case that these two interpretations at least make it
reasonable to believe that the requirement of an equal distribution of burdens would
be satisfied? And even if we ignore the three accounts of burdens which have here
been considered and which are those that can be extracted from the discussions,
might it not be the case that other interpretations could be presented which would
make the distribution requirement a plausible assumption? Even if we assume that
these questions could be answered affirmatively there is, however, a final objection
which in a slightly different way undermines the assumption of an initial equal
distribution of burdens and which holds independently of whatever interpretation of
a burden a fairness theorist is defending. The objection arises from the
considerations we have dealt with in the first section of this chapter.
The core of the challenge of self-defeatingness was the conclusion that the
most likely result of being punished within a proportionalist punishment system is
that one does not receive a punishment which is precisely proportionate to the
gravity of the crime one has committed. Due to all the difficulties of making even
tentatively precise estimations on the many factors which determine the seriousness
of crimes and the severity of punishments, perpetrators will, in most cases, end up
receiving punishments which do not fully fit their crimes. If this conclusion is put in
fairness theoretic terms, it means that the burden which is forced upon a criminal by
punishing him will most likely not perfectly correspond in size to the unfair benefit
he has gained by his criminal act. He is most likely to be punished too hardly or too
mildly which means that he, after having served his sentence, will end up with either
a deficit or a surplus of burdens compared to the level of burdens of law-abiding
members of the society. The punishment will usually not perfectly succeed in
restoring the initial equilibrium. But if this is so, then it has serious consequences
for the application of the fairness theory. If we imagine a hypothetical initial
situation in which no one has yet violated the law and in which the law leaves a
perfectly equal distribution of burdens (and benefits) between the members of the
society, then the punishment system itself will, once it is applied to those who
commit crimes, end up by eliminating the equal distribution. The application will
thus pull the carpet from under the very justification of the system. And notice that
this problem arises independently of how precisely the concept of a burden is
interpreted. Therefore, even if it is correct that an interpretation of a burden can be
given which makes it reasonable to believe that there might exist an initial equal
distribution of burdens, this distribution will be torn apart once the system is
applied, which means that further application will no longer be justified.
The discussion of the justice preconditions on which the fairness theory is
based has thus led us to the following conclusions. With regard to the requirement of
an initial equal distribution of benefits among members of the society, we have seen
that there are good reasons to hold that such a distribution does not exist. The same
is the case with regard to the distribution of burdens. And even those interpretations
of a burden which would apparently make it most reasonable to believe that an
equal distribution might exist would still be hit by the problem that, due to the
imperfection of the punishment system, a punishment practice will itself undermine

such a distribution pattern if it was existing in some initial situation. Thus, the
presumption of an initial situation in which each member of the society enjoys the
same degree of benefits from the law and bears the same degree of burdens is most
likely not satisfied when we turn from mere theory to reality. And two further
comments should be added to this conclusion.
Firstly, it is important to notice that the conclusion is not merely that the
required distribution pattern does not presently exist. If this was all that could be
concluded it would not necessarily be a problem. For instance, it might simply be
suggested that one should effectuate policies designed to realize the required
distribution patterns and which would thus prepare the ground for the application
for a fairness theoretically-based punishment system. However, an examination of
the different accounts of benefits and burdens has made it clear that such policies
46
are doomed to fail. The conclusion, therefore, is the stronger one, namely, that the
distribution pattern cannot be expected to be realized in practice - not merely that it
is not presently realized.
Secondly, it should be noticed that the outlined problems do not arise
merely from the fact that we have assumed an initial equal distribution of benefits
and burdens. It might be suggested that what matters is not that all members of the
society gain the same degree of benefits and the same degree of burdens but rather
that there exists some reasonable initial distribution pattern. For instance, in a
general discussion of the fairness principle, Arneson has suggested that at bottom
the moral intuition behind the principle is the same as the intuition that “it is right to
pay the grocer for our groceries or to pay rent for the use of land improved by the
47
landowner” ; in short, we should pay for the benefits we gain. But if that is so, it
might be held that an unequal distribution in itself does not constitute a problem:
what matters is rather that the benefits a member of the society gains stand in a
reasonable proportion to the burdens he bears. If one gains less from cooperation
one should bear a correspondingly minor burden. Equality is therefore not required.
However, this suggestion clearly does not solve any problems. In the same way as
the analysis of benefits and burdens does not make it reasonable to believe that an
equal distribution can be realized, neither is it reasonable to expect that a
distribution based on proportions between benefits and burden or, for that sake, any
other possible distribution pattern that one might come up with, would be realizable.
But the fairness theory obviously needs to make assumptions on an initial morally
important distribution pattern. Otherwise there would be no sort of pattern which is
disturbed by criminal conduct and no pattern that should be restored by the
imposition of punishments on criminals. The fairness theory of punishment would
fall apart.
If the conclusion we have reached is correct, then we are faced with one
final question, namely, what do all these considerations on justice presuppositions
mean for the application of the fairness theory? The answer depends upon whether
one adopts the view that social justice is of external relevance or of internal
relevance. If the first is the case, that is, if one believes that the existence of an
initial distribution pattern of benefits and burden constitutes a precondition for the
178 JESPER
THE ETHICS RYBERG
legitimate application of the fairness theory, then the consequences are pretty
48
radical. Since the preconditions cannot be satisfied, the fairness theory will not
survive the step from theory to practice. We end up with a theory constructed for the
spheres of philosophical theory but not for the reality of penal practice. This, I
believe, should be regarded as devastating for the theory.
The only way to avoid this conclusion would be to adopt the second
approach, that is, to maintain that social justice is of internal relevance in the sense
that it does not constitute a precondition for the application of the theory but rather
that injustice is something that should be accounted for in an appropriate way when
we turn from theory to practice. For instance, this might be done by suggesting that
if A and B bear the same burdens but A does not gain as much benefit from the law
as does B, then if A commits a crime and thus gains an unfair benefit he should not
be punished as severely as if B had committed the same crime. In this way the social
distribution problems might be accounted for in the distribution of punishments. I
shall not here enter a further discussion of possible ways in which this might be
done, nor engage in analysis of the problems this may raise. But there is one point
which, given the overall purpose of our analysis, it is important to notice, namely,
that this would be tantamount to giving up the idea of proportionality in punishing.
Due to different shares in the distribution of benefits and burdens, one person may
end up being punished more severely than another even though the first has
committed a less serious crime, or two persons who have committed the same crime
may end up receiving punishments of very different degrees of severity. The
proportionality constraint would have to be abandoned.
3. CONCLUSION
Not much attention has been directed to the possible problems of application in the
modern retributivist literature. However, views are split among those who have
commented on the issue. On the one hand, Murphy has resignedly drawn the
conjunctive conclusion that a fairness theoretic version of retributivism is “the only
morally defensible theory of punishment” and that arguments which are Marxist in
spirit “can be formulated which show that social conditions as they obtain in most
societies make this form of retributivism largely inapplicable within those
49
societies” . On the other hand, several retributivists have emphasized applicability
as one of the merits of modern retributivism.
When it comes to the question of punishment distribution, the previous
discussion indicates that the proportionality principle is not exempted from problems
of application. The first issue that was considered was the challenge of self-
defeatingness. The premise that - due to the complexity of meting out punishments
precisely proportionate to the crimes committed - disproportionality in punishing is
the most likely outcome of a retributivist penal practice, led to the conclusion that
what a proportionalist penal system will be doing most of the time will be wrong
according to the proportionalist’s own standards. After having rejected the idea that
this problem should be avoided by reformulating the constraint in terms of mental
attitudes, I argued that the challenge was not devastating to proportionalism.
However, it gave rise to an important, but ignored, question concerning priorities,

i.e. more precisely, how much should be done within a society to ensure that justice
is done. The next aspect of the application issue I turned to was the question of just
punishment in an unjust society. The diversity of the foundations of proportionalism
did not allow for one general conclusion. What we have seen, however, is that while
injustice had implications for expressionism - at least by complicating the
imposition of just punishments - the implications were even more radical for the
fairness theory which was forced into the dilemma of either giving up
proportionalism or turning out as genuinely inapplicable. Thus, independently of
whatever other problems might face proportionalism, the principle does not invite a
simple step from penal theory to penal practice.
180 JESPER
THE ETHICS RYBERG
NOTES
1J. G. Murphy, “Marxism and Retribution”, Philosophy and Public Affairs, vol. 2, 1973. I here quote from the
republication in J. Simmons, M. Cohen, J. Cohen & C. R. Beitz (eds.), Punishment, Princeton University
Press, United States of America, 1995, p. 18.
2For a critical discussion of retributivist penal practice see, for instance, J. Braithwaite and P. Pettit, Not Just
Deserts, Oxford University Press, Great Britain, 1998
3R. O. Lempert, “Desert and Deterrence: An Assessment of the Moral Bases of the Case for Capital
Punishment”, Michigan Law Review, vol. 79, no. 6, 1981.
4See also L. Alexander, “Retributivism and the inadvertent punishment of the innocent”, Law and
Philosophy, vol. 2, 1983, p. 235.
5See A. Wertheimer, “Punishing the innocent - unintentionally”, Inquiry. vol. 20, 1997; M. Philips, “The
inevitability of punishing the innocent”, Philosophical Studies, vol. 48, 1985.
6See, for instance, B. Williams, “The Macopulos Case: Reflections on the Tedium of Immortality”, in
Problems of the Self, Cambridge University Press, Cambridge, 1973; T. Nagel, “Death”, in Mortal Quetions,
Cambridge University Press, Cambridge, 1978; J. McMahan, Ethics, vol. 99, 1988.
7See chapter 2 section 1.
10M. S. Moore, Placing Blame, Clarendon Press, Oxford, 1997, p. 158.
11T. Nagel, The View from Nowhere, Oxford University Press, New York, 1986, p. 179.
12Along the same lines, Duff says that we are not “committed to the kind of perversion of punishment which
a consequentialist must allegedly sanction, so long as we not aim to produce punishment of the innocent”. A.
Duff, Trials and Punishment, Cambridge University Press, Great Britain, 1986, p. 159.
13For instance - to take Bennett’s example from his excellent general discussion of intentions - if a wife
believes that by wiping her husband’s brow she will make him feel more comfortable and that she will make
the handkerchief wet, it may nevertheless only be the first belief which explains her action. J. Bennett, The
Act Itself, Clarendon Press, Oxford, 1995, chp. 11. See also R. A. Duff, Intention, Agency and Criminal
Liability, Blackwell, Great Britain, 1990, part 1.
14A possible conjecture which might explain what is wrong in the different cases of utilitarian punishing can
be found in formulations of the Doctrine of Double Effect which permits certain conduct that leads to bad
results only if (among other conditions) it is true that: the good which follows does not flow from the bad
and/or the agent does not intend the bad as a means to the good. This principle - sometimes referred to as “the
means principle” - refers both to a subjective condition: what is being intended, and an objective condition:
the bad effect being a means to the good effect. It might be claimed that the latter condition itself is what
makes utilitarian punishing wrong and that this condition must therefore be added to the formulation of
proportionalism. However, firstly, it is not difficult to construct examples which show that the objective
constraint on the causal structure leads to absurdity (see, for instance, the excellent example in A. Norcross,
“Intending and Foreseeing Death: Potholes on the Road to Hell”, Southwest Philosophy Review, vol. 15, no.
1, 1999). Secondly, even if this condition were to be somehow added to the formulation of what
proportionalism claims, it would still be insufficient to rule out other situations which the proportionalist will
strongly oppose such as if A simply omits punishing B, or if A imposes an extremely tough punishment on B
for reasons which do not make the punishment a means to producing some good.
Another conjecture is made by Dworkin, who points out that the deliberate framing of an innocent
is different from mistakenly convicting him because the former, unlike the latter, involves a lie. However,
even if we adopt this idea, assuming that lying is something morally bad, it does not help much in the present
context. Imposing disproportionate punishment of criminals for reasons of deterrence or incapacitation surely
need not involve lying. Thus, adding this idea to the way proportionalism is put does not help with regard to
the problem that the view may be much too permissive from the proportionalist’s own point of view; see R.
Dworkin, “Principle, policy, procedure”, in C. F. H. Tapper (ed.), Crime, Proof and Punishment, London,
1981.
15H. Sidgwick, The Methods of Ethics, Dover Publications (republication of the 1907 edition), New York,
1966, p. 202.
16It might be suggested that further specification should be added with regard to what it is one is prohibited
in doing knowingly. As indicated in the above quotation, Moore believes that we are bound with respect to
evils which we knowingly visit on “specified individuals”. In my view, neither does this addition succeeds in
answering the challenge of self-defeatingness in a satisfactory way. However, since in the end I do not believe
that the challenge constitutes a genuine problem, I shall not discuss the suggestion any further. For related
discussions see, for instance, C. Fried, An Antonomy of Values, Harvard University Press, 1970, chap. 12; or
J. Glover, Causing Death and Saving Lives, Penguin Books, 1990, pp. 210-13.
17P. Singer, Practical Ethics, Cambridge University Press, United States of America, 1993, p. 2.
18To reply that proportionalism does not imply any obligations whatsoever, with regard to the collection of
information concerning the nature of a crime, does not seem plausible. Surely, the reason for engaging in
investigations - according to the retributivist - must be to ensure that justice is done.
19For a short presentation of some of the retributivist approaches to the “just punishment in an unjust society”
question see M. Tonry, “Proportionality, Parsimony, and Interchangeability of Punishments”, in A. Duff and
D. Garland (eds.) A Reader on Punishment, Oxford University Press, Great Britain, 1994, pp. 152-54.
20A. von Hirsch, Censure and Sanctions, Clarendon Press, Great Britain, 1993, p. 108.
21H. L. A. Hart, Punishment and Responsibility, Oxford University Press, New York, 1968, p. 51.
22See, for instance, B. A. Hudson, “Mitigation for Socially Deprived Offenders”, in A. von Hirsch and A.
Ashworth (eds.), Principled Sentencing, Hart Publishing, Oxford, 1998.
23See R. A. Duff, Trials and Punishment, Cambridge University Press, Cambridge, 1986, pp. 191-99; and R.
A. Duff, “Retributive punishment - Ideals and actualities”, Israel Law Review, vol. 25, 1991, pp. 441-45.
24R. A. Duff, “Punishment, Communication, and Community”, in M. Matravers (ed.), Punishment and
Political Theory, Hart Publishing, Oxford, 1999, p. 67.
25See, for instance, W. Sadurski, “Social Justice and the Problem of Punishment”, Israel Law Review, vol.
25, 1991, p. 304 ff.
26R. Dagger, “Playing fair with punishment”, Ethics, vol. 103, 1993, p. 477.
27R. Sadurski, “Social justice and the problem of punishment”, Israel Law Review, vol. 25, 1991, p. 311.
28J. G. Murphy, “Three mistakes about retributivism”, Analysis, 1971, p. 166.
30It should be noticed that what we are talking about here is not social deprivation in general, but only a
precondition concerning the initial distribution of benefits and burdens.
33I have here interpreted the question of security as concerning the risk of becoming the victim of some sort
of wrong-doing. However, it might be suggested that a benefit has nothing to do with an eliminated or
decreased risk but rather with whether one actually would have become a victim had the law not existed. Only
those who would have been victims of wrong-doing had the law not existed have genuinely benefited from
the law. But clearly nor would this interpretation make it reasonable to presume an initial equal distribution of
benefits, because obviously it is not all persons who would have been victims of wrong-doing (or to the same
degree) had the law not existed.
34See R. W. Burgh, “Do the guilty deserve punishment?”, Journal of Philosophy, vol. 79, 1982. Of course, a
law against embezzlement might indirectly protect some other people; but still there is not reason to hold that
the benefit from this law is equally distributed.
37It might be suggested that a benefit should not be understood as the risks of actual wrongings one is
spared from by the law, but rather as the freedom within limits which the law provides. And since the law
in principle counts for everyone each member of the cooperation gains the same benefit (if I understand
Sadurski correctly it is something like this he has in mind). However, this suggestion is puzzling because
in that case it is no longer clear that it is at all benefiting to submit to the law in the first place. After all,
since the law formally limits one’s options it would seem that one would be formally better off by not
engaging in cooperation in the first place.
39M. Davis, “How to Make Punishment Fit the Crime?”, Ethics, vol. 93, 1983, p. 743.
40W. Sadurski, Giving Desert its Due, Riedel, Dordrecht, 1985, p. 226.
41G. Dworkin, “Is more choice better than less?”, Midwest Studies in Philosophy, vol. 7, 1982.
42W. Sadurski, “Social justice and the problem of punishment”, Israel Law Review, vol. 25, 1991, p. 317.
43Ibid. p. 317.
45See also D. Dolinko, “Some thoughts about retributivism”, Ethics, vol. 101, 1991, pp. 547-8.
46For instance, no (acceptable) policy would ever succeed in harmonizing everyone’s inclinations for
acting in certain ways.
47R. J. Arneson, “The principle of fairness and free-rider problems”, Ethics, vol. 92, 1982, p. 633.
48As far as I can see, this is the position which several fairness theorists advocate. For instance, Dagger
clearly speaks as if a social order constitutes a precondition for the legitimate application fo the theory.
See R. Dagger, “Playing fair with punishment”, Ethics, vol. 103, 1993, p. 477.
49J. G. Murphy, “Marxism and Retribution”, in J. Simmons, M. Cohen, J. Cohen and C. R. Beitz (eds.),
Punishment, Princeton University Press, United States of America, 1995, p. 7. Murphy’s pessimism is
based on considerations somewhat different from those I have presented.
.
CHAPTER 6
RELAXED PROPORTIONALITY
Discussion in previous chapters has revealed a number of challenges facing

proportionalism; challenges which, if they are genuine, in various ways question the
idea of proportionate punishment. However, when a philosophical theory - as well
as a theory in other fields of research - is faced with problems, there are different
ways in which adherents to the theory might react if they do not straightforwardly
give in and accept the criticism. One obvious possibility is to seek to show that the
criticism should be considered flawed or misplaced and that, therefore, on closer
inspection there are no genuine problems facing the theory after all. Another
approach is to admit that there is something to the criticism but to hold that the
outlined problems are not fatal since, for instance, they can be avoided by slight
modifications of non-essential parts of the theory under attack. During the foregoing
discussions, responses of the first sort have been considered, none of which - at least
so I have suggested - have succeeded in proving the problems insubstantial. The
time has now come to direct attention to the second sort of response, that is, to
consider revisionist accounts of proportionalism.
In the following, I shall start out by recapitulating the main conclusions
that have been reached so far. This is appropriate in light of the fact that we have
arrived at the final chapter. But, furthermore, by crystallizing what I regard as the
three cardinal problems of proportionalism, we will have a background for
evaluating the track which has been followed by revisionist proportionalists. The
summary will be accompanied with a few comments concerning the alleged
attractions of proportionalism. Subsequently, it will be considered whether some of
the theories that have been developed by modern exponents of hybrid theoretical
approaches to punishment - which allow for deviations from strict proportionality -
have succeeded in meeting the challenges. As will be indicated, I believe that if the
identified problems are genuine then they are not avoided by any of the suggested
modifications of the principle.
1. PROBLEMS AND PROMISES
The point of departure in the previous investigations was taken on the question of
how the proportionality principle could possibly be justified: why should a crime,
relative to one that is comparatively less serious, be more severely punished and, the
corollary, why should equally serious crimes be responded to with equally severe
punishments? That this is the first and main question the principle gives rise to is
certainly not surprising. Even though there are, for logical reasons, limits as to how
183
184 JESPER
THE ETHICS RYBERG
far one can reach in any plausible chain of justifications of an ethical theory, it is
indeed reasonable to demand some sort of theoretical underpinning of something so
special as a punishment distribution principle: what we should expect is a logical
relation to a more basic theory of the purpose of punishment. A number of theories
which, despite the fact that they all subscribe to the idea that punishment should be
perceived as just desert, diverge significantly from each other, have been held to
provide the requisite sort of justification. However, on a closer scrutiny it
surprisingly turned out that none of the theories, even if each ex hypothesi was
assumed morally convincing, succeeded in provided a foundation of
proportionalism.
Even though the flaws in the inferences from the more basic theories of
justice to the proportionalist distribution principle differed somewhat from one
theory to another, the theories all failed with regard to establishing the necessity of
such a distribution pattern. As will be recalled, the contention of the simple desert
theory, namely, that suffering, when it appropriately befalls wrongdoers, is
intrinsically valuable, did not entail proportionality. The fact that nonpunitive post-
crime suffering as well as pre-crime suffering may both nullify a post-crime desert
debt, blocked the step from simple desert to proportionality. A similar problem
confronted the fairness-theoretic justification and, as we saw further, the attempt to
establish proportionality by holding that only punitive burdens can restore the
benefit/burden balance that is disturbed by a perpetrator’s gain of an unfair
advantage, did not succeed either. Even though the expressionist theory of
punishment, by providing a background for distinguishing punitive and non-punitive
hardship, seemed in this respect more promising than rival retributive theories, the
theory, due to the fact that the object of desert was no longer suffering, faced other
problems. Either it did not convincingly answer the why-hard-treatment question or
it succeeded in answering this question only at the cost of pointing at goals, beyond
mere condemnation, which did not necessarily require proportionality.
If these considerations are correct they obviously do not constitute a
problem to the retributivist. All that is established is that if you are a retributivist of
the sort which today constitutes the dominant expositions of a retrospectively
oriented approach to punishment, then you should not support proportionality. But it
is equally obvious that this conclusion does constitute a problem for the
proportionalist. If the proportionality principle cannot be shown to rest on a morally
firm ground all we have is a freely dangling principle. Thus, proportionalists have
rightly recognized the demand for a justification; unfortunately they have at this
point not had an answer to hand.
The second major question which we have been investigating concerns the
more precise content of the proportionality requirement. One of the few points of
criticism that has been directed against the proportionality principle by early
commentators of retributivism has been that the principle presupposes something
which is not there, namely, some objective measure of appropriateness between
crime and punishment. Now, as we have seen, this criticism is premature. Except for
a few theorists defending versions of lex talionis, no recent proportionalists have
subscribed to the idea of some direct relation between crime and punishment.
However, this does not change the fact that there is still much to the idea behind the
RELAXED PROPORTIONALITY 185
criticism, namely, that we are not on a safe theoretical ground when it comes to the
question of what proportionality more precisely amounts to. The above discussion
focused on the three points at which proportionalism requires clarification: the
ranking of crimes, the ranking of punishments, and the anchoring of the scales.
As became clear, the idea of a crime ranking was complicated by the fact
that seriousness is held to be a function of both harm and culpability and, at least
according to some theorists, also of prior criminal record. The harm aspect faced
problems concerning detailed issues such as, for instance, how one should estimate
the harm of inchoate crimes or other crimes which do not involve directly
identifiable victims. Moreover, the question of remote harms, and thus of how much
of the harm that follows from a certain criminal act that should be attributed to this
act, raised a number of problems of a more general character. With regard to the
determination of culpability even more problems were identified. These problems
mainly consisted in providing a theoretical framework for determining degrees of
culpability. How different degrees of diminished responsibility should be compared
- no matter whether this was perceived from the perspective of the choice theory or
the character theory of responsibility - was far from clear. Similar problems existed
with regard to the mens rea of the criminal. And when it came to what was referred
to as “the challenge of absolute comparison”, namely, the question of how different
degrees of harm and culpability should be combined into judgements of crime
gravity, there were no morally well-grounded answers available. A challenge that
was certainly not mitigated if one in addition adhered to the view that respects to
prior criminal record should also figure in the assessment of criminal conduct.
Taken together, these problems were further enhanced by the fact of what
proportionalists typically assume is not merely an ordinal but rather some sort of
interval scaling of crimes.
At the outset a ranking of punishments in severity was theoretically less
demanding than the ranking of crimes. The answer as to what determines the
severity of a punishment is that this is the amount of suffering or inconvenience it
imposes on the punished. However, as was argued, this does not allow for a general
ranking of a sort which makes all objectively equal punishments - a year of
imprisonment or a certain fine - count as equally severe for all those who are
punished. Two related problems were brought forward. The first concerned the
Benthamite idea of sensibility: different persons may be very differently affected
even when they are made the victims of the same outer impact. Some are
psychologically and physically more vulnerable than others. The second concerned
the fact that one and the same punishment - say a fine or a prison term - may,
everything considered, result in very different outer impacts and thus affect people
very differently even if they are psychologically and physically equally “sensible”.
The upshot of these observations, combined with the claim that one cannot provide
a plausible conception of punishment severity that is based on something which is
inter-personally invariable, entailed that a traditional interpretation of
proportionality would have to be abandoned in favour of a principle accounting for
differences in impact. However, this was tantamount to a severe loss of simplicity in
application.
186 JESPER
THE ETHICS RYBERG
The problems were no less when attention was directed to the question of
how crimes and punishments, once they have been scaled, should be connected. As
was shown, the answer to the anchor-question depended on what kind of scales the
seriousness of crimes and the severity of punishments were held to be measurable
on. The answer favoured by most proportionalists was some degree of interval
matching of the scales. However, this sort of matching required at least two anchor-
points, exactly what a number of proportionalists have sought to establish. The most
complicated problem of this approach was to clarify how the limit of the maximum
punishment should be set. The idea that this limit should be drawn by reference to
human dignity did not provide a solid base for an anchoring. Moreover, even if it
were possible to specify maximum and minimum punishment levels, the procedure
still lacked reasons justifying the levels actually constituting the points at which the
two scales should be pegged. All in all, the conjecture did not provide much
anchoring guidance. Neither did von Hirsch’s proposal in this respect have much to
offer. What according to him should determine the upper cardinal limits of
punishment severity was not clear; also, the main idea of applying a decremental
strategy according to which one should, in accordance with the principle of
parsimony, make pro rata reductions of the punishment level until a floor,
determined by what will constitute a prudential incentive for compliance, has been
reached was on closer scrutiny equally unclear in its guidance. Thus, with the regard
to the question of how crimes should be punished the suggested proportionalist
answers have not provided theoretically satisfactory solutions.
Now, what does this wide range of technical problems concerning crime
and punishment scaling and not least the anchoring of the scales show us? Strictly
speaking, all that can be concluded is that it is far from clear what exactly it is that
the proportionality principle implies. This, of course, is important in respect of the
fact that certain punishment systems pretend, at least partly, to rely on a
proportionalist rationale. There is no background for claiming that this is the case.
However, at the purely philosophical level, the conclusion is also important. Though
it is an obvious non sequitur to conclude from the fact that satisfactory answers to a
problem have not been provided to the claim that such answers cannot be provided,
it certainly leaves the principle in a weak position if at crucial points it lacks the
requisite theoretical clarification. Moreover, the problem is that several of the
difficulties facing the principle are not only unanswered but that it is hard to see
how they could plausibly be answered. For the present we have no reason to believe
that they can be.
The final set of considerations to which attention was directed in the
previous analyses concerned the applicability of the proportionality constraint as a
guide for actual penal practice. The main problem that was taken up - after having
considered the challenge of self-defeatingness and the priority-problem raised by the
challenge - concerned the question of how justice in punishing relates to different
aspects of social justice. As we have seen, there was no single answer to this
question. Rather was the answer dependant upon the precise content of the different
retributivist positions on which proportionalism is held to be based. Amongst
expressionists, views were somewhat split. However, the contention was at least that
social injustice is something that should be accounted for in punishment distribution.
But it was not made clear how exactly this should be done. When it came to the
fairness theory, the implications of social justice considerations were more far-
reaching. It was argued that the justice presuppositions which the theory is
presuming - namely, an initial equal distribution of benefits and burdens - are not,
given the different interpretations of respectively a burden and a benefit, satisfied.
And more importantly, it did not seem plausible to believe that it ever could be
satisfied, even by effectuating political means specifically designed to this goal.
Furthermore, it was argued that, even if the required initial distribution pattern was
actually realized, this pattern would, due to the problems that led to the challenge of
self-defeatingness, be undermined by the very application of the punishment system
itself. The conclusion was that the fairness theory was either not applicable or that it
might account for initial inequalities by the distribution of punishments which,
however, would be tantamount to giving up proportionality. Thus, the overall
conclusion was that the proportionality principle did not admit of straightforward
application.
The three conclusions concerning respectively the justification of the
proportionality principle, the technical clarification of the principle and, finally, the
applicability problems constitute serious challenges to the principle. Moreover, they
are particularly noteworthy if seen in the light of the attractions which have been
ascribed to the principle and which have often been strongly emphasized in
comparison to the forward-looking punishment theories and practices in opposition
to which modern retributivism was developed.
A first alleged attraction of proportionalism is that the view provides a
principled way of handling sentencing. For most of the 20th century sentencing in
several countries (e.g. the USA and England) has been characterized by a wide
judicial discretion. Normal procedure has been for the legislature to establish
maximum punishment for crimes, leaving it to the courts to exercise discretion -
however, without the existence of any genuine standards for sentencing beneath the
legislative maxima. Furthermore, when there have been requirements for basic
reasons for sentencing, what has sometimes been advocated is what may be referred
to as the “smorgasbord” approach to sentencing, namely, that the aim of sentencing
1
is deterrence, reform, incapacitation, and retribution. Objectives which have often
been recounted as if no further comment would be necessary or perhaps supplied by
the device that all one has to do is to “balance” the objectives against each other
when they are in conflict. A device which is almost obscure when nothing further is
added. In contrast to such approaches which allow the court to choose fairly freely
within a very wide range of punishment options, the just deserts movement with its
emphasis of proportionality seemed to provide a principled way of resolving the
issue. However, if the above conclusions are correct this attraction should be taken
with a grain of salt. Firstly, without a genuine justification of proportionalism, what
we are left with is clearly not principled sentencing in the sense that the distribution
of punishments reflects some basic aim of punishing. And, secondly, both the
ranking of crimes and punishments as well as the anchoring of the scales have left
so many questions open that the reference to principled sentencing at best must be
characterized as premature.
188 JESPER
THE ETHICS RYBERG
A second-alleged
2
attraction of proportionalism is that it avoids
individualization. The thought - closely tight to the rehabilitationist ideal - that a
punishment must be tailored to the individual defendant because criminals, like
everyone else, are variable human beings with different histories and future
prospects, was formed as early as the end of the 19th century. This view led to a
rejection of fixed penalties in favour of the outlook that sentencing judges should be
given the widest leeway in order to succeed in suiting the punishment to the needs of
the individual perpetrator. However, a growing dissatisfaction with individualization
fertilized the ground for the revival of retributivism. On a proportionalist ground,
individuality would be maintained only with regard to the question of the
seriousness of each criminal’s conduct (and in fact this would be so only to a limited
extent because judgements of crime gravity would typically rely on standardization).
However, if the foregoing considerations are correct, this picture constitutes a
misrepresentation. The problem of delimitating punitive suffering and not least the
challenge of sensibility made it clear that the proportionalist - albeit his focus
naturally differs from that of the rehabilitationist - cannot plausibly avoid
individualization in the distribution of punishments.
A third-proclaimed attraction has been that proportionalism would lead to a
reduction in punishment levels. As Galligan says “the attractions of retribution are
easier to understand, for retribution is closely related to notions of desert, justice and
so .... provides
3
the criminal offender with greater protection against over-zealous
state action” . One aspect of this protection has usually been held to consist in the
fact that proportionalism implies lenience in punishing. The anti-draconian character
of proportionalism has, as we have seen, constituted a recurrent theme amongst
modern proportionalists. And several jurisdictions which have adopted
proportionality as the guiding rationale have succeeded, as has been emphasized,
either in reducing penalty levels (e.g.4 Finland) or at least in limiting penalty
increases (e.g. Sweden and Minnesota). Now, whether a deflation of punishment
levels is itself an attraction independently of a theory of punishment can perhaps be
discussed but even if this is the case it is, as we have seen, less obvious that
proportionalists are in this respect justified in claiming merits. The lack of a clear
theory of anchoring questions the genuineness of this attraction.
A final and repeatedly underscored attraction 5
of proportionalism is that the
principle succeeds in providing practical guidance. A theory, it has rightly been
held, must stand up in practice or fall. The “smorgasbord” approach to punishing
should therefore be rejected. But, furthermore, this has been emphasized as one of
the reasons why the utilitarian theory of punishment should be regarded as defective.
As Davis has emphatically put it “[t]he trouble with the utilitarian principle of 6
setting penalties is not so much that it leads us astray as that it leads us not at all” .
Without the requisite insight into the actual or probable consequences of setting
penal levels at one place rather than at another, we simply cannot do what
utilitarianism ideally tells us to do, but only “nervously shift from foot to foot”. In
contrast to such a “scarecrow” of a theory, proportionalism has been drawn forward
as a theory capable of delivering the requisite action-guidingness. However, the
theoretical lacunas in the ranking of crimes and punishments, as well as the lack of a
clear theory of anchoring, leave us without a solid theoretical ground from which
guidance can be extracted. And even if these problems are ignored the step from
theory to practice itself gave rise to additional problems. Thus, once again there are
reasons for some reluctance when it comes to applauding proportionalism.
Admittedly, proportionalists have not contended that the alleged attractions
are themselves sufficient to justify proportionalism. Thus, the more important
overall conclusion which I believe can be drawn on the ground of the previous
considerations is that the principle does not provide a plausible conjecture for
punishment distribution. If a principle is unjustified, theoretically flawed and not
easily applicable then it certainly provides reasons for looking for alternative
conjectures. Each of the three conclusions constitutes a serious problem to the
principle; together, I believe they are fatal. There may, however, at this point be
reasons for hesitation. Even if the overall conclusion is correct the strength of this
conclusion, as mentioned, is also dependant on the question of whether the different
problems could by slight changes in the principle be avoided. What we shall
consider now, therefore, are modifications of proportionalism.
2. MODIFIED PROPORTIONALISM
Thoughts on the justification of punishment have to a very large extent been a story
of two radically different and competing approaches: the forward-looking utilitarian
approach and the backward-looking retributive approach. As it has been the case in
many other areas of philosophy, where the theoretical field has been marked by a
deep division between opposing positions, an obvious and tempting response has
been to attempt to develop theories which manage to overcome the apparent
irreconcilability of the rival conjectures. A number of quite different theories
seeking to combine utilitarian and retributive considerations have been suggested.
The idea behind a compromise approach is obviously that one will thereby avoid the
difficulties with which each of the theories in their pure forms are confronted while
at the same time maintaining the insights of both. Whether the different attempts
have in this respect succeeded is not the question here. However, what is interesting
is whether the idea of a compromise has affected the punishment distribution
question in ways that are relevant with regard to the three basic sorts of problems
with which the genuine proportionalist is faced.
The proportionality principle was initially defined as requiring that a more
serious crime should, in relation to one less serious, be more severely punished and
that equally serious crimes should be equally severely punished, that is, as
prescribing ordinal proportionality and parity in punishing. In order to provide actual
guidance, the principle would have to be supplied by considerations as to which
punishments should be allocated to which crimes but the proportionalist was not,
according to the definition, locked to any particular view on cardinal proportionality.
Now, the question is whether one, by loosening the proportionality constraint in a
non-retributivist direction by in one way or another, letting in other objectives, may
avoid some of the outlined problems. It is in this connection worth recalling that
adherence to the proportionality principle does not commit one to rejecting the idea
that traditional utilitarian values should play a role in the allocation of punishment.
As we have seen, it has been suggested that the anchor problem should be resolved
partly on the ground of the principle of parsimony. And the idea that the amount of
punishment should be determined on the ground of the proportionality principle
could - and according to several theorists should - be combined with a view which,
at least to some extent, allows for punishment substitutability by letting the
punishment method be determined on standard utilitarian grounds. Such conjectures
are obviously fully consistent with the proportionality constraint and therefore do
not constitute a departure from the principle. What we shall consider are only
suggestions which
7
incorporate utilitarian goals in ways which amount to genuine
modifications. A number of such suggestions which to a greater or lesser extent
depart from the principle have been presented. I shall distinguish five sorts of
principles of which the former allow for minor deviations, while the latter more
radically deviate from traditional proportionalism.
(1) A first model that might be mentioned is to let in utilitarian goals only
as a way of conditioning proportionate punishment distribution. In the foregoing, it
has generally been held that the basic idea is that perpetrators should be punished for
their misdeeds and in a way that is proportionate to the crimes committed. However,
it might be contended that the proportionality principle should not be considered as
an implication from the retributivist idea that perpetrators ought to be punished but
rather that it places a constraint in so far as punishing is justified. Whether criminals
actually should be punished should be determined on utilitarian grounds. A theory of
this type would follow if one adopts the Hartian distinction between, on the one
hand, the question of the general justifying aim of a punishment system and, on the
other, the question of punishment distribution and, in addition, if one holds that the
former question should be answered on a purely utilitarian ground while the second
on a retributivist ground employing proportionality. In this version the suggestion
does not constitute a radical step away from proportionality: all that is said is that, if
there is a general utilitarian reason for punishing, one should punish proportionately.
Therefore, unsurprisingly, the outlined problems would remain intact. However, a
more noteworthy modification would occur if the utilitarian condition were not
placed on the general question of whether one should have a punishment system but
on each individual imposition of punishment. What this would amount to would be
that a criminal should be punished if this was better than his not being so, and in that
case he or she should receive a punishment proportionate to the crime. The result of
this sort of view would be that, in contrast to a system that merely imposes
proportionate punishment on criminals, some criminals might now be exempted
from punishment. Thus, this would obviously infringe the requirement of ordinal
proportionality and the parity condition.
(2) Another and perhaps more interesting position involving a departure
from strict proportionality is to hold that punishments should be allocated in
accordance with the proportionality principle but at the same time to allow for
exceptional deviations from this requirement. With such a view, punishment would
basically be proportionalist but, under exigent circumstances, utilitarian objectives
might trump proportionality. A notable variant of this conjecture has been developed
by Robinson. His point of departure is the reasonable claim that any punishment
system which pursues retributivist as well as utilitarian goals will, in order to avoid
moral arbitrariness, need to articulate a governing principle. The hybrid distribution
principle he suggests is that “[d]esert is to be given priority over the combined
utilitarian formulation, except where it causes an intolerable level of crime that the
utilitarian formulation could avoid. At this point utilitarian adjustments can be
made, but no utilitarian adjustment can be made 8
if it generates a formulation that
imposes an intolerable unjust punishment” . In other words, the idea is that
proportionality can be overridden. Upward departures will be permitted, however,
only exceptionally, when the stakes are high enough. Besides allowing for such
deviations, Robinson’s principle contains the further clause that gross deviations
should not be allowed.
Principles which like Robinson’s permit departures from proportionality,
no matter whether it is in the form of upward or downward deviations or both, may
obviously infringe an absolutist conception of ordinal proportionality. A more
serious crime may, under exceptional circumstances, be punished more leniently
than a crime which is less serious. Whether the parity requirement is infringed is not
clear. One possibility is to hold that a certain type of crime should be punished
disproportionately when there are strong utilitarian reasons in favour of it (e.g. for
9
crime preventive reasons drunk driving may be harshly punished) . This would
preserve parity. Another possibility is to accept deviations in individual cases (e.g.
10
where an offender is considered extremely dangerous) . This obviously violates the
parity requirement. Thus, in this respect a view allowing for exceptional departures
from proportionality may be given different interpretations.
(3) As we have seen in an earlier chapter, scales of crimes and punishments
may be linked in different ways. One might defend either a ratio, an interval or an
ordinal matching. Now, if one requires only an ordinal matching then this might
open up a further way in which utilitarian considerations could operate within a
proportionalist framework. All that ordinality requires is that the punishment for a
certain crime should be more severe than that for a less serious one. But this
requirement might well be satisfied while still leaving a leeway for utilitarian aims.
How a particular crime should be punished may be determined on utilitarian
grounds as long as the crime is not punished more severely than more serious crimes
and more leniently than less grave ones. However, as we have also seen, most
proportionalists - for good reasons - will not be satisfied with mere ordinality but
will require a stronger matching. If one requires strict intervality (or a ratio
matching) then obviously their will no longer be a leeway for such ulterior
objectives. Given the requisite anchor points, an interval matching will specify a
precise degree of punishment severity. But the idea that the spacing between the
seriousness of different crimes should be reflected in the punishments may be
satisfied if one does not require strict intervality. For instance, one might defend a
matching which allows one to maintain that the difference in seriousness between
one crime and another is larger than the difference in seriousness between two other
crimes, without being committed to such precise judgements as that the differences
between the former crimes is three times the difference between the latter. That is, it
may be the case that such differences themselves are only measurable on an ordinal
192 JESPER
THE ETHICS RYBERG
scale, which means that all one can say is that one seriousness difference is bigger
or smaller than (or equal to) another difference. Such a matching will leave a leeway
for how different crimes should be punished.
As far as I can see, it is this kind of matching which has been defended for
11
instance by Ten in what he calls a “thick version of the proportionality principle” .
What he believes is that such matching will lay down “fairly broad” limits of
permissible punishment for each crime, and that within those limits “utilitarian
considerations may operate freely in arriving at the punishment which will be most
12
effective in reducing crime” . Now, if one in this way allows for utilitarian
considerations but still maintains the requirement of parity in punishing, then the
suggestion remains fully consistent with proportionality. That is, we will have a
distribution principle which maintains the requirement of ordinal proportionality as
well as the parity condition and which therefore is only of a form which says that
once these requirements are satisfied other considerations should be taken into
account. However, if one abandons parity by allowing for different punishments for
the same crimes when this is supported on utilitarian grounds, and when the
requirements of an ordinal or a weak interval matching are maintained, then the
theory obviously constitutes a genuine modification of proportionality.
(4) Another suggestion that should be mentioned has been labeled “limiting
proportionalism” and constitutes to a larger extent a departure from strict
proportionality. What this theory implies is not that crimes should be scaled in terms
of seriousness and that they should in some way or another each be connected to a
punishment of a particular degree of severity, but rather that there is, for each crime,
a punishment range within the limits of which all punishments are regarded as
proportionate (or not disproportionate). These ranges are related to the crimes in
such a way that at least the upper limit is higher the more serious the crimes are. The
fine-tuning of the punishment severity within each zone should then be determined
on utilitarian (or other non-retributive) grounds. Limiting proportionality has been
advocated by Tonry and perhaps most notably by Morris, who has defended his
13
theory in various writings over several decades. Morris epitomizes his view by
saying that “[t]he upper and lower limits of ‘deserved’ punishment set the range in
14
which utilitarian values .... may properly fix the punishment to be imposed” . It is
important to note that, in contrast to the previous sort of modification according to
which there was within some sort of crime/punishment matching a leeway for other
considerations, the zones of deserved punishments on the limiting proportionalist
view may well be overlapping. This might be the case both if the view is presented
in a form in which there is only a relatively narrow range of deserved punishments
for a particular crime or, oppositely, if the view allows for a wide range of deserved
punishments between upper and lower limits. The standard version of limited
15
proportionality therefore does not only rule out the parity condition - which Morris
16
and Tonry regard as chimerical - but the requirement of ordinal proportionality is
derogated: for utilitarian reasons one crime may be punished more leniently than
another despite the fact that the latter is less serious. In that sense the theory
constitutes, in comparison to the previous modifications, a more radical step in the

direction of utilitarian punishment.
(5) The final proposal might be to take the idea of a zone of not unjust
punishments a step further by holding that what proportionalism requires is that no
one receives a punishment that exceeds an upper limit. The punishment, which is by
a standard proportionalist regarded as the one that should be imposed on a
perpetrator who has committed a crime of a certain degree of seriousness, is now
regarded only as a limit that should not be transgressed. A view along these lines is
apparently what Murphy has in mind when he contends that “ ... I am no longer
inclined to argue that justice requires that criminals suffer retributive punishment.
17
Rather I am now drawn to a weaker view that justice allows or permits this” . By
abandoning the positive aspect of proportionality - that a criminal should be
punished in a fitting manner - this idea of proportionality admits for the possibility
of letting in utilitarian goals to a large extent. The utilitarian punishing will be
restricted only by the upper limits of proportionate punishment. I shall refer to this
sort of limiting principle as “negative” proportionality.
The five outlined variants of theories departuring from proportionality do
not exhaust the number of possible modifications, for instance, they may be
combined in various ways. However, they do capture the modifications which have
been suggested by adherents of compromise approaches and, furthermore, they
seem to constitute the most obvious conjectures as to how utilitarian (or other)
considerations might be combined with a punishment structure which is basically
proportionalist. The question is whether the theories succeed in avoiding some of
the outlined problems facing the proportionality principle.
The basic question as to why we should at all punish in such a way that the
punishment is proportionate to the seriousness of the criminal conduct is unaffected
by letting utilitarian considerations influence the punishment distribution. Even the
most relaxed version of limiting proportionalism will have to explain why the upper
punishment limit is higher for a more serious crime than for one less grave. Since
the modifications do not concern the justification question this is a matter which still
has to be addressed. But what about the problems related to the technicalities of
proportionality?
A distribution theory of type (1) which makes a utilitarian justification a
necessary condition for the imposition of proportionate punishments is obviously
confronted with all the problems, concerning crime and punishment ranking and the
anchoring of the scales, which confront strict proportionality. The same is the case
both with regard to a type (3) conjecture, which lets utilitarian considerations
operate within the proportionalist framework, and with regard to the negative
proportionalist principle of type (5). With regard to a theory of type (2), which
allows for exceptional deviations from proportionality, neither does this conjecture
avoid the problems which must be answered by a genuine proportionalist. On the
contrary, the view seems to add further challenges of clarification. Firstly, it is
necessary to clarify what constitutes a sufficient reason to justify deviations.
18
An
exception-model obviously rejects the idea of fiat justitia, et ruat mundus , but the
more interesting question is what constitutes the minimal but sufficient ground for
deviations? In order not to turn into something very close to utilitarian punishment it
must be assumed that the countervailing utilitarian reasons must be quite weighty for
an override to be warranted. Robinson holds that19a departure is required when this is
necessary to prevent
20
“an intolerable crime level” . A suggestion which is obviously
far from precise. Secondly, a further clarification question obviously arises if one,
like Robinson, contends that the exception-model should be supplied by a further
justice constraint proscribing unconscionable departures from proportionality.
Should such a limit be determined relative to the weight of the countervailing
utilities? And when is this limit reached? The same sort of questions obviously
arises in the case of modifications of type (4) prescribing limiting proportionality.
Where should the desert limits for each crime be located? Theorists who have
defended limiting proportionality have at this crucial point not much to say. Morris
has indicated that the width of the range should be relatively narrow by emphasizing
that his view is primarily retributive and that “the proper punishment
21
to be imposed
should be strongly influenced by what the criminal has done” . However, in other
places he contends
22
that there is a quite broad range of not undeserved punishments
for a crime. The important thing is not that these statements point in different
directions but simply that, in order to be action-guiding, the question requires a more
precise answer and, hopefully, one that is not morally arbitrary. Thus, in sum, it
seems that none of the conjectures resolve the technical matters which have been
discussed; an upshot which is expected in the light of the fact that the introduction
of utilitarian goals within a proportionalist structure will add to, rather than reduce, the
theoretical complexity of a theory.
Then, how are things if we finally turn to the question concerning the
applicability of the principle? As we have seen, the moral significance of social
justice was dependant on the more precise content of the theories underpinning the
proportionality constraint. Nevertheless, the following seems to be the case. In so far
as social deprivation or some other aspects of social justice is something that should
be accounted for in the computation of the proportionate punishment, it was not
clear how this should be done. This fact is clearly not changed by resorting to any of
the modified versions of proportionalism. Even the theories of type (4) and (5),
which allow for a leeway for not unjust punishments, would still have to be able to
draw the line between what constitutes a proportionate and a disproportionate
punitive response to a certain crime. If, on the other hand, social justice plays a
preconditional role in the sense that certain presuppositions on the matter have to be
satisfied in order to make a theory of just punishing applicable in the first place and,
in addition, if these presuppositions are not satisfied then the fact that the different
modifications of proportionality let in forward-looking considerations does not
change the conclusion that the application of the theories remains unjustified.
Moreover, there is in this connection one further point worth noticing. As
we have seen, attempts to effectuate proportionate punishing will most likely result
in the infliction of punishments which are not genuinely proportionate. This gave
rise to the challenge of self-defeatingness which, however, did not turn out to
constitute a genuine problem to the traditional proportionalist. However, the
challenge turns out to have significant implications in relation to the most relaxed
versions of proportionality, namely, a theory of type (5). By holding that the

proportionate punishment of a criminal is not morally required but that it is not
morally permitted to impose a punishment which upwardly deviates in severity from
what is warranted by the gravity of a crime, the problem of punishment mistakes
becomes pertinent. If one imposes punishments on criminals one knows that some
will be punished disproportionately severely. But, since this is not permitted and
since it will not be the case if one abstains from the infliction of punishments, it
follows that in reality criminals should not be punished for their misdeeds. Thus,
negative proportionalism which merely places upper proportionalist limits to the
pursuit of utilitarian goals will, if the considerations presented in chapter 5 are
correct, never be lifted from the ground when we turn from ideal penal theory to
actual penal practice. The theory cannot consistently provide any penal guidance
(except from an extreme form of abolitionism). Whether adherents of the theory
would accept this far-reaching implication is questionable.
Though much more needs to be said in order to provide an adequate
assessment of the distribution principles which might be developed from a
compromise theoretic point of view I, nevertheless, believe that sufficient has now
been said to answer the objection that slight modifications of the proportionality
principle will be sufficient to render it immune to the problems which have been
presented. If one allows for departures from proportionality which are so
comprehensive as to render punishment distribution basically utilitarian then some
of the problems might be avoided. But, as long as the admission to letting in
utilitarian consideration are constrained in a way which renders the distribution
basically proportionalist in structure, the problems remain unimpaired. Basically,
this conclusion is not surprising. The discussion of proportionality has not been
focused on the traditional question as to whether it is at all reasonable to adopt a
retrospectively oriented view on punishing if, alternatively, punishment could be
used as an instrument to preventing the criminal conduct which all theorists would
prefer to live without. If this has been the pivotal theme, then obviously it might be
thought that any step away from pure retributivism in the direction of a forward-
looking theory would constitute an improvement. However, since the purpose has
rather been that of uncovering some of the problems facing proportionalism it is not
so surprising that these problems are intact as long as the distribution remains - even
if some concessions are made to the utilitarians - basically proportionalist.
3. CONCLUSION
As it has often been pointed out in works considering the just desert movement, the
revivification and development of the proportionality principle took place against a
background of an atmosphere of disillusionment and dissatisfaction with rival views
on punishment and the prevailing punishment practice. In this light the
proportionality principle appeared as an attractive conjecture. As pointed out by one
of the few critics of the movement, the principle apparently had something to offer
23
to many different participants in the ongoing controversy on punishment. To the
liberal, the principle would contribute to an elimination of the arbitrariness of
systems which depended upon wide discretion and disparity in sentencing; to the
right-wing, it guaranteed against the leniency of casting criminals into the care of
social workers rather than into the control of the punishment system; and to those
campaigning against excessive use of imprisonment, it promised a general reduction
in the punishment levels reserving imprisonment only for the most serious crimes.
However, more important than these somewhat rough generalizations is that the
principle apparently had two more basic attractions. In contrast to an approach
enunciating that punishment objectives covers retribution, deterrence, reform and
incapacitation, the proportionality principle constituted a theoretically attractive
alternative. And, in contrast to the utilitarian view, the principle seemed capable of
delivering genuine action-guidingness. Thus, the most significant alleged attractions
of proportionalism have consisted in the fact that the principle succeeded in
providing both a principled and workable background for sentencing. However, if
the arguments of this book are correct, these attractions are more apparent than real.
Once we start digging beneath the surface this promising picture falls apart. The
lack of a justification, the problems of theoretical clarification, and the problems
associated with the application of the principle strongly support this judgement.
Perhaps the real merit of proportionalism lies not in the particular way it has sought
to close the initially mentioned gap between philosophers and penologists, but
rather in the fact that the principle has conspicuously succeeded in directing
attention to such a morally important question as the distribution of punishment.
NOTES
1See, for instance, A. Ashworth, “Criminal Justice and Deserved Sentences”, in N. Lacey, Criminal
Justice, Oxford University Press, England, 1994.
2See, for instance, A. von Hirsch, Doing Justice, Hill and Wang, United States of America, 1976, section
4. Or M. E. Frankel, “The Quest for Equality in Sentencing”, Israel Law Review, vol. 25, 1991.
3D. J. Galligan, “The Return of Retribution in Penal Theory”, in C. F. Tapper (ed.), Crime, Proof and
Punishment, Butterworth & co., Great Britain, 1981, p. 145. Galligan believes that the protection also
consists in the fact that the criminal is not being used as an instrument in the promotion of other goals.
4See A. von Hirsch, “The Future of the Proportionate Sentence”, in T. G. Blomberg and S. Cohen (eds.),
Punishment and Social Control, Aldine De Gruyter, United States of America, 1995, p. 130.
5See, for instance, ibid. p. 125.
6M. Davis, “How to Make Punishment fit the Crime”, Ethics, vol. 93, 1983, p. 733.
7As mixed theorists typically do, I shall talk about letting utilitarian goals into the proportionalist
structure. But, obviously, it could just as well be other non-retributivist values. The important thing here
is not what the ulterior goals precisely consist in but rather how the proportionalist structure may be
relaxed.
8P. H. Robinson, “Hybrid Principles for the Distribution of Criminal Sanctions”, Northwestern University
Law Review, vol. 82, no. 1, 1987, p. 38.
9See A. von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993, p. 48.
10See, in this connection, A. E. Bottoms & R. Brownsword, “Dangerousness and Rights”, in J. W. Hinton
(ed.), Dangerousness: Problems of Assessment and Prediction, George Allen & Unwin, Great Britain,
1983.
11C. L. Ten, Crime, Guilt, and Punishment, Clarendon Press, Oxford, United States of America, 1987, p.
159.
12Ibid. p. 160. Ten basically defends a sort of Hartain compromise theory which says that punishing a
person is justified if and only if she is a criminal who has voluntarily violated a legitimate law and
punishing her is justified on utilitarian grounds (ibid. 78-80). To this he adds that punishing, under certain
conditions, is justified even when a punishment is not optimific and, further, that innocents may, under
certain conditions, be punished. The latter thought indicates that Ten’s suggestion might also be
considered an instance of modification (2).
13N. Morris, The Future of Imprisonment, The University of Chicago Press, United States of America,
1974; N. Morris, Madness and Criminal Law, University of Chicago Press, United States of America,
1982; N. Morris and M. Miller, “Predictions of Dangerousness, in Crime and Justice: An Annual Review
of Research, vol. 6, 1985; N. Morris and M. Tonry, Between Prison and Probabtion, Oxford University
Press, New York, 1990; M. Tonry, “Proportionality, interchangeability, and Intermediate Punishments”,
in R. Dobash, R. A. Duff and S. Marshall (eds.), Penal theory and penal practice, Manchester University
Press, England, 1993.
14N. Morris,” Incapacitation within limits”, A. von Hirsch & A. Ashworth, Principled Sentencing, Hart
Publishing, Oxford, 1998, p. 110.
15Obviously, one might within limiting proportionality hold that parity should still be maintained;
however, to my knowledge no one has defended this version of the theory.
16See N. Morris and M. Tonry, Between Prison and Probation, Oxford University Press, United States of
America, 1990, p. 94.
17J. Murphy and J. Hampton, Foregiveness and Mercy, Cambridge University Press, New York, 1988, p.
180.
18”Let the world perish so long that justice be done”.
19P. H. Robinson, “Hybrid Principles for the Distribution of Criminal Sanctions”, Northwestern
University Law Review, vol. 82, 1988, p. 38.
20See the discussion in A. von Hirsch, Censure and Sanctions, Clarendon Press, Oxford, 1993, chap. 6.
198 JESPER
THE ETHICS RYBERG
21N. Morris, “Incapacitation within limits”, in A. von Hirsch & A. Ashworth (eds.), Principled
Sentencing, Hart Publishing, Oxford, Great Britain, 1998, p. 110.
22N. Morris, Madness and the Criminal Law, Chicago University Press, United States of America, 1982,
p. 151.
23B. Hudson, Justice through Punishment, Macmillan Education, Hong Kong, 1987, p. 37 .
BIBLIOGRAPHY
Acton, H. B. (ed.), The Philosophy of Punishment, St. Martin’s Press, Great Britain,
1969.
Adams, K., “Adjusting to Prison Life”, Crime and Justice, vol. 16, 1992.
Alexander, L., “The Doomsday Machine: Proportionality, Punishment and

Prevention”, The Monist, vol. 63, no. 2, 1980.
Alexander, L., “Retributivism and the inadvertent punishment of the innocent”, Law
and Philosophy, vol. 2, 1983.
Alexander, L., “Consent, Punishment and Proportionality”, Philosophy and Public

Affairs, vol. 15, no. 2, 1986.
Alexander, L., “Self-Defense, Punishment, and Proportionality”, Law and

Anscombe, G. E., Intention, Basil Blackwell, Oxford, 1963.
Arenella, P., “Character, Choice and Moral Agency: The Revelance of Character to
our Moral Culpability Judgments, Social Philosophy and Policy, vol. 7, 1990.
Armstrong, K. G., “The Retributivist Hits Back”, in Acton, H. B. (ed.), 1969.
Ashworth, A., Sentencing and Penal Policy, Weidenfeld and Nicolson, London,
1983.
Ashworth, A., “Sharpening the subjective element in criminal liability”, in Duff, R.

A. and Simmonds, N. (eds.), 1984.
Ashworth, A., “The elasticity of mens rea”, in Tapper, C. F. H. (ed.), 1991.
Ashworth, A., “Taking the consequences”, in Shute, J., Gardner, J. and Horder, J.
(eds.), 1993
Ashworth, A., “Criminal Justice and Deserved Sentencing”, in Lacey, N. (ed.),

1994.
Ashworth, A., Principles of Criminal Law, Clarendon Press, Oxford, 1995.
199
Ashworth, A., Sentencing and Criminal Justice, Butterworths, Great Britain, 1995.
Ashworth, A., “Four Techniques for Reducing Sentence Disparity”, in von Hirsch,
A. and Ashworth, A (eds.), 1998.
Ashworth, A and von Hirsch, A., “Desert and the Three Rs”, in von Hirsch, A. and
Ashworth, A. (eds.), 1998.
Ashworth, A.and Wasik, M. (eds.), Fundamentals of Sentencing Theory, Clarendon

Bagaric, M., “Double Punishment and Punishing Character: The Unfairness of Prior
Convictions”, Criminal Justice Ethics, Winter, 2000.
Bagaric, M., Punishment & Sentencing, Cavendish Publishing, Great Britain, 2001.
Baldwin, T., “Punishment, Communication, and Resentment”, Matravers, M. (ed.),

1999.
Barnett, R. E., “Restitution: A New Paradigm of Criminal Justice”, Ethics, vol. 87,
1977.
Barnett, R. E., “The Justice of Restitution”, The American Journal of

Jurisprudence, vol. 25, 1980.
Bayles, M., “Character, Purpose and Criminal Responsibility”, Law and Philosophy,
vol. 1, 1982.
Beccaria, C., On Crimes and Punishment, in Manzoni, A. (ed.), The Column of

Imfany, Oxford University Press, Oxford, 1964.
Bedau, H. A., Concessions to Retribution in Punishment”, in Cederblom, J. B. and

Blizek, W. L. (eds.), 1977.
Bedau, H. A., “Retribution and the Theory of Punishment”, Journal of Philosophy,

vol. 75, 1978.
Bedau, H. A., “Classification-based Sentencing: some Conceptual and Ethical

Problems”, New England Journal of Criminal and Civil Confinement, vol. 10,
1984.
Belliotti, R. A., “Gloom and Doom: Executing the Eighth Amendment”, The
International Journal of Applied Philosophy, vol. 3, 1986/7.
Benn, S. I., and Peters, R. S., Social Principles and the Democratic State, George
Allen & Unwin LTD, London, 1959.
BIBLIOGRAPHY 201
Bennett, J., The Act Itself, Clarendon Press, Oxford, 1995.
Bentham, J., The Principles of Moral and Legislation, Prometheus Books, United
States of America, 1988.
Bentham, J., Principle of Penal Law, Works, Bowring, J. (ed.), Willliam Trait,
Edinburgh.
Blomberg, T. G. and Cohen, S., Punishment and Social Control, Aldine De Gruyter,
new York, 1995.
Bottoms, A. E. and Brownsword, R., “Dangerousness and Right”, in Hinton, J. W.

(ed.), 1983.
Bottoms, A., “Five Puzzles in von Hirsch’s Theory of Punishment”, in Ashworth,

A. and Wasik, M. (eds.), 1998.
Brandt, R., Ethical Theory, Prentice-Hall, Englewood Cliffs, 1959.
Braithwaite, J. and Pettit, P., Not Just Deserts, Oxford University Press, Great
Britain, 1998.
Brownless, I., “Hanging, judges and wayward mechanics: reply to Michal Tonry”,
in Duff, R. A., Marshall, S., Dobash, R. E., Dobash, R. P., (eds.), 1995.
Buchner, D., “Scale of Sentencing Severity”, The Journal of Criminal Law and
Criminolgy, vol. 70, 1979.
Burgh, R. W., “Do the Guilty Deserve Punishment?”, Journal of Philosophy, vol.
79, 1982.
Byrne J. M., Lurigio, A. J. and Petersilia, J. (eds.), Smart Sentencing, Sage

Publications, United States of America, 1992.
Calvert, B., “Retribution, Arbitrariness and the Death Penalty, Journal of Social
Card, C., “Retributive Penal Liability”, American Philosophical Quarterly, vol. 7

(monograph), 1973.
Cederblom J. B. and Blizek, W. L. (eds.), Justice and Punishment, Ballinger

Publishing Company, Massachusetts, 1977.
Clark, M., “The Sanctions of the Criminal Law”, Proceedings of the Aristotelian
Society, vol. XCVII, 1997.
Clark, R. S., “The United States Federal Sentencing Guidelines: Who is Watching
the Makers of Grids”, Criminal Law Forum, vol. 11, 2000.
Clarkson, C. and Morgan, R. (eds.), The Politics of Sentencing Reform, Clarendon

Cooper, D. E., “Hegel’s theory of punishment”, in Pelczynski, Z. A. (ed.), Hegel’s

Political Philosophy, Cambridge University Press, Great Britain, 1971.
Cottingham, J. G., “Varieties of Retribution”, Philosophical Quarterly, vol. 29,

1979.
Cottingham, J. G., “Punishment and Respect for Persons”, in Stewart, M. A. (ed.),

1979.
Cragg, W., Retributivism and its Critics, Franz Steiner Verlag, Stuttgart, 1992.
Dagger, R., “Playing Fair with Punishment”, Ethics, vol. 103, 1993.
Davis, L. H., “They Deserve to Suffer”, Analysis, vol. 32, 1971-2.
Davis, M., “How to Make Punishment Fit the Crime”, Ethics, vol. 93, 1983.
Davis, M., “Death Penalty, Civilization, and Inhumaneness”, Theory and Practice,
vol. 16, 1990.
Davis, M., “Punishment as Language: Misleading Analogy for Desert Theorists”,

Law and Philosophy, vol. 10, 1991.
Davis, M., To Make the Punishment Fit the Crime, Westview Press, United States
of America, 1992.
Davis, M., “Criminal Desert and Unfair Advantage”, Law and Philosophy, vol. 12,
1993.
Dolinko, D., “Some Thoughts About Retributivism”, Ethics, vol. 101, 1991.
Dolinko, D., “Mismeasuring “Unfair Advantage”: A Response to Michael Davis”,

Law and Philosophy, vol. 13, 1994.
Dolinko, D., “Retributivism, Consequentialism, and the Intrinsic Goodness of

Punishment”, Law and Philosophy, vol. 16, 1997.
BIBLIOGRAPHY 203
Duff, R. A. and Simmonds, N. (eds.), Philosophy and the Criminal Law, Franz
Steiner Verlag, Wiesbaden, 1984.
Duff, R. A., Trials and Punishment, Cambridge University Press, Cambridge, 1986.
Duff, R. A., “A Reply to Bickenbach”, Canadian Journal of Philosophy, vol. 18,

1988.
Duff, R. A., Intention, Agency and Criminal Liability, Blackwell, Great Britain,
1990.
Duff, R. A., “Punishment, Expression and Penance”, in Jung, H, Müller-Dietz, H.,

Neumann, U. (eds.), Recht und Moral, Baden-Baden, 1991.
Duff, R. A., “Alternatives to Punishment - or Alternative Punishments?”, in Cragg,

W. (ed.), 1992.
Duff, R. A., “Choice, Character, and Criminal Liability”, Law and Philosophy, vol.
12, 1993.
Duff, R. A. and Garland, D., Punishment, Oxford University Press, United States,
1994.
Duff, R. A., Marshall, S., Dobash, R. E., Dobash (eds.), Penal Theory and Penal
Practice, Manchester University Press, Great Britain, 1995.
Duff, R. A., “Desert and Penance”, in von Hirsch, A. and Ashworth, A. (eds.),
1998.
Duff, R. A., Philosophy and Criminal Law, Cambridge University Press, United
Duff, R. A,, “Punishment, Communication, and Community”, in Matravers, M.

(ed.), 1999.
Dunbar, I. and Langdon, A., Tough Justice, Blackstone Press Limited, Great Britain,
1998.
Durham III, A. M., “Justice in Sentencing: The Role of Prior Record of Criminal
Involvement”, The Journal of Criminal Law and Criminology, vol. 78, no. 3,
1987.
Ellis, A., “Punishment and the Principle of Fair Play”, Utilitas, vol. 9, 1997.
Erikson, M. L. and Gibbs, J. P., “On the Perceived Severit of Legal Penalties”, The
Journal of Criminal Law and Criminology, vol. 70, no.1, 1979.
Ewing, A. C., The Morality of Punishment, Patterson Smith, Montclair, 1970 (org.
1929).
Ezorsky, G., Philosophical Perspectives on Punishment, State University of New

York Press, Albany, 1972.
Freeman, M. D. A., “Retributivism and the Death Sentence”, in Stewart, M. A.

(ed.), 1979.
Feinberg, J., “The Expressive Function of Punishment”, in Doing and Deserving,

Princeton University Press, 1970.
Feinberg, J., Harm to Others, Oxford University Press, New York, 1984.
Finnis, J., “The Restoration of Retribution”, Analysis, vol. 32, 1971-2.
Finnis, J., Natural Law and Natural Rights, Clarendon Press, Oxford, 1980.
Fletcher, G., Rethinking Criminal Law, Little, Brown, Boston, 1978.
Fletcher, G., “The Recidivist Premium”, Criminal Justice Ethics, vol. 1, 1982.
Frankel, M. E., “Lawlessness in Sentencing”, Cincinnati Law Review, vol. 41,

1972; reprinted in an excerpted version in von Hirsch, A. and Ashworth, A.
(eds.), 1998.
Frankel, M. E., “The Quest for Equality in Sentencing”, Israel Law Review, vol. 25,
1991.
Frase, R. S., “Sentencing Principles in Theory and Practice”, Crime and Justice, vol.
22, 1997.
Freeman, M. D. A., “Retributivism and the Death Sentence”, in Stewart, M. A.

(ed.), 1979.
Frey, R. G., and Morris, C. W. (eds.), Liability and Responsibility, Cambridge

Fried, C., An Anatomy of Values, Harvard University Press, 1970.
Galligan, D. J. “Guidelines and Just Deserts: A Critique of Recent Trends in

Sentencing Reform”, Criminal Law Review, 1981.
Galligan, D. J., “The Return of Retribution in Penal Theory”, in Tapper, C. F. H.

(ed.), 1981.
BIBLIOGRAPHY 205
Gardner, J., “Crime: in Proportion and in Perspective”, in Ashworth, A. and Wasik,

M. (eds.), 1998.
Gerstein, R. S., “Capital Punishment - Cruel and Unusual”?: A Retributivist

Response”, Ethics, vol. 84, 1974.
Gibbs, M. R., “Eighth Amendment - Narrow proportionality Requirement Preserves

Defence to Legislative Judgment”, The Journal of Criminal Law and
Glover, J., Causing Death and Saving Lives, Penquin Books, 1990.
Golash, D., “The Retributive Paradox”, Analysis, vol. 54, 1994.
Goldman, A. H., “The Paradox of Punishment”, in Simmons, A. J., Cohen, M.,

Cohen, J., Beitz, C. R. (eds.), 1995.
Gotesky, R. and Laszlo, E., Human Dignity. This Century and the Next, Gordon and
Breach, United States of America, 1970.
Griset, P. L., Determinate Sentencing. The Promise and Reality of Retributive

Justice, State University of New York Press, United States of America, 1991.
Gross, H., A Theory of Criminal Justice, Oxford University Press, New York, 1979.
Gross, H. and Ashworth, A., The English Sentencing System, Butterworths, London,
1981.
Gross, H., “Culpability and Desert”, in Duff, R. A. and Simmonds, N. (eds.), 1984.
Gross, H. and Harrison, R. (eds.), Jurisprudence, Clarendon Press, Oxford, 1992.
Hampton, J., “A New Theory of Retribution”, in Frey, R. G. and Morris, C. W.

(eds.), 1991.
Hampton, J., “An Expressive Theory of Retribution”, in Cragg, W. (ed.), 1992.
Hampton, J.,”The Moral Education Theory of Punishment”, in Simmons, J., Cohen,

M., Cohen, J., Beitz, C.R. (eds.), 1995.
Harlow, R. E., Darley, J. M. and Robinson, H. P. , “The Severity of Intermediate

Penal Sanctions: A Psychophysical Scaling Approach for Obtaining Community
Perceptions”, Journal of Quantitative Criminology, vol. 11, 1995.
Hart, H. L. A., “Are There Any Natural Rights?”, in Quinton, A. (ed.), Political
Philosophy, Clarendon Press, oxford, 1967.
Hart, H. L. A., Punishment and Responsibility, Oxford University Press, New York,
1968.
Hegel, F., Philosophy of Right, translated by T. M. Knox, Oxford University Press,

United States of America, 1967.
Hestevold, S., “Disjuctive Desert”, American Philosophical Quarterly, vol. 20,

1983.
Hill, T. E., Dignity and Practical Reason in Kant’s Moral Philosophy, Cornell
Hinton, J. W., Dangerousness: Problems of Assessment and Prediction, Allen &

Unwin, London, 1983.
Holmgren, M., “Justifying Desert Claims: Desert and Opportunity”, Journal of

Value Inquiry, vol. 20, 1986.
Holmgren, M., “The Backward-looking component of weak retributivism”, Journal

of Value Inquiry, vol. 23, 1989.
Honderich, T., Punishment, Polity Press, Great Britain, 1989.
Hudson, B., Justice through Punishment, Macmillan Education, Hong Kong, 1987.
Husak, D., “Already Punished Enough”, Philosophical Topics, vol. 18, 1990.
Husak, D., “Why Punish the Deserving?”, Nous, vol. 26, 1992.
Husak, D., “Is Drunk Driving a Serious Offence?”, Philosophy and Public Affairs,
vol. 23, 1994.
Jacobs, J., “Luck and Retribution”, Philosophy, vol. 74, 1999.
Jareborg, N., Essays in Criminal Law, Iustus, Uppsala. 1988.
Jareborg, N., “Why Bulk Discounts in Multiple Offence Sentencing?”, in Ashworth,

A. and Wasik, M. (eds.), 1998.
Johnson, C. A. H., “Entitled to Clemency: Mercy in the Criminal Law”, Law and
BIBLIOGRAPHY 207
Kadish, S. H., Blame and Punishment, Macmillan Publishing Company, New York,
1987.
Kahan, D. M., “Punishment Incommensurability”, Buffalo Criminal Law Review,

vol. 1, no. 2, 1998.
Kellogg, F., “From Retribution to Desert”, Criminology, vol. 15, 1977.
Kemp, P., Rendtorff, J. and Johansen, N. M., (eds.), Bioethics and Biolaw, Rhodos,
Copenhagen, 2000.
Kershnar, S., “George Sher’s Theory of Deserved Punishment, and the Victimized
Wrongdoer”, Social Theory and Practice, vol. 23, 1997.
Kershnar, S., “Objections to the Systematic Imposition of Punitive Torture”,

International Journal of Applied Philosophy, vol. 13, 1999.
Khatchadourian, H., “Just Ethical Punishment”, Journal of Value Inquiry, vol. 16,
no. 3, 1982.
Kleinig, J., “The Concept of Desert”, American Philosophical Quarterly, vol. 8, no.
1, 1971.
Kleinig, J., Punishment and Desert, Martinus Nijhoff, The Hague, 1973.
Kleinig, J., “The Hardness of Hard Treatment”, in Ashworth, A. and Wasik, M.

(eds.), 1998.
Kolnai, A., “Dignity”, Philosophy, vol. 51, 1976.
Kuhse, H., “Is There a Tension Between Autonomy and Dignity?”, in Kemp, P.,
Rendtorff, J. and Johansen, N. M. (eds.), 2000.
Lacey, N., State Punishment, Routledge, Great Britain, 1988.
Lacey, N., Criminal Justice, Oxford University Press, Oxford, 1994.
Lempert, R. O., “Desert and Deterrence: An Assessment of the Moal Bases of the
Case for Capital Punishment”, Michigan Law Review, vol. 79, no. 6, 1981.
Lucas, J. R., On Justice, Clarendon Press, Oxford, 1990.
Lyons, W., “Deterrent theory and Punishment of the Innocent”, Ethics, vol. 84,
1974.
Mabbott, J. D., “Punishment”, in Acton, H. B. (ed.), 1969.
Mackie, J. L., “Morality and the Retributive Emotions”, Criminal Justice Ethics,
Winter/Spring, 1982,
Maclagan, W. G., “Punishment and Retribution”, Philosophy, vol. 14, 1939.
Martin, R., “Justifying Punishment and the Problem of the Innocent”, Journal of
Social Philosophy, vol. 20, 1989.
Matravers, M., Punishment and Political Theory, Hart Publishing, Oxford, 1999.
Matravers, M.,”’What to Say?’: The Communicative Element in Punishment and

Moral Theory”, in Matravers, M. (ed.), 1999.
McCloskey, H. J., “The Complexity of the Concepts of Punishment”, Philosophy,

vol. 37, 1962.
McCloeksy, H. J., “A Note on Utilitarian Punishment”, Mind, vol. 72, 1963.
McCloskey, H. J., “A Non-utilitarian Approach to Punishment”, Inquiry, vol. 8,

1965.
McMahan, J., “Death and the Value of Life”, Ethics, vol. 99, 1988.
Meyer, M. J., “Dignity , Rights, and Self-Control”, Ethics, vol. 99, 1989.
Meyer, M. J., “Dignity, Death and Modern Virtue”, American Philosophical

Quarterly, vol. 32, 1995.
Miller, F. G., “Restitution and Punishment : A Reply to Barnett”, Ethics, vol. 88,
1978.
Mitias, A. H., “Is Retributivism Inconsistent Without Lex Talionis?”, Revista

Internazionale di Filosofia del Diritto, vol. 60, 1983.
Moberly, W., The Ethics of Punishment, Faber and Faber, London, 1968.
Moore, M. H., Estrich, S. R., McGillis, D., Spelman, W., Dangerous Offenders,
Harvard University Press, United States of America, 1984.
Moore, M. S., “Choice, Character, and Excuse”, Social Philosophy and Policy, vol.
7, 1990.
Moore, M. S., “Justifying Retributivism”, Israel Law Review, vol. 27, 1993.
BIBLIOGRAPHY 209
Moore, M. S., Placing Blame, Clarendon Press, Oxford, 1997.
Morison, J., “Hart’s Excuse: Problems with a Compromise Theory of Punishment”,

in Leith, P. and Ingram, P. (eds.), The Jurisprudence of Orthodoxy, Routledge,
London, 1988.
Morris, H., “Persons and Punishment”, Monist, vol. 52, 1968.
Morris, N., The Future of Imprisonment, The University of Chicago Press, United
Morris, N., Madness and Criminal Law, The University of Chicago Press, United
Morris, N. and Miller, M., “Predictions of Dangerousness”, Crime and Justice, vol.
6, 1985.
Morris, N. and Tonry, M., Between Prison and Probation, Oxford University Press,
New York, 1990.
Morris, N., “Incapacitation within limits”, in von Hirsch, A. and Ashworth, A.

(eds.), 1998.
Mundle, C. W. K., “Punishment and Desert”, in Acton, H. B. (ed.), 1969.
Murphy, J. G., “Three Mistakes about Retributivism”, Analysis, 1971.
Murphy, J. G., “Marxism and Retribution”, Philosophy and Public Affairs, vol. 2,
1973.
Murphy, J. G., Retribution, Justice, and Therapy, Reidel, Dordrecht, 1979.
Murphy, J. G., “Cruel and Unusual Punishments”, in Murphy, J. G., 1979.
Murphy, J. G., Forgiveness and Mercy, Cambridge University Press, Cambridge,

1989.
Nagel, T., “Death”, in Mortal Questions, Cambridge University Press, Cambridge,

1978.
Nagel. T., The View from Nowhere, Oxford University Press, New York, 1986.
Narayan, U., “Appropriate Responses and Preventive Benefits: Justifying Censure

and Hard Treatment in Legal Punishment”, Oxford Journal of Legal Studies, vol.
13, no. 2, 1993.
Nino, C. S., “Does Consent Override Proportionality?”, Philosophy and Public

Affairs, vol. 15, no. 2, 1986.
Norcross, A., “Intending and Foreseeing Death: Potholes on the Road to Hell”,
Southwest Philosophy Review, vol. 15, no. 1, 1999.
Nozick, R., Philosophical Explanations, Harvard University Press, Cambridge,

1981.
Oldenquist, A., “An Explanation of Retribution”, The Journal of Philosophy, 1988.
Parent, W. A., “The Whole Life View of Criminal Desert”, Ethics, vol. 86, 1975.
Philips, M., “The Inevitability of Punishing the Innocent”, Philosophical Studies,

vol. 48, 1985.
Pincoff, E. L., “Are Questions of Desert Decidable?”, in Cederblom, J. B. and

Blizek, W. L. (eds.), 1977.
Primoratz, I., “Utilitarianism and Self-sacrifice of the Innocent”, Analysis, vol. 38,
1978.
Primoratz, I., “On some arguments against the retributive theory of punishment”,
Revista Internazionale di Filosofia del Diretto, vol 56., 1979
Primoratz, I., “Utilitarianism and Punishment of the Innocent”, Rivista

Internazionale di Filosofia del Diritto, 1980.
Primoratz, I. “On Retributivism and the Lex Talionis”, Rivista Internazionale di

Filosofia del Diritto, vol. 61, 1984.
Primoratz, I., Justifying Legal Punishment, Humanities Press International, London,

1989.
Primoratz, I, “Punishment as Language”, Philosophy, vol. 64, 1989.
Quinton, M., “On Punishment”, in Acton, H. B. (ed.), 1969.
Rachels, J., “Punishment and Desert”, in LaFollette, H. (ed.), Ethics in Practice,

Blackwell, Oxford, 1997.
Rainbolt, G.,”Mercy: In Defence of Caprice”, Nous, vol. 31, 1997.
Rawls, J., “Two Concepts of Rules”, Philosophical Review, vol. 64, 1955.
BIBLIOGRAPHY 211
Reiman, J. “The Death Penalty, Deterrence, and Horribleness: Reply to Davis”,

Theory and Practice, vol. 16, 1990.
Reiman, J. and van den Haag, E., “On the Common Saying that it is Better that Ten
Guilty Persons Escape than that One Innocent Suffer: Pro and Con”, Social
Philosophy and Policy, vol. 7, no. 2, 1990.
Reiman, J., “Justice, Civilization, and the Death Penalty: Answering van den Haag”,
in Simmons et al (eds.), 1995.
Roark, A. P., “Retribution, the Death Penalty, and the Limits of Human Judgment”,
International Journal of Applied Philosophy, vol. 13, no. 1, 1999.
Robinson, P. H., “A Sentencing System for the 21st Century?”, Texas Law Review,
vol. 66, no. 1, 1987.
Robinson, P. H., “Hybrid Principles for the Distribution of Criminal Sanction”,

Northwestern University Law Review, vol. 82, 1987.
Robinson, P. H., “Desert, crime control, disparity, and units of punishment”, in

Duff, R. A., Marshall, S., Dobash R. E., Dobash, R. P. (eds.), 1995.
Robinson, P. H. and Darley, J. M., Justice, Liability & Blame, Westview Press,
United States of America, 1995.
Rosen, F., “Utilitarianism and the Punishment of the Innocent: The Origins of a
False Doctrine”, Utilitas, vol. 9, 1997.
Rothman, D. J., “More of the Same: American Criminal Justice Policies in the
1990s”, in Blomberg, T. G. and Cohen, S. (eds.), 1995.
Ryberg, J., “The Expressionist Theory of Punishment and the Problem of Fallible
Communication”, Readings in Philosophy and Science Studies, vol. 1, 2001.
Ryberg, J., “Recidivism, Multiple Offending and Legal Justice”, Danish Yearbook
of Philosophy, vol. 36, 2001
Sadurski, W., Giving Desert Its Due, Reidel, Dordrecht, 1985.
Sadurski, W., “Social Justice and the Problem of Punishment”, Israel Law Review,
vol. 25, 1991.
Schedler, G., “Can Retributivists Support Legal Punishment?”, The Monist, vol. 63,
no. 2, 1980.
Schoeman, F., Responsibility, Character, and the Emotions, Cambridge University

Press, United States of America, 1987.
Sebba, L., “Some Explorations in Scaling of Penalties”, Journal of Research in

Crime and Delinquency, vol. 15, 1978.
Sebba, L. and Nathan, G., “Further Explorations in the Scaling of Penalties”, The
British Journal of Criminology, vol. 23, 1984.
Sellin, T. and Wolfgang, M. E., The Measurement of Delinquency, Patterson Smith,

New Jersey, 1978.
Scheid, D. E., “Kant’s Retributivism”, Ethics, vol. 93, 1983.
Scheid, D. E., “Davis and the Unfair-Advantage Theory of Punishment: A

Critique”, Philosophical Topics, vol. 18, 1990.
Scheid, D. E., “Davis, Unfair Advantage Theory, and Criminal Desert”, Law and
Scheid, D. E., “Constructing a Theory of Punishment, Desert, and the Distribution

of Punishments”, The Canadian Journal of Law and Jurisprudence, vol. 10,
1997.
Sher, G., Desert, Princeton University Press, Princeton, 1987.
Shute, S., Gardner, J. and Horder, J. (eds.), Action and Value in Criminal Law,
Clarendon Press, Oxford, 1993.
Sidgwick, H., The Methods of Ethics, Dover Publication (republication of the 1907
edition), New York, 1966.
Simester, A. P., “Why Distinguish Intention from Foresight?”, in Simester, A. P.

and Smith, A. T. (eds.), 1996.
Simester, A. P. and smith, A. T. (eds.), Harm and Culpability, Clarendon Press,

Oxford, 1996.
Simmons, A. J., Cohen, M., Cohen, J., Beitz, C. H. (eds.), Punishment, Princeton
Singer, G., Just Deserts, Ballinger Publishing Company, United States of America,
1979.
Skillen, A. J., “How to Say Things with Walls”, Philosophy, vol. 55, 1980.
BIBLIOGRAPHY 213
Smart. J. J. C., and Williams, B., Utilitarianism: For and Against, Cambridge
University Press, Cambridge, 1973.
Smart, J. J. C., “Utilitarianism and Punishment”, Isreal Law Review, vol. 25, 1991.
Smilansky, S., “Utilitarianism and the ‘Punishment’ of the Innocent: The General
Problem”, Analysis, vol. 50, 1990.
Snare, A., Beware of Punishment, Scandinavian Studies in Criminology, vol. 14,

1995.
Sparks, R., Genn, H and Dodd, D. J., Surveying Victims, Wiley, 1977.
Spiegelberg, H., “Human dignity: A Challenge to Contemporary Philosophy”, in

Gotesky, R. and Laszlo, E. (eds.), 1970.
Sprigge, T. L. S., “A Utilitarian Reply to Dr. McCloskey”, Inquiry, vol. 8, 1965.
Stewart, M. A. (ed.), Law, Morality and Rights, Reidel, Dordrecht, 1979.
Stuart, J. D., “Retributive Justice and Prior Offences”, Philosophical Forum, vol.
18, 1986.
Tapper, C. F. H., (ed.), Crime, Proof and Punishment, Butterworth, London, 1991.
Ten, C. L., Crime, Guilt and Punishment, Clarendon Press, oxford, 1987.
Tonry, M., “Proportionality , parsimony, and interchangeability of punishments”, in

Duff, R. A, Marshall, S, Dobash, R. E., Dobash R., (eds.), 1995.
Tonry, M., Sentencing Matters, Oxford University Press, United States of America,
1996.
Tonry, M. and Hatlestad, K., Sentencing Reform in Overcrowded Times, Oxford

Tonry, M., “Interchangeability, Desert Limits and Equivalence of Function”, in von

Hirsch, A. and Ashworth, A. (eds.), 1998.
Tonry, M. (ed.), The Handbook of Crime and Punishment, Oxford, University

Press, New York, 1998.
Tamburrini, C. M., Crime and Punishment, Akademitryck AB, Sweden, 1996.
von Hirsch, A., Doing Justice, Hill and Wang, New York, 1976.
von Hirsch, A., “Desert and Previous Convictions in Sentencing”, Minnesota Law
Review, vol. 65, 1981.
von Hirsch, A., “Commensurability and Crime Prevention: Evaluating Formal

Sentencing Structures and Their Rationale”, The Journal of Criminal Law and
von Hirsch, A., Past and Future Crimes, Rutgers University Press, New Brunswick,
1985.
von Hirsch, A., “Deservedness and Dangerousness in Sentencing Policy”, Criminal

Law Review, 1986.
von Hirsch, A., “Hybrid Principles in Allocating Sanction: A Reply to Professor

Robinson”, Northwestern University Law Review, vol. 82, 1987
von Hirsch, A. and Jareborg, N., “Provocation and Culpability”, in Schoeman, F.,
(ed.), 1987.
von Hirsch, A., Knapp, K. A., and M. Tonry, The Sentencing Commission and Its
Guidelines, Northeastern University Press, Boston, 1987.
von Hirsch, A., Wasik, M., and Green, J. A., “Punishments in the Community and
the Principles of Desert”, Rutgers Law Review, vol. 20, 1989.
von Hirsch, A., and Jareborg, N., “Gauging Criminal Harm: A Living-Standard
Analysis”, Oxford Journal of Legal Studies, vol. 11, 1991.
von Hirsch, A., “Scaling Intermediate Punishments: A Comparison of Two

Models”, in Byrne, J. M., Lurigio, A. J., and Petersilia, J. (eds.), 1991.
von Hirsch, A. , “Proportionality in the Philosophy of Punishment: From “Why

Punish?” to “How Much?””, Israel Law Review, vol. 25, 1991.
von Hirsch, A., “Criminal Record Rides Again”, Criminal Justice Ethics, vol. 11,
1991.
von Hirsch and Jareborg, N., Strafmass und Strafgerechtigheit, Forum Verlag
Godesberg, Germany, 1991.
von Hirsch, A. and Ashworth. A., “Not Not Just Deserts: A Response to Braithwaite
and Pettit”, Oxford Journal of Legal Studies, vol. 12, 1992.
von Hirsch, A., “Proportionality in the Philosophy of Punishment”, Crime and

Justice, vol. 16, 1992.
BIBLIOGRAPHY 215
von Hirsch, A., Censure and Sanctions, Clarendon Press, Oxofrd, 1993.
von Hirsch, A., “Sentencing reform: its goals and prospects”, in Duff, R. A.,
Marshall, S., Dobash, R. E., Dobash, R. P. (eds.), 1994
von Hirsch, A., “The Future of Proportionate Sentencing”, in Blomberg, T. G. and

Cohen, S. (eds.), 1995.
von Hirsch, A., “Extending the Harm Principle: ‘Remote’ Harms and Fair
Imputation”, in Simester, A. P. and Smith, A. T. (eds.), 1996.
von Hirsch, A., “Desert and Previous Convictions”, in von Hirsch and Ashworth, A.
(eds.), 1998.
von Hirsch, A., “Seriousness, Severity and the Living Standard”, in von Hirsch, A.
and Ashworth, A. (eds.), 1998
von Hirsch, A., “Penal Theories”, in Tonry, M. (ed.), 1998.
von Hirsch, A. and Ashworth, A., Principled Sentencing, Hart Publishing, Oxford,
1998.
Walker, N., Punishment, Danger, and Stigma, Blackwell, Oxford, 1980.
Walker, N., Why Punish?, Oxford University Press, Oxford, 1991.
Walker, N., “Modern Retributivism”, in Gross, H. and Harrison, R. (eds.), 1992.
Walker, N., “Legislating the Transcendendal: von Hirsch’s Proportionality”,

Cambridge Law Journal, vol. 51, no. 3, 1992.
Walker, N., “Even More Varieties of Retribution”, Philosophy, vol. 74, 1999.
Wasik, M. and Pease, K., Sentencing Reform. Guidance or Guidelines?,

Manchester, 1987.
Wasik, M. and von Hirsch, A., “Non-Custodial Penalties and the Principles of
Desert”, Criminal Law Review, 1988.
Wertheimer, A., “Should Punishment Fit The Crime?”, Social Theory and Practice,
Vol. 3, no. 4, 1975.
Wertheimer, A., “Punishing the Innocent - Unintentionally”, Inquiry, vol. 20, 1997.
Williams, B., “The Macropulos Case: Reflections on the Tedium of Immortality”, in

Problems of the Self, Cambridge University Press, Cambridge, 1973.
Zedner, L., “Reparation and Retribution: Are They Reconcilable?”, Modern Law
Review, vol. 57, 1994.
Zimring, F., “Making Punishment Fit the Crime: A Consumer’s Guide to

Sentencing Reform”, in Duff, R. A. and Garland, D. (eds.), 1994.
INDEX
Acton, H. B. 10, 53, 150, 199, 208 Cohen, J. 153, 182, 217
Adams, K. 120, 199 Cohen, M. 153, 182, 217
Alexander, L. 180, 199 Cooper, D. E. 202
anchor problem 7, 48, 123, 129ff, Cottingham, J. G. 4, 53, 12, 202
138ff, 142, 144, 147ff, 163, 190 Cragg, W. 202f
Anscombe, G. E. 199 crime 2ff, 6f, 11f, 13ff, 21, 24ff, 30ff,
Arenella, P. 73, 75, 96f, 199 38f, 45ff, 51ff, 101ff, 110ff,
Armstrong, K. G. 10,199 124ff, 130ff, 135ff, 142ff, 147,
Arneson, R. J. 177, 182 158ff, 166ff, 184ff, 191ff
Ashworth, A. 10, 54, 57, 63, 70, 78, culpability 6, 59, 68ff, 75ff, 83f, 85,
95ff, 108, 120, 151f, 182, 197f, 93, 127, 158f, 167f, 185
199f, 209, 213ff Dagger, R. 38, 55, 170f, 202
attempts 50, 63f, 66, 87, 90ff, 189, Darley, J. M. 202
194 Davis, L. H. 15, 202
Bagaric, M. 10, 53, 200 Davis, M. 5, 39, 87ff, 91ff, 114, 141,
Baldwin, T. 24, 54, 200 173, 188, 202
Barnett, R. E. 200, 208 desert 3ff, 12, 14ff, 19, 22, 32, 39,
Bayles, M. 96, 200 44ff, 71ff, 77, 82, 93, 106, 114,
Beccaria, C. 12f, 93, 99, 200 123, 125f, 132, 142f, 184, 188,
Bedau, H. A. 200 194f
Belliotti, R. A. 151f, 200 deterrence 12, 19, 45, 93, 164, 169,
Benn, S. I. 99, 111, 200 187, 196
Bennett, J. 180, 201 dignity 7, 61, 126, 131ff, 137ff,
Bentham, J. 10ff, 53, 103, 109, 120, 141ff, 148
160, 201 Dolinko, D. 15, 53, 56, 98, 151, 182,
Blomberg, T. G. 201 202
Bottoms, A. E. 197, 201 draconian punishment 143f
Brandt, R. 96, 201 Duff. R. A. 10, 21, 25f, 28ff, 54ff,
Braithwaite, J. 152, 180, 201, 214 95f, 120, 169, 180, 199, 203
Brownless, I.201 Dunbar, I. 10, 203
Brownsword, R. 197, 201 Durham III, A. M. 80, 97, 203
Buchner, D. 120, 201 Dworkin, G. 174
Burgh, R. W. 38, 40, 43, 55ff, 181, Ellis, A. 55, 98, 204
201 Erikson, M. L. 120, 204
burglary 13, 59 Ewing, A. C. 206
Byrne, J. M. 99, 201, 214 excuses 27, 68, 71ff, 76f, 83, 159,
cardinal proportionality 13f, 48, 56, 168
124, 143ff, 149, 152, 156, 189 expressionism 6, 19ff, 28ff, 33f, 51,
Calvert, B. 201 115, 125, 143, 167ff, 179
Cederblom, J. B. 200f, 210 Ezorsky, G. 18f, 204
censure 21, 25ff, 30f, 36, 78, 143f, fairness theory 6, 8, 12, 15, 36ff,
Clark, R. S. 10, 202 40ff, 50f, 87, 93, 114, 125, 169ff,
Clarkson, C. 10, 202 175ff
condemnation 20, 22f, 25, 50, 184 Feinberg, J. 19, 36, 204
217
Fine 2, 6, 13, 19, 47, 51, 66, 101ff, Kadish, S. H. 96, 207
110, 112f, 124, 129f, 42, 149, 185 Kant, I. 4, 44, 72, 116, 134ff, 155
Finnis, J. 55, 296 Kahan, D. M. 207
Fletcher, G. 75, 77, 96f, 204 Kellogg, F. 207
Frankel, M. E. 10, 204 Kemp, P. 151, 207
Frase, R. S. 204 Kershnar, S. 42, 56, 152, 207
Freeman, M. D. A. 204 Khatchadourian, H. 207
Fried, C. 181, 204 Kleinig, J. 15, 125, 131, 137f, 207
Galligan, D. J. 13, 53, 188, 197, 204 Kolnai, A. 151, 207
Gardner, J. 95, 199, 205 Kuhse, H. 133, 207
Gerstein, R. S. 137, 152, 205 Lacey, N. 96, 182, 199, 207
Gibbs, M. R. 120, 203 Lempert, R. O. 156f, 161, 207
Glover, J. 181, 205 lex talionis 14, 68, 125ff, 130, 148
Golash, D. 205 living-standard 61f, 05
Goldman, A. H. 205 Lucas, J. R. 20, 22, 24, 35, 207
Gotesky, R. 152, 205 Lyons, W. 56, 207
Griset, P. L. 10, 205 Mabbott, J. D. 4f, 88, 208
Gross, H. 5, 46, 205 Mackie, J. L. 2, 208
Hampton, J. 54, 116, 121, 197, 205 Maclagan, W. G. 127, 208
harm 6, 14, 58, 60ff, 65ff78, 83ff, 87, Martin, R. 208
93, 107, 127, 133, 148 Matravers, M. 54, 202, 208
Harlow, R. E. 120, 205 McCloskey, H. J. 44ff, 50, 208
Harrison, R. 205 McMahan, J. 180, 208
Hart, H. L. A. 22, 36f, 45, 72, 84, mercy 7, 117ff
111, 168, 206 Meyer, M. J. 134, 208
Hegel, F. 125, 155, 202, 206 Miller, F. G. 208
Hestevold, S. 121, 206 Mitias, A. H. 56, 208
Hill, T. E. 151, 206 mitigation 72, 78, 81f, 116, 168
Hinton, J. W. 197, 201 Moberly, W. 48, 50, 208
Holmgren, M. 206 Moore, M. S. 49f, 161, 163, 208
Honderich, T. 206 Morison, J. 209
Hudson, B. 10, 121, 181, 198, 206 Morris, H. 40, 170, 209
Husak, A. N. 64, 206 Morris, N. 119, 192, 194, 209
hybrid theories 7, 14 Mundle, C. W. K. 15, 124, 209
individualization 188 Murphy, J. G. 123, 312, 135ff, 155,
intention 66, 69ff, 113, 158. 162ff 170, 178, 193, 209
intermediate sanction 101, 133 Nagel, T. 12, 209
Jacobs, J. 206 Narayan, U. 29ff, 209
Jareborg, N. 61ff, 66ff, 206 Nathan, G. 120, 212
Johnson, C. A. H. 121, 206 negligence 69, 158
just deserts 3, 5, 49f, 187 Nino, C. S. 210
justice: criminal 4, 107, 117, 158, Norcross, A. 180, 210
169 Nozick. R. 20f, 23f, 34f, 37, 68, 70f,
justice: social 7, 37, 156, 167, 177, 73, 84, 210
186f Oldenquist, A. 210
INDEX 219
ordinal proportionality 13f, 110, 124, Schedler, G. 211

126, 130, 136, 142f, 147ff, 157, Scheid, D.E. 84, 91f, 131f, 141, 212
189ff Schoeman, F. 57, 212
Parent, W. A. 54, 210 Sebba, L. 101f. 212
parity condition 33, 108, 110, 113, Sellin, T. 60, 86, 212
140, 190, 192 sensibility 102ff, 113, 160, 185, 188
penance 23ff seriousness of crime 59f, 76, 90, 93,
Pettit, P. 152f, 180 105, 125, 127ff, 131, 157, 176,
Philips, M. I. 180, 212 186
Pincoff, E. L. 210 sentencing system 4, 13, 17, 52, 65,
poena naturalis 51f, 112, 114 77, 84, 123, 157
Primoratz, I. 23, 59, 86, 149, 210 severity of punishment 13, 33, 59,
principled sentencing 87, 187 101f, 106, 125, 127f, 131, 157,
proportionality: departures from 19, 176, 186
194f; justification of 24, 49; Sher, G. 40ff, 114, 171f, 175, 209,
principle of 4, 12, 14, 34 212
prevention 7, 30f, 142ff, 147ff Shute, S. 95, 199, 212
punishment: capital 156f; of Sidgwick, H. 164, 212
innocents 44, 157; justification of Simester, A. P. 95, 212
19f, 30, 33, 36, 44, 68, 77, 111, Simmons, A. J. 205, 212
189 simple desert theory 6, 12, 14ff, 19,
Quinton, M. 53, 206, 210 22, 39f, 44, 51, 106, 124, 125,
Rachels, J. 53, 210 184
Rainbolt, G. 121, 210 Singer, G. 77, 123, 212
rape 13, 44ff, 49, 69f, 79, 126, 128 Singer, P. 165, 212
Rawls, J. 36, 210 Skillen, A. J. 54, 212
reform 4, 18, 21f, 25ff, 28f, 33, 35, Smart, A. 46, 213
178, 187, 196 Smart, J. J. C. 46, 213
rehabilitation 3, 12, 52, 93, 188 Smilansky, S. 213
Reiman, J. 126f, 211 Snare, A. 213
remote harms 164ff, 185 social justice, see justice
responsibility: character theory 72ff, Sparks, R. 95, 213
159, 185; choice theory 72ff, 76, Spiegelberg, H. 137, 213
185 Sprigge, T. L. S. 56, 213
retributivism 4ff, 11f, 15, 20, 45, Stewart, M. A. 202, 204, 213
48ff, 59, 110, 116, 125, 127, Stuart, J. D. 213
155ff, 161, 167, 178, 184, 187f, Tamburrini, C. M. 213
195 Tapper, C. F. H. 53, 95, 199, 213
Roark, A. P. 211 tax evasion 43, 92, 146
Robinson, P. H. 104, 190ff, 194, 211 Ten, C. L. 85, 128, 192, 215
Rosen, F. 211 Tonry, M. 119, 192, 201, 209, 213
Rothman, D. J. 211 torture 89, 92, 1126, 133, 138f, 141
Ryberg, J. 55, 97, 211 treatment 3, 22, 23f, 27f, 30f, 69, 80,
Saduski, W. 39f, 114, 170, 173ff, 211 104f, 107ff, 137 143f
sanction, see punishment
unfair advantage theory, see fairness

theory
utilitarianism 13, 44ff
victims 61, 78, 137, 158, 171, 185
Voltaire, J.F. M. A. 44
von Hirsch, A. 21ff, 29ff, 61ff, 65ff,
76, 78f, 80ff, 104, 107, 123,
142ff, 148, 158, 167f, 186, 201,
203, 213f
Walker, N. 11, 110, 215
Wasik, M. 104f, 200, 205, 214
Wertheimer, A. 180, 215
Williams, B. 56, 213, 216
Wolfgang, M. E. 60, 86, 216
Zedner, L. 216
Zimring, F. 216
LIBRARY OF ETHICS AND APPLIED PHILOSOPHY
1. O. Lagerspetz: Trust: The Tacit Demand. 1998 ISBN 0-7923-4874-5
2. W. van der Burg and T. van Willigenburg (eds.): Reflective Equilibrium. Essays in
Honour of Robert Heeger. 1998 ISBN 0-7923-5066-9
3. J. G. Murphy: Character, Liberty, and Law. Kantian Essays in Theory and Practice.
1998 ISBN 0-7923-5275-0
4. B. C. Postow: Reasons for Action. Toward a Normative Theory and Meta-Level
Criteria. 1999 ISBN 0-7923-5700-0
5. D. Richter: Ethics After Anscombe. Post "Modern Moral Philosophy". 2000
ISBN 0-7923-6093-1
6. G. den Hartogh (ed.): The Good Life as a Public Good. 2000 ISBN 0-7923-6167-9
7. T. van den Beld (ed.): Moral Responsibility and Ontology. 2000 ISBN 0-7923-6255-1
8. M.J. Almeida (ed.): Imperceptible Harms and Benefits. 2000 ISBN 0-7923-6464-3
9. J.A. Corlett: Responsibility and Punishment. 2001 ISBN 0-7923-7167-4
10. M. Gore Forrester: Moral Beliefs and Moral Theory. 2002 ISBN 1-4020-0687-X
11. A.W. Musschenga, W. van Haaften, B. Spiecker and M. Slors (eds.): Personal and
Moral Identity. 2002 ISBN 1-4020-0764-7
12. C. Wilks: Emotion, Truth and Meaning. In Defense of Ayer and Stevenson. 2002
ISBN 1-4020-0916-X
13. M. Schermer: The Different Faces of Autonomy. Patient Autonomy in Ethical Theory
and Hospital Practice. 2002 ISBN 1-4020-0984-4
14. H. Dyke (ed.): Time and Ethics: Essays at the Intersection. 2003
ISBN 1-4020-1312-4
15. J. Ryberg and T. Tännsjö (eds.): The Repugnant Conclusion. Essays on Population
Ethics. 2004 ISBN 1-4020-2472-X
16. J. Ryberg: The Ethics of Proportionate Punishment. A Critical Investigation. 2004
ISBN 1-4020-2553-X
KLUWER ACADEMIC PUBLISHERS – DORDRECHT / BOSTON / LONDON

The Ethics of Proportionate Punishment

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Ethics of Proportionate Punishment

Uploaded by

Copyright:

Available Formats

THE ETHICS OF PROPORTIONATE PUNISHMENT

LIBRARY OF ETHICS AND APPLIED PHILOSOPHY

Govert A. den Hartogh, University of Amsterdam, The Netherlands

KLUWER ACADEMIC PUBLISHERS

ISBN 1-4020-2553-X (HB)

Published by Kluwer Academic Publishers,

Sold and distributed in North, Central and South America

In all other countries, sold and distributed

Printed on acid-free paper

All Rights Reserved

Printed in the Netherlands.

Chapter 1 Proportionalism and its Justifications 11

Chapter 2 The Seriousness of Crimes 59

Chapter 3 The Severity of Punishments 101

Chapter 4 The Anchor Problem 123

Chapter 5 Proportionalism and Penal Practice 155

Chapter 6 Relaxed Proportionality 183

The philosophical discussion of state punishment is well on in years. In contrast

1. WHY CONSIDER PROPORTIONALISM?

The question of how punishments should be distributed in a society is obviously of

by the development of retributive emotions. Whether it is correct that retributivist

assessment of degree of severity, cannot be provided. Considerations on the possible

proportionality are considered and it is argued that none of these conjectures

PROPORTIONALISM AND ITS JUSTIFICATIONS

Though the idea of proportionalism is susceptible to different interpretations and has

What I shall do is to distinguish three overall accounts of retributivism: the

That the question of punishment distribution should be answered by adopting a

proportionality (though, of course, not if it does both). However, such punishments

2. THE SIMPLE DESERT THEORY

The cardinal concept in the various versions of retributivist theories of punishment

Western in which an irremediable wicked villain meets an unpleasant end. The

(1) A criminal deserves to suffer proportionately to the seriousness of the crime

in comparison to other criminals who have perhaps committed equally serious

3. THE EXPRESSIONIST THEORY

A dominant view amongst theorists contemplating criminal sanctions has been to

communicative dimension of punishment. With them, punishment might be seen as

assertion, whereas the goal of the teleological retributivist corresponds to the

An interesting preliminary question which naturally arises is, why should

transformation of the criminal. The punishment should be a vehicle for repentance

Another question relating to the double purpose of conveying blame and

person determine the severity of a punishment without failing in the communication

“as tigers might be treated in a circus, as beings that have to be restrained,

4. THE FAIRNESS THEORY

At first glance the principle seems reasonable. If we think of a cooperative

proportionalism, would no longer be blocked by a position which is retributive only

Besides the introductory indication of what adherents of retributivism standardly

wrongdoers is of basic ethical importance, or whether the fairness theory constitutes

What both of these responses do is to abandon proportionality as a simple

THE SERIOUSNESS OF CRIMES

Since the claim of proportionalism is that the severity of punishment should be

becomes more demanding the larger the number of different seriousness-

1. THE HARM DIMENSION

harm gravity: grave, serious, upper-intermediate, lower-intermediate, and lesser.

That the question of the allocation of punishment should be determined on the

CRIME HARMMENS REA RANKING

murder death intention murder

different degrees of capacity-excuses and fair opportunity-excuses: what is required

character, it will probably be necessary to answer the intricate question of what a

CRIME EXCUSE RANKING

purposefulness or carelessness and to develop theories of partial excuses in order to

specifically in a non-recidivist discount is not further explained, though it can hardly

Another comment on the ranking problem is provided by Ten. He believes

Davis sets out by suggesting a seven-step procedure for the assignment of