http://plato.stanford.

edu/entries/criminal-law/

Duff, Antony, "Theories of Criminal Law", The Stanford Encyclopedia of Philosophy (Summer
2013 Edition), Edward N. Zalta (ed.), URL =
<http://plato.stanford.edu/archives/sum2013/entries/criminal-law/>.

Theories of Criminal Law

Antony Duff

First published Mon Oct 14, 2002; substantive revision Tue May 14, 2013
Philosophical ‘theories of criminal law’ may be analytical or normative (§ 1). Once we
have identified the salient features that distinguish criminal law from other kinds of law
(§2), we ask whether and why we should maintain such an institution (§3).
Instrumentalist answers to this question portray criminal law as an efficient technique
that helps us achieve worthwhile ends; non-instrumentalist answers portray it as an
intrinsically appropriate response to certain kinds of wrongful conduct (§4). By
considering the question of how the criminal law should address citizens (§5), we can
discern the truth in the non-instrumentalist perspective. The next question concerns the
proper scope of the criminal law: what kinds of conduct should be criminalized?
Several candidate principles of criminalization are critically discussed (§6), including
the Harm Principle, and the claim that the criminal law should be concerned with
‘public’, rather than merely ‘private’, wrongs. Further questions are raised, however
(§7), by the increasingly important phenomenon of international criminal law.

 1. Different Kinds of Theory

 2. Aspects of Criminal Law
 3. Should We Abolish the Criminal Law?
 4. Instrumental and Moralistic Conceptions of Criminal Law
 5. The Law's Voice
 6. Crimes as Public Wrongs
 7. International Criminal Law
 8. The Internal Structure of the Criminal Law
 9. Normative Theory and Political Reality

1
  Bibliography
 Academic Tools
 Other Internet Resources
 Related Entries

1. Different Kinds of Theory

‘Theories of criminal law’ could just be general theories of law applied to the particular
case of criminal law: proponents of legal positivism, of natural law, of economic
analysis of law, of Critical Legal Studies and other schools of legal theory will expect to
be able to say about the criminal law what they say about law in general (for examples
of the last two approaches, see Posner 1985; Kelman 1981). Questions raised by
theories of this kind will figure in what follows—for instance whether it is part of the
essence of criminal law that it must satisfy, or make, certain kinds of moral demand;
whether criminal law can be adequately understood in purely instrumental terms;
whether we should take the criminal law's apparent pretensions to rationality and
principle seriously, or should rather see it as an oppressive exercise of political or
economic power, or as the site of conflicts which produce an irredeemably
contradictory, unprincipled set of doctrines and norms (see Norrie 2001). Such
questions are important, but we will not begin with them. We should, instead, begin by
asking what is distinctive about criminal law. What marks it out from other kinds or
aspects of law? What are its distinctive institutional structures, purposes, or content?
Philosophical theories of criminal law can be analytical, or normative (see Husak 1987:
20–26). Analytical theorists seek to explain the concept of criminal law, and related
concepts such as—most obviously—that of crime (metaphysically more ambitious
theorists might seek an account not merely of the concept of criminal law, but of its
real, metaphysical nature; see Moore 1997: 18–30). They need not look for a strict,
ahistorical definition—an account of the necessary and sufficient conditions given, and
only given, which a human practice counts as a system of criminal law; we have no
reason to think that any such definition will be available. But they can hope to identify
and explain the central or salient features of systems of criminal law—features at least
some of which will be exhibited by anything we can count as a system of criminal law;
and to develop an account of a paradigm of criminal law, on the basis of which we can
recognise as systems of criminal law other practices that resemble that paradigm
sufficiently closely, even though they do not quite fit it.
Normative theorists seek an account not just of what criminal law is, but of what it
ought to be (and whether it ought to be at all). Should we maintain a system of criminal

2
law? If so, what goals should it serve, what values should inform it, what should its
scope and structure be? Any such normative theory must presuppose some analytical
account of that whose goals, values, scope and structure are being discussed. Whether
analytical and normative theorising are related more closely than this will depend on
what kind of analytical theory we develop: a legal positivist will insist that, here as
elsewhere, the question of what law ought to be is quite separate from, and left open by
answers to, the question of what law is; a Natural law theorist will argue that an
adequate analysis of the concept or the metaphysical nature of criminal law will reveal
the moral purposes or values that a practice must serve (or at least claim to serve) if it is
to count as a system of criminal law at all (see Moore 1997: 23–35).
Philosophical theories of criminal law, whether analytical or normative, cannot subsist
in isolation. For one thing, they cannot be wholly separate from other branches of
philosophy. They must draw, most obviously, on political philosophy, since they must
depend on some conception of the proper aims of the state and of the proper
relationship between a state and its citizens (see e.g. Pettit 1997, 2002, forthcoming;
Duff 2001: chs 2–3; Dubber 2005; Dagger 2008, 2011; Matravers 2011; Thorburn
2011a; Dimock forthcoming). They must draw on moral philosophy, insofar as the
criminal law properly aims to define types of moral wrong and to punish those who
culpably commit them (see e.g. Moore 1997; Tadros 2005). They must draw on
philosophy of action and on philosophy of mind, if they are to explicate ideas of
wrongdoing and of fault that are appropriate to law's wrong-defining role (see, e.g.,
Moore 1993; Green 2005, 2012; Gardner 2007; Yaffe 2010).
Philosophical theories must also, however, draw on the resources of other disciplines.
They must attend to the empirical actualities of that which they theorise: to the histories
of the different systems of criminal law, and to sociological inquiries into their actual
operations. Some critical theorists believe that such historical or sociological inquiries
will undercut the pretensions of philosophical theorising: that what needs analysing is
not the superstructure or superficial self-presentation of the criminal law, on which
philosophers tend to concentrate, but the social, political and economic realities lying
beneath that surface; and that given the oppressive or conflictual nature of those
realities, philosophical theories cannot amount to anything more than doomed attempts
to rationalise what is inherently irrational or a-rational (see Kelman 1981; Norrie 2001;
also Law and Ideology). The only adequate reply to such critiques of philosophical
theorising is to show how such theorising can assist both an understanding of what
criminal law is, and the discussion of what it ought to be, by taking seriously the
concepts in terms of which it presents itself: that is the task on which we embark in
what follows.
There remain some large questions, however, about the character and the proper
amibitions of such theorising. How far must it be, and remain, relatively local, as

3
theorising about what criminal law can and should be in particular kinds of society, in
particular historical and political settings? How far can it hope to transcend both its own
history, as theorising from within a particular tradition, and that of the systems of
criminal law from which it begins, towards a more general (if not a universal) account
of what criminal law should be? Although we cannot pursue such questions here (but
see Farmer 1996, 2010; Lacey 2000, 2001, 2009; Brown 2009), they challenge anyone
seeking to develop a normative theory of criminal law (see Duff and Green 2010b)

2. Aspects of Criminal Law

We can usefully begin by identifying some of the salient features of the systems of
criminal law with which we are familiar (contemporary systems of municipal criminal
law): features by which we can distinguish the criminal law both from non-legal
phenomena and from other types of law. It would be unproductive to ask whether all
these are strictly necessary features of criminal law, or whether we might still count a
practice that lacked one or more of them as a system of criminal law; the most we can
sensibly claim is that these are central features of the paradigm of criminal law as we
understand and experience it.
The criminal law deals with crimes: but what is a crime—and how does the criminal
law deal with it?
Crimes, we might initially say, are kinds of conduct that are defined by the law as
wrong. However, even this crude initial approximation must be qualified. First, we can
say that crime always involves ‘conduct’ only if we stretch the meaning of that term so
far as to empty it of substantial content (see Husak 1987: ch. 4): we can (whether justly
or not) be held criminally liable not merely for what we do, or fail to do, but for what
we are, perhaps even for what we think—for what we intend, for instance. But for the
moment we can talk of ‘conduct’, since it captures the most familiar kinds of crime.
Second, we must not, or must not yet, read ‘wrong’ here as ‘morally wrong’: it will be a
further question whether the criminal law either must of its nature, or should as a matter
of normative theory, portray the conduct it criminalizes as morally wrong; all we should
say so far is that it portrays it as being in some way wrong or defective, something that
those bound by the law should not do (this point is often expressed by saying that the
criminal law ‘prohibits’ the conduct that it defines as criminal, but we will see in
section 5 that this is misleading). That is, for instance, the defining difference between a
law that defines a certain kind of conduct as a crime which is punishable by a fine, and
one that subjects that conduct to a tax: both laws might be intended to reduce the
incidence of the conduct, but the former, unlike the latter, does so by defining and
punishing it as wrong.
Crimes differ from extra-legal wrongs in that they are defined as wrongs by the law:
they are not, or not just, wrongs in terms of some extra-legal social standard of

4
morality, prudence, or etiquette, but wrongs that are defined and recognised as such by
the law. (This leaves open the question of whether the criminal law can create wrongs,
or whether it rather gives formal recognition to wrongs whose wrongfulness is initially
determined by extra-legal standards.) But not all legally defined wrongs are criminal
wrongs.
First, some legal systems distinguish between ‘crimes’ properly speaking and other
kinds of penalised conduct. So German law distinguishes ‘Strafrecht’ and ‘Straftaten’
(criminal law and crimes) from ‘Ordnungswidrigkeitenrecht’ and
‘Ordnungswidrigkeiten’ (regulations and violations; Weigend 1988); and the American
Law Institute's Model Penal Code distinguishes ‘crimes’ from ‘violations’ (s. 1.04).
Violations might include conduct that other legal systems count as criminal, although
even in systems in which it counts formally as criminal, it is often seen as not ‘really’
criminal (thus German Ordnungswidrigkeitenrecht includes many traffic violations that
English law defines as crimes, although many drivers would deny that they are ‘real’
crimes). They are distinguished from (‘real’) crimes by the procedures for dealing with
them, by the relative mildness of the sanctions they attract, and by the absence of some
of the other consequences that typically attach to conviction for a crime—such as a
criminal record. I will not discuss this distinction further here, save to note that it might
be justified (if it is justifiable) either on pragmatic grounds—mere violations are not
dangerous enough to justify mobilising the expensive resources of the criminal justice
system; or on principled grounds—they do not involve serious enough wrongs to merit
the condemnation that a criminal conviction, as we will see, involves (for further
discussion, see Weigend 1988; Steiker 1997; Duff et al 2007: ch 6.5; Tadros 2010).
Second, most legal systems distinguish criminal from civil wrongs: wrongs that ground
a criminal prosecution, from those that ground a civil case for damages brought by the
injured party. We can clarify the concept of crime by focusing on this distinction. The
same conduct often constitutes both a criminal and a civil wrong, as is shown most
dramatically when, after a failed prosecution or a decision not to prosecute, the victim
or her family bring a civil case for damages against the alleged wrongdoer: but we can
still usefully ask what the difference is between defining and treating conduct as a
criminal wrong and defining and treating it as a civil wrong (see generally Murphy &
Coleman 1990: ch. 3; Boston University Law Review 1996).
Civil wrongs are typically treated as ‘private’ matters in the sense that it is for the
victim to investigate what happened, to identify the alleged wrongdoer, and to bring a
case against him. The law provides the institutions (the courts, arbitration panels)
through which that case can be brought; it lays down the norms by reference to which
the case is decided; it specifies what remedies are available; it might also help
successful plaintiffs to extract damages from unwilling defendants. But it is for the
injured party to bring, or to decide not to bring, a case; to pursue, or to abandon, that

5
case; to insist on extracting the damages the court awarded, or to forgo them. The case
is described and understood as ‘P v D’: P sues D, and the case thus belongs to her. The
criminal law, however, provides for the public investigation, prosecution and
punishment of crimes: for a police force, tasked with investigating (as well as
preventing) crime and detecting criminals; for a system of criminal courts, in which
defendants are tried for the crimes that they are alleged to have committed (and whose
workings are structured by a complex array of procedural rules and requirements); for a
system of punishments that will be imposed by the courts, and administered by other
institutions and officials. Now the police act in the name and with the authority not just
of the victim, but of the whole polity; it is for the prosecuting authority, not for the
victim, to decide whether, and on what charge, anyone will be prosecuted. If the victim
does not want the case to go to court, the prosecutors will in fact often not proceed with
it—because it would be hard to do so without the victim's willing co-operation, or out
of concern for the victim's feelings; but cases can be prosecuted despite the victim's
unwillingness (this can be an important issue for prosecutors dealing with domestic
violence; see Dempsey 2009). When the case comes to court, it is described not as
‘P vD’, but as ‘State v D’, or ‘People v D’, or ‘Queen v D’: D is prosecuted not by an
individual victim, but by the polity—or, in societies that have not yet shaken off the
trappings of undemocratic monarchy, by its sovereign. (Some legal systems allow the
possibility of private prosecutions; this is one of several ways in which the distinction
between criminal and civil law is neither sharp nor watertight.)
The difference between the public character of criminal wrongs and the private
character of civil wrongs is also evident in the outcomes of the two kinds of legal
process. A civil case typically results in a finding either for the plaintiff, or for the
defendant; if the plaintiff wins, the defendant may have to pay her damages, as
compensation for the harm that she suffered, and for which she has sued. Criminal
cases, by contrast, result in a conviction or an acquittal; if the defendant is convicted, he
is liable to suffer a punishment. Criminal convictions express an explicit condemnation
of the defendant: he has been proved guilty of doing wrong, and the verdict is focused
on that wrong. A verdict for the plaintiff in a civil case will typically imply that the
defendant acted wrongfully, but the focus of the case, and thus of the verdict, is often
more on who should pay for whatever harm was caused (see Ripstein 1999: chs. 2–4).
Finally, the punishments imposed for crimes differ from the damages that are awarded
as a result of a civil suit—and not just in the fact that whereas the successful plaintiff
can forego the damages she is awarded, it is not for the victim of a crime to decide
whether the sentence imposed by the court should be carried out. Often punishments
take a different material form from civil damages, as when an offender is imprisoned or
put on probation. Even when their material form does not differ, however (as when a
convicted offender is fined £1,000, and the defendant who loses a civil case is ordered
to pay £1,000 in damages), their meanings differ. First, even if the severity of criminal

6
punishments is to some degree determined by the extent of the harm caused (itself a
controversial matter), it typically also depends on the nature and degree of the
offender's culpability for that harm: someone who kills or injures recklessly can expect
to be punished more severely than someone who causes death or injury by a negligent
act or omission. Civil damages, however, are typicaly proportioned to the harm actually
caused; some kind of culpability, such as negligence, might be a threshold requirement,
in that the plaintiff must prove that the defendant was at least negligent in relation to the
harm he caused; but the damages are not proportioned to the degree of the defendant's
culpability, since their purpose is simply to provide compensation for the harm caused.
Second, punishments are intended to be burdensome (indeed, this is standardly cited as
a defining feature of punishment), whereas civil damages are not. If I am ordered to pay
£1,000 compensation for damage that I negligently caused to your property, making
that payment might be burdensome for me, if I am not well off, or no burden at all, if I
am rich (or have suitable insurance): but the damages serve their purpose in either case.
If, however, I am fined as punishment for a crime, that fine is intended to be
burdensome: that is why many sentencing authorities aim to proportion fines to the
offender's means, to ensure that both rich and poor offenders are burdened fairly and
proportionately; and that is what is wrong with the idea that one might take out
insurance to cover the cost of fines.
These distinctions between criminal and civil outcomes are often blurred in practice.
For instance, English criminal courts can include a ‘compensation order’ in the sentence
that they impose, thus bringing a dimension of civil law into the criminal process (see
Powers of Criminal Courts (Sentencing) Act 2000 s.130); in some systems, the victim
of a crime can pursue a claim for compensation at the same time, in the same court, as
the criminal prosecution. Some systems of tort law allow for the award of ‘punitive
damages’: these are intended to burden the defendant, and are typically awarded in
cases in which the defndant was egregiously at fault, or acted maliciously. Furthermore,
while many accounts of tort law portray it, as it was portrayed above, as being
essentially concerned with allocating the costs of harm (on the basis of considerations
of justice, or of efficiency), others portray it as being more directly concerned with
wrongs—for instance as enabling those who have been wronged to seek ‘civil recourse’
against those who wronged them (see e.g. Zipursky 1998; Goldberg and Zipursky
2010; Theories of Tort Law s. 3.2). Such an account seems particularly suitable for the
kinds of case that do not sit easily in a cost-allocation model: cases (such as those in
which a pliantiff sues for libel, or bereaved parents sue the hospital or corporation that
negligently caused their child's death) in which what the plaintiff properly seeks is not
compensation for any material harm that they have suffered, but an apologetic
recognition of and response to the wrong they have suffered. When damages are
awarded in such cases, they seem more like punishments that are intended to burden the

7
wrongdoer, as a way of making moral reparation to the person wronged, than like an
allocation of the costs of harm (see Duff 2001: 94–96).
Insofar as tort law can be portrayed as a matter of allocating the costs of harm, we can
draw a straightforward analytical distinction (albeit one that is often blurred in practice)
between a tort paradigm which focuses on harm that has been caused and on the
question of where the costs of that harm should fall; and a criminal paradigm, which
focuses on a wrong that has been done and on the question of who—if anyone—should
be condemned and punished for that wrong. The tort paradigm is a matter of private law
in that it aims to provide compensation and satisfaction for the aggrieved plaintiff, if she
chooses to pursue the case; the criminal paradigm is, by contrast, a matter of public law
in the sense that the case is brought and the punishment is imposed in the name and on
behalf of the whole polity rather than any individual victim. Insofar as we should
instead or also see tort law as a matter of civil recourse, that analytic distinction is less
straightforward, since both processes are now concerned with holding wrongdoers to
account; but a key difference remains—that criminal law makes the wrong a public
matter, that is to be pursued and punished in the name and on behalf of the polity rather
than the individual victim (if there is one).
We now have a sketch of the criminal law as a distinctive kind of human institution.
This then raises three further questions. First, should we have such an institution at all?
Second, if we should, what goals or purposes should it serve? Third, what should its
content be, or how should that content be determined: what kinds of conduct should be
criminal, and how should we go about deciding that issue?

3. Should We Abolish the Criminal Law?

Although the criminal law is a pervasive, and might seem to be an inescapable, feature
of the developed societies in which we live, there are those who argue that, precisely in
virtue of the paradigmatic features identified in the previous section, it is an institution
that we should seek to abolish: this is a central strand of the ‘abolitionist’ movement
which, whilst often focusing most directly on the abolition of criminal punishment, also
incorporates a critique of criminal law (seeLegal Punishment ss. 2, 7). Abolitionist
critics focus on three aspects of criminal law which, they argue, make it an utterly
unsuitable institution for the kinds of social life and the kinds of relationship that we
should seek (see Christie 1977; Hulsman 1986; Bianchi 1994).
First, the criminal law purports to declare and enforce authoritative standards of value,
in particular of moral value: it claims the authority to tell us how we should live, and to
enforce its demands on us if we disagree or disobey. But this, critics argue, amounts to
an illegitimate attempt to impose a moral consensus—inevitably the consensus of those
with political power—on societies which are rather characterised by radical moral

8
disagreement; it denies to those who do not share that consensus the freedom to think
and live as they see fit.
Second, the criminal law ‘steals conflicts’ from those to whom they properly belong. Of
course citizens often find themselves in conflict with one another; their relationships are
often impaired by various ‘troubles’. Such conflicts and troubles must be resolved; any
harms that have been done must be repaired. But that is a task for those most directly
involved—for the ‘victim’ and the ‘offender’ (though we should be cautious about such
notions), with the help of their local community. The criminal law, however, in defining
such conflicts or troubles as criminal wrongs to be dealt with by a public criminal
process, steals them: it transfers them to the professionalized context of a criminal
justice system in which neither victims nor offenders are allowed really to participate; it
thus denies those to whom the conflict belongs the chance to work it out for themselves.
Third, the criminal law deals in punishment—in ‘pain delivery’—when what is needed
is instead a process that will repair whatever harm was caused, reconcile the people
involved in the conflict, and thus restore the relationships that the conflict damaged.
Criminal punishment cannot contribute to those appropriate ends: it reflects a primitive,
backward-looking concern with retributive justice, whereas we should rather be seeking
a forward-looking restorative or reparative justice.
I will not discuss the third objection here, since it belongs with the discussion of
criminal punishment (see Legal Punishment ss. 2, 7): but we will see that the familiar
consequentialist and retributivist models for the justification of punishment have
analogues in accounts of the proper aims of criminal law more generally. The following
sections will constitute an answer to the first two objections. Abolitionists are right to
highlight these two features of a system of criminal law: it claims the authority to
declare certain public norms of conduct (norms that must, as we will see, claim a moral
foundation), and to insist on respect for those norms even from those who do not share
them; and it makes breaches of those norms its business, and so the business of the
whole polity in whose name and on whose behalf the law claims to speak and to act,
rather than leaving them as the business purely of those who are most immediately
involved. The question is whether and how we can justify maintaining institutional
practices of this kind.
The first objection is, as it stands, unimpressive, and sometimes expresses an incoherent
moral relativism which makes the moral demand that we should not make moral
demands of others (see B. Williams 1976: 34–9). It does reflect two general issues that
face any attempt to justify systems of political authority and law: the question of how
far a polity depends for its legitimacy on a normative consensus, at least a Rawlsian
overlapping consensus, amongst its members, and how far law and polity are possible
in contexts of radical disagreement; and the question of whether and how a polity can
claim legitimate authority over those who reject its central values and its normative

9
claims. We cannot pursue these questions here, although we may note that they are as
urgent for abolitionists as they are for advocates of the criminal law, since their
favoured practices and institutions depend, just as a system of criminal law does, on the
legitimacy and authority of the polity that sustains them. However, this first objection
does also raise a question that is more specific to the criminal law, and that must be
answered by those who would defend the criminal law: what kinds of norm, with what
kind of claimed authority, does or should the criminal law declare—and should we
maintain an institution that seeks to declare and support such norms?
The second objection focuses our attention on the distinction between civil and criminal
law sketched in the previous section. We might agree with the abolitionists that our
existing criminal procedures do not allow either victims or offenders the actively
participatory roles that they should be able and encouraged to play, but the basic
question is still this: should we maintain a system of law that defines and responds to a
category of ‘public’ wrongs—wrongs that concern not only the particular victim and
offender, but the whole polity; wrongs which are ‘our’ business collectively as a polity,
and which must therefore be investigated and dealt with by a public process—which
inevitably involves taking them out of the hands of those most immediately affected by
them?

4. Instrumental and Moralistic Conceptions of

Criminal Law
We can begin to tackle these two questions by distinguishing two radically different
ways of conceptualising criminal law. We might decide, in the end, that a plausible
account will have to draw on both kinds of conception; but we can usefully begin by
contrasting simple, pure versions of each.
One conception is instrumental. The criminal law is a technique or instrument that can
be used to serve various possible ends. We are justified in maintaining a system of
criminal law if it is an efficient technique for achieving worthwhile ends; its structure
and content should then be determined by asking how it can serve those ends most
efficiently.
What worthwhile ends could a system of criminal law serve? We cannot simply say that
it should prevent or reduce crime, since without the criminal law there would be no
crimes—no conduct would count as criminal. However, a number of plausible goals
could be posited, reflecting a range of views both about human goods and about the
proper roles and functions of the state. The American Model Penal Code, for instance,
declares that:
The general purposes of the provisions governing the definition of offenses are:

10
a) to forbid and prevent conduct that unjustifiably and inexcusably inflicts or threatens
substantial harm to individual or public interests [s. 1.01(1)].
We begin with a set of individual and public interests that merit protection, given their
role in human welfare: they can be protected by various methods, including various
state activities; a system of criminal law makes its distinctive contribution to their
protection by forbidding and thus preventing conduct that threatens substantial harm to
them. German criminal law theory posits a similar starting point: a set of individual and
collective Rechtsgüter (a Rechtsgut is a good which the law properly recognises as
being necessary for social peace or for individual well-being, and as therefore meriting
legal protection) which the criminal law protects against conduct that seriously
threatens them (see Roxin 2006: 8–47; for critical discussion of the utility of the idea
of Rechtsgütersee Wohlers et al 2003). As we will see in § 6, it is not yet clear whether
or how individual as distinct from public or collective interests should figure in an
account of the protective aims of the criminal law, and some accounts certainly
emphasise the collective dimension. Thus on Walker's ‘pragmatic’ account, the criminal
law should aim to further the “smooth functioning of society and the preservation of
order” (Walker 1980: 18, quoting Devlin 1965: 5)—collective or shared goods which
provide essential preconditions for individual flourishing.
Two aspects of such instrumentalist accounts are worth noting here. First, they typically
limit the criminal law's concern to serious harms to the specified kinds of interest,
which cannot be otherwise prevented: thus the Model Penal Code refers to “substantial
harm”, and German theorists argue that criminal law should be used only as a last resort
against seriously harmful conduct (see Roxin 2006: 45–7; also, more generally, Husak
2004; Jareborg 2005). This kind of limitation can itself be rationalised in instrumental
terms. The criminal law is a blunt and oppressive technique, which impinges seriously
on the interests of those who are subjected to its coercive attention: not just those who
are convicted and punished, but also those who are caught up in police investigations,
or who are tried and acquitted, A consequentialist calculus of costs and benefits is
therefore unlikely to favour its use unless it is the only feasible method of preventing
quite serious harm.
But, second, the Model Penal Code also limits the criminal law's concern to conduct
that “unjustifiably and inexcusably inflicts or threatens substantial harm” (see also
Feinberg 1984: 31–6), and most criminal codes include similar limitations. The
‘unjustifiably’ limit might still be justified instrumentally: we should not want to
prevent conduct that justifiably causes harm. Some theorists argue that we can also
justify the ‘inexcusably’ limit in instrumental terms (e.g. Braithwaite & Pettit 1990):
that the criminal law's goals are not efficiently served by criminalizing faultless or
excusable conduct. Others, however, ground this limit in a non-instrumental side-
constraint on the aim of harm-prevention: a purely instrumentalist theory cannot justify
criminalizing only culpable conduct; we must instead appeal to a non-instrumentalist

11
demand of justice, that those who lack fault should not be liable to criminal punishment
(see Hart 1968: 17–24, 28–53).
What emerges here is a familiar difference between two types of instrumentalist theory
(see Braithwaite & Pettit 1990: 26–36). A pure instrumentalist seeks to explain every
aspect of a justified system of criminal law in consequentialist terms: in designing a
system, we need only ask which doctrines, practices and rules will efficiently serve the
goals we have posited. A side-constrained instrumentalist, by contrast, argues that our
pursuit of those goals is also constrained by non-consequentialist values (for instance by
requirements of justice) which might preclude some practices (for instance the
criminalization of faultless conduct) even if those practices would efficiently serve the
system's goals.
For any instrumentalist theory, whether pure or side-constrained, it is an open empirical
question whether we should maintain a system of criminal law at all: we should do so
only if this is an efficient means to whatever goals the theory posits. For a pure
instrumentalist, the proper structure and contents of a system of criminal law also
depend on an empirical inquiry into how those goals can be most efficiently served,
whilst side-constrained instrumentalists must also attend to whatever non-
consequentialist constraints bear on these issues. We cannot pursue the debates between
these two kinds of account, but should note one set of questions that must figure in
them, about the relation between criminal and moral wrongs.
Some instrumentalists hold that we should criminalize only conduct that is in some way
immoral, and should punish only agents who are morally culpable for such conduct:
thus, for instance, Braithwaite and Pettit “assume … that only persons who are morally
culpable for a prescribed [sic] encroachment upon the dominion of others should be
convicted” (1990: 99), whilst the Model Penal Code declares another purpose of the
criminal law to be “to safeguard conduct that is without fault from condemnation as
criminal” (s. 1.02(1)(c)). Others seem less sure about this. Walker, for instance, sees
reason to criminalize conduct that provokes social disorder, even if we would, speaking
morally, blame that disorder on the intolerance of others rather than on the moral
wrongness of the conduct (1980: 21). As for moral culpability, Hart argues that we
should explain excuse doctrines not as aiming to exempt the morally faultless from
criminal liability, but as aiming to protect individual freedom by subjecting to liability
only those who had a fair opportunity to avoid it (1968: 17–24).
Thus an instrumentalist approach to the justification of criminal law seems to leave it as
something of an open question whether the law should criminalize only immoral
conduct, or should subject only morally culpable agents to criminal liability. At the
other extreme of the spectrum of theories of criminal law, by contrast, we find accounts
that make immorality and moral culpability central to the proper concerns of the
criminal law.

12
Criminal law, Stephen notoriously argued, “is in the nature of a persecution of the
grosser forms of vice”; conduct is properly criminalized
not only because [it is] dangerous to society, and so ought to be prevented, but also for
the sake of gratifying the feeling of hatred—call it revenge, resentment or what you
will—which the contemplation of such conduct excites in healthily constituted minds
(1873: 152).
One could read Stephen as offering what is still an instrumentalist account of criminal
law; it is important to satisfy that “feeling of hatred and the desire of vengeance” in “a
regular public and legal manner” (loc. cit.), because otherwise they will find more
violent, uncontrolled and socially harmful expression (compare Gardner 1998: 31–32).
But he clearly also believed that such feelings and desires were intrinsically appropriate
responses to the grosser forms of vice, which deserved to be thus satisfied; and we find
a contemporary version of this kind of view in Moore's claim that criminal law should
be understood as a functional kind, whose function is to achieve retributive justice by
punishing “all and only those who are morally culpable in the doing of some morally
wrongful action” (Moore 1997: 35). This is, as it stands, a wholly non-instrumentalist,
intrinsicalist account of the proper purpose of the criminal law: it has no purpose
beyond itself, beyond the punishment of culpable agents for their immoral conduct; it
does not even, apparently, aim to reduce the incidence of such conduct.
Moore offers what looks at first like an extreme and simple version of ‘Legal
Moralism’, the view that “all and only moral wrongs should be criminally prohibited”
(1997: 662), but in fact the implications of his account are less dramatic than this might
suggest. The immorality of a given kind of conduct creates a presumption in favour of
criminalizing it—it ‘should be criminally prohibited’. However, that presumption can
be defeated by other considerations to do with the impact of criminalization; in
particular, a proper regard for individual liberty will dissuade us from actually
criminalizing much wrongful conduct (see Moore 1997: ch. 18, and 2009).
There are, of course, other types of Legal Moralism than Moore's. Any version of Legal
Moralism claims that the immorality of a given kind of conduct is significantly relevant
to the question of whether it should be criminalized. We can then distinguish positive
from negative versions. Positive Legal Moralists hold that immorality is a good reason
for criminalization—not necessarily that it creates a presumption in favour of
criminalization, but that it provides a reason that should carry some weight in our
deliberations (see Feinberg 1984: 27; 1988: 324). Negative Legal Moralists hold instead
that immorality constitutes only a necessary condition for criminalization: we must not
criminalize conduct unless it is immoral, but its immorality does not give us any
positive reason to criminalize it. Negative Legal Moralism, like negative retributivism
(see Dolinko 1991: 539–43), acts as a side-constraint on our pursuit of the goals that
provide our positive reasons for maintaining a system of criminal law, whereas a

13
positive Legal Moralism helps to set those goals (see also Simester and von Hirsch
2011: ch. 2; Duff forthcoming). We should note too that a positive Legal Moralist as
defined here need not be a negative Legal Moralist: one can believe that immorality
provides a good reason for criminalization whilst also believing that there are other
reasons, including reasons for criminalizing conduct that is not immoral. (We will also
look later at the argument that whilst immorality as such provides no reason for
criminalization, immorality of the right kind does provide a good reason.) Furthermore,
even positive Legal Moralists need not think, with Moore, that the reason for
criminalizing immoral conduct is precisely and only to secure its retributive
punishment: she could instead believe, as Feinberg's Legal Moralist does, that we
should criminalize it in order to prevent it, and therefore only if criminalizing it would
be likely to reduce its incidence (see Feinberg 1988: 324).
It might seem that negative Legal Moralism is straightforwardly undermined by the fact
that many of the offences defined by a modern criminal law constitute mala
prohibita rather than mala in se.Mala in se, as normally understood, are crimes
consisting in conduct that is wrong independently of the criminal law—that would have
been wrong even had there been no criminal law. Mala prohibita, on the other hand,
consist in conduct that is not wrongful independently of the law that prohibits it: if they
are wrong, their wrongfulness depends essentially on their illegality. The distinction
between mala in se and mala prohibita is neither clear nor uncontroversial, but does
point to something important: malicious killing, for instance, is wrong, something that
we all have very good reason not to do, independently of the law of murder; by contrast,
there is nothing wrong with driving from north to south down a narrow street in the
absence of a regulation making it a south to north one-way street. However, Legal
Moralists can easily justify a category of mala prohibita. If the legislature is justified in
creating the kinds of regulation that mala prohibitainvolve (such as traffic regulations),
to serve some aspect of the common good, breaches of such regulations might be moral
wrongs (though their weongfulness will need to be shown, not just assumed); that
wrongfulness provides, for the Legal Moralist, a reason to criminalize such breaches—
to define them not merely as morally neutral breaches of a regulation, but as criminal
(see further Green 1997; Duff 2007: chs 4.4, 7.3; and, for criticism, Husak 2005, 2007:
103–119).
Were we faced by a stark choice between an instrumentalist view of the criminal law
and a view like Moore's, we might think that some form of instrumentalism has to be
right. Even if we think that, once we have a system of criminal law, we must justify
criminal punishment in retributivist terms (which is itself controversial), it seems much
less plausible to think that we should create and maintain the whole edifice of criminal
law simply in order to condemn and punish immoral actions; surely at least a central
part of the purpose of a modern system of criminal law must be to protect citizens
against various kinds of harm, by preventing kinds of conduct that cause such harm. We

14
must also ask whether positive Legal Moralists are right to believe that every kind of
immoral conduct is, in principle, the law's business—that even if in the end the balance
of reasons argues against criminalizing some kind of immoral conduct, its immorality
provided a good reason in favour of its criminalization. Suppose that I betray a friend
by frivolously revealing a secret that she had entrusted to me and that I knew mattered a
lot to her: I have done her a grievous moral wrong, and might indeed have fatally
damaged our friendship; but are we really to say that such conduct should (ceteris
paribus) be criminal, or that its immorality gives us good reason to criminalize it?
We will return to Legal Moralism (in § 6 below), but should note here that there are
also some serious questions for instrumentalists. A purely instrumentalist account faces
the same questions, the same moral worries, as does any purely consequentialist theory
of moral, social or political action: put crudely, the general worry is that any such
theory will fail to do justice to individuals and their rights, since it will too easily
sanction unjustly sacrificing individuals to the greater social goods that it posits as the
justifying aim of our actions. Side-constrained instrumentalists avoid that kind of
objection, since the side-constraints that they recognise are precisely intended to rule
out such injustice, such infringements of individual rights: but there is a serious
question about their conception of criminal law—whether we should see it simply as a
technique whose positive justification lies solely in its beneficial effects.
For instrumentalists, whether pure or side-constrained, it is a contingent, empirical issue
whether the criminal law is an appropriate institution: it is appropriate if and because it
does, as a matter of contingent fact, make an efficient contribution to whatever ends we
posited for the state. Now we can agree that this instrumental dimension is crucial to the
justification of a system of criminal law: we must surely believe, for instance, that a
system which on balance did more harm than good could not be justified. However
(leaving aside the question of what is to count as ‘harm’ or ‘good’ for the moment), it
does not follow from this that instrumental efficiency is the only positive justifying
reason for maintaining a system of criminal law: we could still also believe that such an
institution can be justified only if it can be shown to be an intrinsically appropriate way
of dealing with and responding to the kinds of conduct that fall within its proper ambit.
Consider what I will argue is an apt analogy. If I believe that a friend ought (morally) to
go to visit her sick aunt, I might try to persuade her to do so; and if she is initially
unwilling, I will then try to work out how I can best persuade her. What counts as ‘best’
here is in part an instrumental matter: I want to find something to say, or do, that will in
fact persuade her. But it is not just an instrumental matter, since I should rule out some
possibly effective means of persuading her to visit her aunt—bribery, for instance, or
blackmail, or deception. What rules such means out is not that they would not be
effective (they might well be), nor merely that they are inconsistent with some non-
consequentialist side-constraint, but rather that they are intrinsically inappropriate to the

15
end that I should be pursuing. For if I am to show my friend the respect that is due to
her as a moral agent (and as my friend), my aim must be not merely to persuade her, by
whatever means will be effective, to act as I think she ought to act: it must be to bring
her by a process of rational moral discussion to see for herself that that is how she ought
to act; but bribery, blackmail or deception cannot count as means to that end (see Duff
1986: 47–54).
To see how this is an apt analogy to the questions we are pursuing here, about the
proper aims of a system of criminal law, we must turn to a question which is not
addressed as often as it should be, about the voice—the tones, and the terms—in which
the criminal law should address those whom it claims to bind.

5. The Law's Voice

On some accounts, the law is not addressed to the citizens at all: it is, rather, addressed
to the courts, laying down what actions they should take (what punishments they should
impose, for instance) when certain conditions are satisfied (see Hart 1994: 35–38, on
Kelsen). Perhaps the law should also be made known to, or easily knowable by, the
citizens on whom it is liable to impinge, as a matter of fairness to them: but they are not
its direct addressees. Such a view is no doubt true for some aspects of law, including
some aspects of criminal law: laws that deal, for instance, with sentencing, or that
define various legal excuses, seem to be addressed to courts rather than to the citizens
(on the distinction implied here between ‘rules for courts’ and ‘rules for citizens’ see
Fletcher 1978: chs. 6.6–8, 7, 9; Dan-Cohen 1984; Robinson 1997). But it is not a
plausible view of law as a whole, or of the central, offence-defining aspects of criminal
law in particular: the law speaks to all of us, as citizens. We may hear its voice most
loudly, most dramatically, if we find ourselves as defendants in a criminal court, when
we are called to answer a charge of criminal wrongdoing, and to hear the law's
condemnation of our conduct if we are convicted: but in defining which kinds of
conduct are criminal, and which are legally permissible, the law speaks to all of us,
about what we may or may not do. (Some aspects of the substantive offence-defining
criminal law are not addressed to all citizens, but only to those engaged in particular
activities: only drivers are addressed by most of the road traffic laws, for instance, and
only those who deal in shares are directly addressed by the laws concerning insider
trading.)
In what tones and terms, then, does or should the criminal law address the citizens? One
view, familiar from the classical positivist theories of Austin and Bentham (see The
Nature of Law, § 2), tells us that the law, as addressed to the citizens, consists in a set of
commands or orders backed by threats to secure obedience from those who might
otherwise disobey. The law says to us “Don't do this!” (or, less frequently, “Do this!”);
and if we ask why we should obey that command, the answer will refer either to the

16
law's authority (“Because it is the law and you ought to obey the law”) or to its power
(“Because the law will make you suffer if you do not”)—though for classical positivists
like Austin and Bentham the law's authority seems to reduce to its power. That simple
positivist view of law is no longer widely held, but we can see a vestige of it in the very
widespread view that the substantive, offence-defining criminal law consists essentially
in a set of ‘prohibitions’ (rules that ‘forbid’ certain kinds of conduct), which citizens are
supposed to ‘obey’—which, indeed, they supposedly have an obligation to obey.
Now this might indeed be how the law's voice sounds to those who feel no allegiance to
the polity whose law it is, and it is how the law's voice should sound to those who
relationship to it and to the polity is that of oppressed subject to alien sovereign: the law
does speak to them in the threatening coercive tones of one who demands, and claims to
have the power to exact, their obedience. But it is not how the law should speak to the
citizens of a liberal polity (see further Duff 2001: 56–68). As citizens, we are members
of the normative community whose values the law purports to express: if it is to address
us as citizens, and as responsible agents, it must speak to us not in the peremptory,
coercive voice of a sovereign who commands our obedience, but in the rational,
normative voice of values which demand our allegiance as the values of our polity. The
law of a liberal polity, that is to say, must aim to be a common law: a law which
belongs to the citizens, as a reflection of the values they share, rather than a law which
is imposed on them by an alien sovereign (compare Cotterrell 1995: ch. 11).
The law, or the legislators who create and declare the law, must claim that there are
good reasons to criminalize the kinds of conduct it defines as crimes. Since to
criminalize conduct is to declare that it should not be done, that claim must be that there
are good reasons why the citizens should not engage in such conduct—reasons
reflecting the polity's values. If the law is to address us as responsible members of the
normative political community, it must address us in terms appropriate to those reasons.
In the example offered in § 4, I treat my friend as a responsible agent only if the reasons
I offer her for going to visit her aunt are of the right kind—the very reasons that, as I
see it, make it right for her to do this. Similarly, I am now suggesting, if the law is to
address us as responsible citizens, it must address us in terms that appeal to the right
kind of reason for refraining from the conduct that it defines as criminal: in terms that
appeal, that is, to the reasons which justified criminalizing such conduct in the first
place.
What kinds of reason could those be? We will return to this question in the following
two sections, but should note here that it will be hard to resist the initial conclusion that
they must be moral reasons, to do with the moral wrongfulness of the conduct that is
criminalized. For, first, the law's voice is an insistent one. It declares that these things
must not be done, even if (it implies) it might suit our individual interests to do them; it
attaches significant penalties to the conduct it criminalizes: how could such a voice be

17
justified other than by claiming that it is speaking to us of moral duties that we owe to
each other and to the polity? Second, the law speaks in terms that appear closely related
to the extra-legal languages of morals. It speaks of guilt, of fault, of culpability and
wrongdoing; it speaks of murder, rape, dishonesty, theft and the like: unless we are to
say that these terms are systematically ambiguous as between their legal and their extra-
legal uses (in which case the law would not be making itself accessible or readily
intelligible to its citizens), we must conclude that the law's definitions of offences are
meant to be legal definitions of moral wrongs—of kinds of conduct that are wrong
either pre-legally, as mala in se are; or as breaches of legal regulations which, once they
are created, citizens have a moral obligation to obey (see Green 1997). The criminal
law's definitions of offences will not always aspire to match precisely our extra-legal
understanding of the relevant moral wrongs: there will often be good reasons, to do
with the practical and moral constraints of law enforcement and the criminal process,
for the law's definitions to diverge from extra-legal moral understandings. But the law's
definitions must be grounded in those extra-legal moral understandings. What the
criminal law must say to the citizens is therefore not that they must refrain from such
conduct because the law forbids it and can demand their obedience, but that they should
refrain from such conduct because it is wrong.
Why should we maintain an institution that speaks to its citizens in such terms of
wrongs that should not be committed? Part of the reason is obviously to dissuade the
citizens (if they need dissuading) from committing such wrongs—that is the truth in the
instrumentalist view. Indeed, nothing said so far rules out the familiar suggestion that a
central purpose of a system of criminal law is to reduce the incidence of the relevant
kinds of wrongdoing by threatening those who might commit them with punishments
that will deter them—whether punishment should be justified as a deterrent is a further
issue. But this is not to say that instrumentalists are wholly right, or that Moore is
wholly wrong to think that the sole purpose of criminal law is to provide for the
retributive punishment of those who culpably commit such wrongs. For, first, even if
we are in the end justified in using punishment as a deterrent for those who will not
otherwise be dissuaded from crime, the law's initial appeal to the citizens must be in the
moral language of wrongdoing, not simply in the coercive language of deterrence
(see Legal Punishment, s. 6): not because such a moral appeal is likely to be
instrumentally effective, but because it is intrinsically appropriate to the law's dealings
with the citizens of a liberal polity. Second, we can now plausibly suggest that another
purpose of the criminal law is to provide a suitable response to criminal wrongs that are
committed. It publicly recognises and condemns them as wrongs by defining them as
crimes; it calls those who are alleged to have committed them to account, to answer for
that alleged wrongdoing, through a process of criminal trials; it condemns those who
are proved to have committed such wrongs by convicting them—and by punishing
them, if we understand punishment as involving the communication of censure (see

18
again Legal Punishment, s. 6). The truth in Moore's view is that such responses to crime
are justified not merely as instrumentally efficient means to the reduction of harmful
conduct, or to other further ends, but as intrinsically appropriate responses to the kinds
of wrongdoing that properly concern the criminal law. We must take such wrongdoing
seriously, if we take seriously the values against which it offends, the victim's standing
as one who has suffered such a wrong, and the wrongdoer's standing as a responsible
agent who has done wrong: but to take it seriously is to be prepared to declare it to be
wrong it, and to call to account and to condemn those who engage in it.
I have suggested in this section the central purpose of criminal law, as a distinctive kind
of law marked out from the other kinds and aspects of law by the features discussed in
s. 2, is to define, and to declare the wrongfulness of, certain kinds of wrongdoing, in
order not only to dissuade citizens from committing such wrongs, but also to provide
appropriate responses to those who commit, or are alleged to have committed, such
wrongs. In defining conduct as criminal, the law identifies it as conduct from which we
have good reason to refrain, and thus also as conduct for which we will be called to
public account, and condemned and punished, if we engage in it. To ask whether we
should have a system of criminal law is therefore to ask whether there are kinds of
wrongdoing that the state should identify and respond to in such a way—kinds of
wrongdoing that the state should take seriously as wrongdoing, and expect its citizens to
take similarly seriously.
But what kinds of wrongdoing could these be? We noted that the simple Legal
Moralist's claim, the claim that we have good reason to criminalize any kind of immoral
conduct simply in virtue of its immorality, seems implausible. My betrayal of my
friend, wrong though it is, does not seem like the kind of wrong that merits public
denunciation by the criminal law, or for which I should be called to account by the
whole polity through its criminal process; it is, surely, a private matter between me and
my friend (and perhaps the circle of friends to which we both belong), not a public
matter that concerns the state, or my fellow citizens as such.
This natural response to this example points us towards one common way of identifying
the kinds of wrong that do properly concern the criminal law—the idea that conduct
which is to be criminalized ought to constitute a ‘public’, rather than a merely ‘private’,
wrong.

6. Crimes as Public Wrongs

What makes simple Legal Moralism seem implausible is not just the thought that some
moral wrongs are not serious enough to attract the attention of the criminal law: though
that is true, the wrong I do my friend is not a trivial one—it might destroy our
friendship. Nor is it simply the thought that there are very good reasons against
criminalizing such wrongdoing—reasons that will forcibly strike us as soon as we begin

19
to think how a law criminalizing such conduct could be drafted and enforced: Legal
Moralists themselves will, as we have seen, argue that such countervailing reasons can
often outweigh the reasons in favour of criminalization; but my example was supposed
to make us doubt whether the moral wrongfulness of this kind of conduct constituted
any reason at all for criminalizing it. The objection to Legal Moralism is naturally
expressed by saying that this kind of wrongdoing is a ‘private’ matter that is simply ‘not
the law's business’ (Wolfenden 1957: para. 61).
Must we then reject every species of positive Legal Moralism outright, and insist that
the immorality of a kind of conduct is never in itself a good reason for criminalizing it?
If so, we might still be able to preserve some form of negative Legal Moralism, and
hold that conduct that is not immoral cannot properly be criminalized; but we would
need to look elsewhere for positive reasons for criminalization. This, however, seems
equally implausible—especially if the argument of the previous section was sound. For
I argued there that the criminal law must aim to identify and condemn kinds of morally
wrongful conduct, which implies that it is the very immorality of the conduct that gives
us reason to criminalize it; and it would anyway be very strange if the reasons for
counting such central mala in se as murder and rape as criminal had nothing to do with
the moral wrongfulness of such actions. There is, however, a third way between the two
extremes of holding that any kind of immorality provides a good reason for
criminalization, and holding that immorality itself never provides a good reason in for
criminalization: we can hold that immorality of the right kind provides a good reason
for criminalizing conduct that involves it. Moral wrongdoing that could in principle be
justifiably criminalized would thus form a subcategory of the larger category of moral
wrongdoing.
But how could we identify that subcategory? One familiar slogan is that the criminal
law is properly concerned only with ‘public’ wrongs, whereas merely ‘private’ wrongs
are either not matters for the law at all, or matters for a civil rather than a criminal legal
process (see s.2 above). To gain any help from this slogan, however, we need to know
what ‘public’ means in this context, and how to distinguish ‘public’ from ‘private’
wrongs; in the rest of this section, I will explore some suggestions (see also The Limits
of Law; Lamond 2007).
A pure instrumentalist will of course argue that this is a pragmatic issue: supposing that
we do have good reason to criminalize only conduct that is in some way immoral, we
decide which kinds of immorality to criminalize, and which to deal with in other ways
(or to ignore) by asking which techniques are likely to constitute efficient means to our
preferred ends. The range of possible techniques is very wide, even if our only aim was
to reduce the incidence of such conduct (which it would not be): extra-legal techniques,
such as education, advertising, and situational crime prevention (on which see von
Hirsch, Garland & Wakefield 2000); the taxation system (we might as efficiently

20
reduce the incidence of a certain kind of conduct by taxing it as by criminalizing it); and
the threat of civil liability to pay damages for any harm that was caused. But an
implication of the argument of the previous section is that the choice between these
techniques should not be a purely pragmatic one: although issues of efficiency are
clearly relevant and important, we must first ask which kinds of measure are
intrinsically appropriate to the kind of conduct, to the kind of wrong, we are dealing
with. We have to ask, that is, whether the conduct in question involves a kind of
wrongdoing that merits the public calling to account and condemnation of those who
engage in it that the criminal law involves: if it does, we then have at least a good
reason to criminalize it; if it does not, we do not. This might be only the first stage in a
long and complex deliberation about whether a given type of wrongful conduct should
be criminalized, and pragmatic issues would certainly need to loom large in later stages
of that process: but my suggestion here is that it is an essential first stage. (Compare the
‘filtering’ model of how we should decide questions of criminalization offered by
Schonsheck 1994. See also Ashworth 2006: chs. 2–3, on the range of ‘principles and
policies’ that should bear on questions of criminalization; and Husak 2007 for a very
useful general discussion).
So what should count as a ‘public’ wrong? A first and familiar suggestion, implicit in
the way I have talked throughout of criminalizing ‘conduct’, is that ‘mere thought’
should not be criminalized: for thought is private; only action or conduct is public. This
suggestion captures at least one central aspect of the slogan that criminal liability
requires an act (or a ‘voluntary act’, as it is sometimes put), and it does seem to have
some force: although mere thoughts can be morally improper (entertaining sadistic
fantasies about my opponents, for instance), surely only what actually impinges on our
shared social or material world can properly be of interest to the state or its criminal
law; but mere thought which is not expressed or acted upon has no such impact. We
must note, however, first, that this sets only extremely modest limits on the scope of the
criminal law—it does not even protect speech from criminalization, since speech
certainly has an impact on the world. Second, if the ‘act requirement’ is to do any
substantive work, we need an account of the concept of ‘action’ or ‘conduct’—an
account which will also need to deal with such questions as that of criminal liability for
omissions (see Hughes 1958; Feinberg 1984: ch. 4), or for statuses such as being drunk
or an addict (see Glazebrook 1978): such an account is notoriously difficult to provide.
(For contrasting views on the act requirement see Moore 1993: chs. 2–3; Husak 1998;
Duff 2007: ch. 5.)
Within the realm of ‘conduct’ as distinct from thought, and leaving aside the question
of omissions, how might we try to identify the subcategory of ‘public’ wrongs? Another
familiar suggestion is suggested by Mill's Harm Principle—“the only purpose for which
power can be rightfully exercised over any member of a civilised community against
his will is to prevent harm to others” (Mill 1859, ch. 1, para. 9): could we not say that

21
only conduct that wrongfully harms or threatens to harm others is a suitable candidate
for criminalization; that the criminal law is properly concerned only
withharmful immorality? (See generally Feinberg 1984; Holtug 2002; Duff 2007: ch. 6;
Persak 2007; Baker 2011; Simester and von Hirsch 2011: chs. 3–5; see also The Limits
of Law.)
An initial question about the Harm Principle concerns its focus. In its classical
formulation by Mill (see also Feinberg 1984: 26), what justifies criminalizing a
particular kind of conduct (or, for Mill, other kinds of coercive intervention) is that
doing so will prevent harm to others. As the Principle is often applied by its proponents,
however, it justifies criminalizing a given kind of conduct only if that kind of conduct
itself causes (or threatens to cause—see below) harm to others. Very often, of course,
these two versions of the Principle will produce the same results: we efficiently prevent
harm by criminalizing harmful conduct. But they can diverge, especially in relation to a
range ofmala prohibita that are created to improve safety. For instance, in order to show
that using the criminal law to enforce strict speed limits on the roads is fully consistent
with the second version of the Harm Principle, we would need to show that every driver
who exceeds the speed limit thereby creates an unreasonably increased risk of harm—
which seems implausible. To show it to be fully consistent with the first version,
however, we would need to show only that the enforcement of strict speed limits is
itself a more efficient way of preventing harm than alternative kinds of regulation (for
instance than a law that criminalized only dangerous driving, or one that allowed highly
competent drivers to exceed the speed limit); that seems a more plausible argument to
make.
Whichever version of the Harm Principle is offered, we also clearly need an account of
the concept of harm itself, which raises some difficult questions (see e.g. Kleinig 1978;
Feinberg 1984; Raz 1986: ch 15,1987; Simester and von Hirsch 2011: ch. 3). For
instance, can we plausibly so define ‘harm’ as to rule out the argument that any
immoral conduct is itself harmful, either to those whom it wrongs (since I am harmed
by being wronged), if it does wrong anyone, or by doing moral harm to the society or
culture in which it is done (see Devlin 1965; Dworkin 1994)? Can we so define ‘harm’
as to exclude purely trivial wrongdoings from the scope of the criminal law, for
instance by focusing on setbacks to ‘welfare interests’ (Feinberg 1984: ch. 1)? We
cannot pursue these issues here (nor the question of whether it is only harm to
others that can justify criminalization—the question of paternalist legislation: see
Feinberg 1986; Simester and von Hirsch 2011: chs. 9–10); but we should note two ways
in which the Harm Principle fails to set any very tight constraints on the scope of the
criminal law.
First, as noted above, any plausible version of the Harm Principle must permit the
criminalization of conduct that threatens, as well as conduct that actually causes, harm:

22
whether our interest is in preventing harm or in criminalizing harmful conduct, we have
good reason to criminalize conduct that creates a serious and unjustifiable risk of harm
to others, even if that risk is not in fact actualised. Even the more limited category of
conduct that causes harm becomes problematic when we ask what counts as causing
harm (see Simester and von Hirsch 2011: chs. 4–5): our existing laws criminalize not
only conduct that immediately causes harm, but also kinds of conduct whose causal
relationship to the harm is more remote—for instance conduct that enables or assists
another's commission of a crime; so we need to ask how far we should extend the law's
reach in this direction. The problems multiply when we turn to conduct that, whilst it
does not actually cause harm, is criminalized because it threatens harm, or creates a
danger of harm: to the broad category of ‘nonconsummate’ offenses (see Husak 1995).
This category includes conduct that is intended to cause harm (attempts to commit
crimes, for one obvious instance); conduct that, though not intended to cause harm,
recklessly or negligently endangers others (dangerous driving, for instance, or breaches
of health and safety rules); and conduct that, although it might not itself be dangerous,
is of a kind that is usually dangerous (speeding, for instance). Again, we need to ask
how far the criminal law should reach in this direction (see Harcourt 1999; also Dubber
2001).
Second, even if we limit our attention to conduct that obviously and directly causes
what must surely count as non-trivial harm, not all such conduct seems even in
principle to be apt for criminalization. The friend whom I betray would surely count
herself as seriously harmed by my betrayal; someone whose spouse betrays their
marriage by committing adultery might reasonably claim to be seriously harmed by that
betrayal. But we should not therefore see any good reason to criminalize betrayals of
friendship; and although adultery is indeed a crime in some states, arguments about
whether it is in principle apt for criminalization do not typically depend on claims that it
is, or is not, harmful to the betrayed spouse—the reality of that harm is taken for
granted. Nor will it help to suggest, for instance, that the kinds of wrongful harm that
are apt for criminalization are those and only those that set back ‘welfare interests’ (see
Feinberg 1984: 37–38, 61–63): if our welfare interests concern those goods that
constitute the ‘basic requisites of [our] well-being’ (Rescher 1972: 6, quoted
approvingly in Feinberg 1984: 37), they surely include such goods as friendship and
other loving relationships.
Some argue that, given the indeterminacy of the notion of harm, and the ease with
which legislatures can stretch the criminal law to cover kinds of conduct only remotely
connected to the harms that supposedly warrant their criminalization, the Harm
Principle can do no substantial work in guiding or constraining the scope of the criminal
law (see especially Harcourt 1999). Even if it can do some substantial work in limiting
the scope of the criminal law, it seems clear that it is still over-inclusive, since there are
kinds of seriously harmful conduct that we do not think should, even in principle, be

23
criminalized. Others argue that it is also under-inclusive, since we have good reason to
criminalize kinds of conduct that do not cause or threaten harm (and whose
criminalization is not aimed at preventing harm): conduct, for instance, that causes
offence rather than harm (see Feinberg 1985; Simester and von Hirsch 2006, and 2011:
chs. 6–8); conduct that violates another's ‘sovereignty’ (see Ripstein 2006), or ‘dignity’
(see Dan-Cohen 2002), or rights (see Stewart 2010); or conduct that is so egregiously
immoral that, even if it causes no harm, we have reason to criminalize it (see Feinberg
1988; Dworkin 1994).
Perhaps, if crimes are ‘public’ wrongs, we could meet the charge of over-inclusiveness
by distinguishing within the category of wrongfully harmful conduct between public
and private harms: the betrayal of a friendship or a marriage is not apt for
criminalization if and because it causes only private harm. But what is to count as a
public harm, or public wrong?
On one familiar reading, a wrong or harm is ‘public’ if and because it affects, i.e.
wrongs or harms, ‘the public’, rather than only an individual victim (see Blackstone
1765–9: Bk. IV, ch. 1, 5); wrongs or harms that affect only individual victims are—if
they are matters for the law at all—appropriately pursued by those individual victims
through the civil courts.
We can understand some crimes as harming or wronging ‘the public’—‘the public’
being understood either a set of individuals among whom we cannot identify
determinate individual victims, or as a collectivity with shared goods that crime
impairs. Three examples will serve to illustrate this point.
First, ‘public order’ offences involving violent, riotous conduct are injurious to the
public in that they pose a threat of serious harm to any of the indeterminate number of
individuals in the area, and might threaten to undermine that shared sense of assured
security on which our civic life depends. (Compare Braithwaite & Pettit 1990: 60–68,
on ‘dominion’ as the central civic good.) There is of course enormous scope for the
political abuse of public order laws—see Lacey et al. 2003: ch. 2; but our concern here
is with the issue of whether we have any in principle reason to define a category of
public order crimes. Similarly, offences of endangerment that involve no ‘disorder’
(driving offences which endanger other road users generally, for instance, and offences
involving public health and safety) often threaten the public, rather than determinate
individuals.
Second, some crimes attack or threaten the polity's own institutions, and thus threaten
or harm ‘the public’ as a collectivity. This category includes such crimes as perjury,
attempts to pervert the course of justice, the offering of bribes to, or their acceptance by,
public officials, and various kinds of electoral malpractice. In some such cases a
determinate individual might be wrongfully harmed—an innocent person might be

24
wrongly convicted, or might lose a civil case, because a witness commits perjury: but
whether or not we can identify any such individual victim, the crime attacks a public
institution which is crucial to the public interest.
Third, other kinds of wrongful conduct are apt for criminalization because they involve
serious unfairness towards one's fellow citizens. Someone who evades their taxes might
cause no identifiable consequential harm, either to any individual or to the social
institutions which are funded by taxation; if asked to explain the wrong she commits,
we would appeal to some version of ‘What if everyone did that?’, rather than trying to
identify any consequential harm that she causes. We would appeal, that is, to the unfair
advantage that she takes over all those who pay their taxes: she gains the benefits that
accrue to all citizens from the taxation system, but refuses to make her appropriate
contribution to that system. (Note, however, that the appeal here seems to be to
unfairness rather than to harm, unless we can say that such wrongs as tax evasion are
criminalized because they are aggregratively harmful: see Feinberg 1984: 193–9.)
So we can explain why some kinds of conduct are properly criminalized by showing
how they wrong or harm ‘the public’, or ‘the public interest’: but can we
explain all crimes in this way? Can we sustain the general claim that conduct should, in
principle, be criminalized only if and because it wrongs or harms ‘the public’ in this
sense? There are two ways in which we might try to do this—by appealing either to the
idea of public order and stability, or to that of unfairness.
Consider, first, the idea of public order, and the suggestion that the criminal law's
proper purpose is to protect the “smooth functioning of society and the preservation of
order” (Devlin 1965: 5). We find relatives of this suggestion in Becker's argument
(1974) that the criminal wrongfulness of crimes consists in their tendency to cause
‘social volatility’, and in Dimock's argument (1997) that it lies in their tendency to
undermine the kinds of trust upon which civic life depends. What makes crimes—
including such crimes as murder and rape—wrongful in a way that properly concerns
the criminal law is, on such accounts, not the wrongful harm that they do to their
immediate individual victims, but their wider effects on social stability or trust.
Consider second the idea of unfairness. According to one well-known theory of
punishment (see Murphy 1973; Dagger 1993), crimes deserve punishment because the
offender takes an unfair advantage over all his law-abiding fellow citizens: he accepts
the benefits of their law-abiding self-restraint (the mutual security provided by an
effective system of law) but refuses to make his proper contribution to that system by
exercising such self-restraint himself. Might we ground a theory of criminalization on
such a theory of punishment? We should criminalize murder, rape and other
central mala in se because, apart from the wrongful harm that they do to their individual
victims, they wrong ‘the public’ (the generality of law-abiding citizens) by taking unfair
advantage of them.

25
The obvious objection to such ways of explaining the idea of crimes as public wrongs
or harms is that, precisely by portraying crimes as wrongs done to ‘the public’, they
distort their character as wrongs that merit criminalization. We are now to criminalize
murder or rape, not because of the wrongs that they do to their individual victims, but
because of their effects on social stability or trust, or the unfair advantage they take over
the law-abiding; from which it follows, if the criminal law should address the citizens in
terms of the reasons and values that inform its definitions of crimes (see s. 5 above),
that a murderer or rapist is to be condemned and punished not for what he did to his
individual victim, but for acting in a way that created social volatility, or undermined
trust, or took unfair advantage over his law-abiding fellows. This is surely not how we
should understand the criminal wrongfulness of such crimes.
To illustrate this point, consider the example of domestic violence and abuse (on
domestic violence, see Dempsey 2009). In English law intra-marital rape was
recognised as a crime only in 1991 (see R[1991] 4 All ER 1981); until then, a husband
who forced sexual intercourse on his wife without her consent was not guilty of rape.
Similarly, although domestic violence (typically husbands violently beating up their
wives) was formally speaking a crime, it was often not taken seriously as a crime by the
criminal justice system: the police were often unwilling to intervene in ‘domestic
disputes’ or to prosecute domestically violent men, seeing it rather as an issue for the
couple to work out for themselves. No doubt part of what lay behind these practices was
a view that the wrongs being done were, if wrongs at all, not that serious; but we can
also discern the view that these were ‘private’ rather than ‘public’ wrongs. If we then
ask what justified the change towards recognising intra-marital rape as genuine,
criminal rape, and domestic violence as a genuine crime that should be prosecuted, it is
not plausible to answer in terms of either of the accounts noted above. It would not be
plausible to argue that domestic violence or intra-marital rape is liable to create social
volatility (indeed, such crimes are often committed by men who in their lives outside
the home are models of peaceful conformity); or that it undermines the kinds of trust on
which social life depends (such crimes, if confined to the home, do not undermine the
trust that we can have in our dealings with strangers, which is the kind of trust that is
relevant here); or that it takes unfair advantage over all those law-abiding people (or
men) who refrain from these or other kinds of crime—as if the law-abiding would love
to commit such wrongs if only they were not restrained by the demands of fairness.
To explain why such domestic abuse should be criminal, and taken seriously as
criminal, we must look not for some ‘public’ harm or wrong that it involves distinct
from the wrongful harm it does to its individual victims, but at that wrongful harm
itself. What matters is that we come to see the wrongs suffered by abused wives not just
as their private business, but as our collective business as citizens of a polity to which
we belong with them; we come to recognise that they have as strong a claim to the
protection and support of their fellow citizens as do the victims of attacks by

26
strangers—a claim grounded simply in their fellow membership of the polity, as our
fellow citizens. The wrongs done to them are ‘public’ wrongs not because they wrong
the ‘public’, but because they are wrongs that properly concern the public—their fellow
citizens; even when they are committed in what might count, empirically, as ‘the
privacy of the home’, they belong in what should count, normatively, as the ‘public’
realm. We could then also say, if we wish, that such wrongs are wrongs against, or
injurious to, the public—the polity and its members: they implicitly deny the core
values by which the polity defines itself, and the basic normative bonds by which we
define our civic relationships with each other; they are wrongs not just against their
individual victims, but against all of us insofar as we identify with those victims as our
fellow citizens—they are wrongs in which we collectively share, and which we make
‘ours’ (see Marshall & Duff 1998). But to talk in this way of ‘public’ wrongs or injuries
is not to try to ground the claim that such wrongs should be criminal: the appeal to the
idea of a ‘public’ wrong now expresses, rather than trying to ground, the claim that it is
a wrong that concerns us all, and that is therefore apt for criminalization.
If that is right, however, we cannot look to the idea of public wrongs or harms to
provide criteria or principles of criminalization. We can say that the criminal law should
be concerned with ‘public’, rather than ‘private’, wrongs, but that is because to call a
wrong ‘public’ in this sense is already to class it as a kind of wrong that is apt for
criminalization—a kind of wrong which should be publicly denounced and whose
perpetrators should be publicly investigated, prosecuted, condemned and punished; a
kind of wrong whose perpetrators should be called to account by the polity as a whole,
not just by the individual victim.
The upshot of this section is that we still lack any clear criteria or principles by appeal
to which we can try to determine which kinds of conduct should be criminal. However,
though this might be frustrating, we should at least by now be clearer about what kinds
of claim we must be able to make about kinds of conduct that we want to show are apt
for criminalization. First, we must be able to show that and how they involve
wrongdoing: for as we saw in s. 2, the criminal law focuses on wrongs that should be
condemned, rather than just on harms that need to be repaired or compensated; and as
we saw in s. 5, the criminal law must speak to us of wrongs that we should not commit.
Second, we must be able to claim that the wrong is of such a kind that it should concern
us all as citizens—we should not leave it to the individual victim to pursue, or not to
pursue, a civil case against the wrongdoer. I have not suggested determinate criteria by
which we can identify such wrongs, nor do I think that any determinate criteria can be
provided; although theorists might yearn to find a single principle, or a single set of
principles, by reference to which we could determine the (in principle) proper scope of
the criminal law, such yearnings are doomed to be frustrated (see Husak 2007; Duff
2005). We can, however, identify the main kinds of consideration that should be
relevant. Is the wrong one that injures ‘the public’ rather than any individual victim? Is

27
it one that flouts or implicitly denies the core values by which we define ourselves as a
polity, and which supposedly underpin our civic relationships? Is it one from which we
should be able to expect the protection of our fellow citizens (which is to ask whether it
is a wrong from which we should be able to expect to be categorically safe as we go
about our normal lives, rather than a kind of wrong that we can be expected to risk on
condition that we can seek compensation if we suffer it)? Answers to these questions
will be contestable, and will properly emerge only from a collaborative attempt to
understand what joins us as citizens and what we owe to each other as citizens—an
attempt which will lead to different results in different political communities: but we
have made progress if we have at least identified more clearly the questions that we
must ask.
It might also be argued, however, that the line of thought sketched above begins in the
wrong place. It begins with a supposedly pre-legal, perhaps even pre-political category
of wrongdoing, and then asks how we can determine which kinds of wrong within that
(very large) category we have reason to criminalize; one answer to that question is that
we have reason to criminalize those wrongs that count as ‘public’. Now there is room
for argument about the very possibility of such a starting point (see Farmer 2010); but
even if it is possible, it might be the wrong place to start. Perhaps, taking seriously the
political character of criminal law as a state institution, we should instead begin with
the ‘public’: with an account of the state and its proper aims, and of the institutions that
should be created and maintained to serve those aims for different versions of this
suggestion (see Thorburn 2011a, 2011b; Farmer forthcoming); or with an account of the
political community and its shared or civic life (see Duff forthcoming). We might then
still insist that the criminal law's proper purpose is to define, and to provide for an
appropriate response to, certain kinds of wrong: but those wrongs will now be identified
within, and take their criminalizable character from their place within, the political and
institutional structure of the polity.

7. International Criminal Law

We have focused so far, explicitly or implicitly, on the domestic criminal law of nation
states—a criminal law that typically claims jurisdiction only over crimes committed
within the state's territory, by and against citizens of or visitors to that state (for some of
the complications here, see Hirst 2003; Duff 2007: ch. 2.2). However, recent years have
also seen significant developments in international criminal law, culminating in the
creation of the International Criminal Court in 2002 (see Rome Statute of the
International Criminal Court 1998; Cassese 2008: 317–335). Behind a range of
particular questions about the proper scope of international criminal law (over what
kinds of crime should the ICC have jurisdiction?), about the relationship of the ICC to
domestic courts and systems of criminal justice, and about the appropriate procedures
for international criminal trials, lies a deeper question about the moral authority or

28
legitimacy of any such court: by what moral right does it claim jurisdiction over this
range of wrongs and over these wrongdoers? This question becomes particularly acute
in relation to crimes whose impact is intra- rather than inter-national. Crimes committed
by one state against another, or against another's population (war crimes, the crime of
aggression; Rome Statute arts. 5, 8), are clearly better dealt with by an international
court: but why should such a court have jurisdiction over ‘widespread or systematic
attack[s]’ committed within a state (very often by or with the connivance of the state's
officials) against its own citizens (Rome Statute art. 7, defining ‘crimes against
humanity’)? What could give such a body the moral authority to hold those who
commit such wrongs to account—and just what kinds of wrong can it claim as its
business?
One answer to these questions appeals to the impersonal demand of retributive justice
that those who commit such wrongs should not escape punishment: the ICC acts in the
name of justice. Another answer is that the ICC acts in the name and on behalf of the
more local polity within and against which the crimes were committed, when the
domestic courts will not or cannot act effectively. Another answer takes seriously the
idea of ‘crimes against humanity’, and portrays the ICC as acting in the name and on
behalf of ‘humanity’: those who commit such wrongs must answer not merely to their
particular polities, but to humanity itself, since their crimes ‘deeply shock the
conscience of humanity’, and are ‘of concern to the international community as a
whole’ (Rome Statute, Preamble). Each of these answers is problematic: a central task
for theorists of criminal law is to work towards a clearer understanding of the questions
to which such answers are offered (see generally Altman & Wellman, 2004; Luban,
2004, 2010; May 2005; Renzo 2012, 2013).

8. The Internal Structure of the Criminal Law

Once we have articulated an account of the proper aims and limits of the criminal law,
we can tackle various issues about its internal structure—about the general principles
and conditions of criminal liability (the so-called ‘general part’ of the criminal law), and
about its definitions of specific offences (the ‘special part’). I cannot pursue these issues
here, but our approach to them must clearly be based on our account of the proper
purposes of the criminal law. Thus if the criminal law should aim to define, to
condemn, and to call perpetrators to public account for wrongs whose character and
implications are such that they properly count as ‘public’, its definitions of crimes and
its principles of liability must be apt to identify such wrongs and the conditions under
which agents can be justly condemned for them. In discussing such issues as whether
the law should contain offences of ‘strict liability’, which can be committed by those
who are not even negligent as to the harm they cause or risk; or what kind of ‘fault
elements’ should be required for criminal liability, either in general or for particular
crimes; or whether criminal liability should depend solely on the ‘subjective’ character

29
of an action (on what the agent intended to do or believed herself to be doing), or also
on its ‘objective’ character (its actual connection with and impact on the world); or
whether and how the law should distinguish offences from defences, either
substantively or procedurally; or what kinds of excuse or justification the law should
recognise, and how they should be defined: we must first ask which doctrines and
principles would capture the relevant kinds of wrongdoing and identify the culpable
agents of such wrongs. This can only be the first stage of the discussion, since we will
need to go on to ask whether those doctrines and principles can satisfy the wide range
of other normative and practical constraints that must bear on a system of criminal law
and justice, or how they could be adapted so as to satisfy those constraints: but it is
where we must start. (For introductions to these issues see Fletcher 1978; Robinson
1997; Tadros 2005; Duff 2007; Gardner 2007; Alexander and Ferzan 2009; Ashworth
2009; Simester et al 2010; Dressler 2012.)
Theorists of criminal law must also attend, however, more seriously than they have
often attended, to the criminal process that leads from (alleged) crime to punishment,
and in particular to the criminal trial. It is through the criminal trial that criminal
responsibility and liability are formally assigned, and the norms and doctrines of the
substantive criminal law are articulated and applied. We therefore need an account of
the proper aims and values of the criminal trial, and of the larger criminal process of
which it is part—an account that can then underpin a more adequately grounded
critique of our existing criminal processes. Should we, for instance, see the trial as an
attempt to establish the truth (but what truth?), albeit an attempt that is constrained by a
range of independent principles and rules that aim to protect us against the potentially
oppressive and intrusive power of the state? Or should we see it as a process through
which alleged wrongdoers are called to answer the charges that thay face and to answer
for their crimes if their guilt is proved? What kinds of criminal process are appropriate
to a liberal democracy that aims to treat all its members as responsible citizens? (See
generally Burns 1999; Duff et al 2007; Ashworth & Redmayne 2010; Dzur 2012.)

9. Normative Theory and Empirical Reality

To those with first hand knowledge of the empirical realities of our systems of criminal
justice, normative theorising of the kind displayed in previous sections is likely to seem
so far removed from those realities that its practical relevance and utility come into
serious doubt. The quick response to such doubts about normative theory is to say that
what such drastic gaps between theory and existing practice show is not the inadequacy
or irrelevance of the theory, but the radical imperfection of the practice: that the proper
way to try to close the gap is not to revise the theory so as to bring it closer to existing
practice, but to seek to reform and improve existing practice so that it comes close to
what, according to the theory, it ought to be. Some such response is certainly warranted:
the proper aim of normative theory is not to describe or to justify existing practices, but

30
to articulate the goals, principles, and ideals against which those practices are to be
assessed; and it should not surprise us when such assessments show our practices to be
seriously defective. However, normative theorists must at least be aware of the nature
and extent of the gaps between theory and practice; and they must be alert to the
possibility that the gap might become so dramatic that it is no longer clear whether the
theory can be seen as a theory of that practice. We should note here three aspects of that
gap between theory and practice, in the context of criminal law, which should give
theorists pause for serious thought (see generally Ashworth 2000; Husak 2007: ch. 1;
Stuntz 2011).
First, criminal law theorists often focus their discussion on a fairly small range of
traditional ‘mala in se’ offences involving conduct that can plausibly be seen as being
pre-legally wrongful, and that typically either actually brings about or is very closely
related to the mischief that warrants the conduct's criminalization. Thus we discuss such
crimes as murder, rape, and other attacks on the person, as well as such property
offences as theft, fraud and criminal damage; and we discuss completed versions of
these crimes, or ‘non-consummate’ offences that are closely linked to those complete
offences, such as attempts to commit those offences. Such offences also typically
involve a substantial ‘fault element’, in that they are committed only if the agent brings
about the relevant effect (death or injury to the person, for instance) intentionally, or
knowingly, or at least recklessly: theorists can therefore focus their discussions of
criminal fault on the proper analysis of these kinds of fault and how they should be
proved. (But see Wells and Quick 2010 for a salutary exception.) However, a vast range
of criminal offences in contemporary systems of criminal law do not fit this classical
paradigm. Many of them are ‘regulatory’ offences, which consist in the breach of some
legal regulation rather than in conduct that could be seen as pre-legally wrongful (see
Horder forthcoming); many involve conduct quite remote from the mischief with which
the law could plausibly be supposed to be concerned (see e.g. Dubber 2001); and many
impose liability that is ‘strict’ in the sense that one can be convicted without proof of
intention, knowledge or recklessness as to all the elements of the offence (see Ashworth
and Blake 1996; Simester 2005). Such offences challenge the classical principle that
criminal liability should depend on proof of culpable wrongdoing, since at least many
of them seem to impose liability without any such proof. A normative theory of
criminal law that is to deal with contemporary criminal law must have something to say
about such offences: it must be able to show that and how at least some of them can be
justified as imposing criminal liability for proven culpable wrongdoing; or argue that
they have no place in the criminal law—and confront the dramatic implications that
such an argument would have for our existing systems of criminal law; or show how
classical theory can and should be adapted to accommodate such offences.
Second, both the theory and the rhetoric of criminal law often talk as if criminal liability
is properly imposed only given proof of culpable wrongdoing—proof that is supposedly

31
provided and tested at a criminal trial. The the Presumption of Innocence, that ‘golden
thread’ that supposedly runs through the criminal law, is typically formulated as the
principle that anyone accused of a criminal offence must be presumed to be innocent
until guilt is proved; and if one asks how guilt is to be proved, the obvious answer is
that it must be proved at a criminal trial by the prosecution (see e.g. Article 6 of the
European Convention on Human Rights; Woolmington v DPP [1935] A.C. 462, at p.
481). The vast majority of criminal convictions, however, involve no such process of
public proof: instead, the defendant pleads ‘Guilty’ to the charge, thus relieving the
prosecution of the burden of proof; and such pleas are very often the result of a plea-
bargain between defence and prosecuting counsel. A complete normative theory of
criminal law must include an account of the criminal process, and any such account
must tackle the problem of plea-bargaining, which seems inconsistent with the idea that
criminal liability should be imposed only when the truth and justice of the criminal
charge have been publicly tested in a criminal trial (see Lippke 2011).
Third, the borders between criminal law and other modes of legal regulation or control
are being increasingly eroded by practices of ‘preventive justice’. Instead of directly
criminalizing conduct that brings about (or increases the risk of) a relevant kind of
mischief, and convicting and punishing those who engage in such conduct,
governments seek more effective ways of preventing such conduct by imposing legal
constraints on those thought likely to engage in it—constraints that are often themselves
backed by the criminal law, in that it becomes a criminal offence to violate them. Good
examples of such measures are the Antisocial Behaviour Orders that English courts can
impose on those accused of various kinds of antisocial conduct (due to be replaced by
Criminal Behaviour Orders), and the various restrictive orders that courts can impose
on those suspected of involvement in terrorist activity (see Terrorism Prevention and
Investigation Measures Act 2011). The structure of this kind of provision is that there is
an initial, formally non-criminal process, in which a court is given reason to believe that
a person has been engaged in, and/or is likely in the future to engage in, some kind of
undesirable, usually criminal, activity (antisocial behaviour; terrorism), and that it is
necessary to subject him to restrictions in order to prevent (or to reduce the risk of)
future behaviour of that kind. The court can then impose a range of restrictions: on
where the person may go or when he may travel (including imposing a curfew), on
whom he may meet, and on a range of activities in which he might otherwise engage.
Once the restrictive order is made, it is a criminal offence to breach it. A normative
theory of criminal law must also have something to say about these kinds of measure.
Are they, or can they be rendered, consistent with the principles of justice and legality
that are supposed to structure the criminal law; or do they mark the subversion of those
principles in the interests of effective prevention? (See generally Ashworth and Zedner
2010, 2011, 2012; Ramsay 2012; Ashworth, Zedner and Tomlin 2013.)

32
 Bibliography
Ashworth 2009, Simester et al. 2010, and Wells & Quick 2010 are useful introductory
texts on criminal law; Duff & Green 2011 is a good collection of philosophical essays.

 Alexander, L., and K.K. Ferzan (with S. Morse), 2009, Crime and Culpability: A
Theory of Criminal Law, Cambridge: Cambridge University Press.
 Altman, A., and C.H. Wellman, 2004, “A Defense of International Criminal
Law”, Ethics, 115: 35–67.
 Ashworth, A. J., 2000, “Is the Criminal Law a Lost Cause? ”, Law Quarterly Review,
116: 225–56.
 –––, 2009, Principles of Criminal Law, 6th edition, Oxford: Oxford University Press.
 Ashworth, A. J., and M. Blake, 1996, “The Presumption of Innocence in English
Criminal Law”,Criminal Law Review, May: 306–317.
 Ashworth, A. J., and M. Redmayne, 2010, The Criminal Process, 4th edition, Oxford:
Oxford University Press.
 Ashworth, A. J., and L. Zedner, 2010, “Preventive Orders: A Problem of Under-
Criminalization?”, in R.A. Duff, et al. (eds.), The Boundaries of the Criminal Law,
Oxford: Oxford University Press, 59–87.
 –––, 2011, “Just Prevention: Preventive Rationales and the Limits of the Criminal
Law”, in Duff and Green (eds.) 2011a, 279–303.
 –––, 2012, “Prevention and Criminalization: Justifications and Limits”, New Criminal
Law Review, 15: 542–71.
 Ashworth, A. J., with L. Zedner and P. Tomlin, 2013, Prevention and the Limits of the
Criminal Law, Oxford: Oxford University Press.
 Baker, D. J., 2011, The Right Not to be Criminalized, Farnham: Ashgate.
 Becker, L., 1974, “Criminal Attempts and the Theory of the Law of
Crimes”, Philosophy and Public Affairs, 3: 262–94.
 Bianchi, H., 1994, Justice as Sanctuary: Toward a New System of Crime Control,
Bloomington: Indiana University Press.
 Blackstone, W., 1765–9, Commentaries on the Laws of England, 4 volumes, Oxford:
Clarendon Press. [available online].
 Boston University Law Review, 1996, “Symposium: The Intersection of Tort and
Criminal Law”, Volume 76: 1–373.
 Braithwaite, J. and P. Pettit, 1990, Not Just Deserts, Oxford: Oxford University Press.

33
 Brown, D., 2009, “ History's Challenge to Criminal Law Theory”, Criminal Law and
Philosophy, 3: 271–87.
 Burns, R. P., 1999, A Theory of the Trial, Princeton, NJ: Princeton University Press.
 Cassese, A., 2008, International Criminal Law, 2nd edition, Oxford: Oxford University
Press.
 Christie, N., 1977, “Conflicts as Property”, British Journal of Criminology, 17: 1–15.
 Cotterrell, R., 1995, Law's Community, Oxford: Oxford University Press.
 Dagger, R., 1993, “Playing Fair with Punishment”, Ethics, 103: 473–88.
 –––, 2008, “Republican Punishment, Consequentialist or Retributive?”, in C. Laborde
& J. Maynor (eds.), Republicanism and Political Theory, Indianapolis: Wiley, 219–45.
 –––, 2011, “Republicanism and the Foundations of Criminal Law”, in Duff and Green
(eds.) 2011a, 44–66.
 Dan-Cohen, M., 1984, “Decision Rules and Conduct Rules: On Acoustic Separation in
Criminal Law”, Harvard Law Review, 97: 625–77.
 Dan-Cohen, M., 2002, “ Defending Dignity”, in M. Dan-Cohen, Harmful Thoughts:
Essays on Law, Self and Morality, Princeton, NJ: Princeton University Press, 150–171.
 Dempsey, M. M., 2009, Prosecuting Domestic Violence, Oxford: Oxford University
Press.
 Devlin, P., 1965, The Enforcement of Morals, Oxford: Oxford University Press.
 Dimock, S., 1997, “Retributivism and Trust”, Law and Philosophy, 16: 37–62.
 –––, forthcoming, “Contractarian Criminal Law Theory and Mala Prohibita Offenses”,
in R.A. Duff, et al. (eds.), Criminalization: The Aims and Limits of the Criminal Law,
Oxford: Oxford University Press.
 Dolinko, D., 1991, “Some Thoughts about Retributivism”, Ethics, 101: 537–59.
 Dressler, J., 2012, Understanding Criminal Law, 6th edition, New York: LexisNexis
 Dubber, M. D., 2001, “Policing Possession: The War on Crime and the End of Criminal
Law”,Journal of Criminal Law and Criminology, 91: 829–996.
 Dubber, M. D., 2005, The Police Power: Patriarchy and the Foundations of American
Government, New York: Columbia University Press.
 Duff, R. A., 1986, Trials and Punishments, Cambridge: Cambridge University Press.
 –––, 2001, Punishment, Communication, and Community, New York: Oxford
University Press.
 –––, 2005, “Theorizing Criminal Law”, Oxford Journal of Legal Studies, 25: 353–67.

34
 –––, 2007, Answering for Crime: Responsibility and Liability in the Criminal Law,
Oxford: Hart Publishing.
 –––, forthcoming, “Theorizing Criminal Law”, Criminal Law and Philosophy, 8.
 Duff, R. A., with L. Farmer, S. Marshall, and V. Tadros, 2007, The Trial on Trial (3):
Towards a Normative Theory of the Criminal Trial, Oxford: Hart Publishing.
 Duff R. A., and S. P. Green (eds.), 2011a, Philosophical Foundations of Criminal Law,
Oxford: Oxford University Press.
 Duff, R. A., and S. P. Green, 2011b, “ Introduction: Searching for Foundations” , in
Duff and Green (eds.) 2011a, 1–18.
 Dworkin, G. (ed.), 1994, Morality, Harm and the Law, Boulder, Colorado: Westview
Press.
 Dzur, A. W., 2012, Punishment, Participatory Democracy, and the Jury, New York:
Oxford University Press.
 Farmer, L., 1996, Criminal Law, Tradition and Legal Order, Cambridge: Cambridge
University Press.
 –––, 2010, “ Criminal Wrongs in Historical Perspective” , in R.A. Duff, et al.
(eds.), The Boundaries of the Criminal Law, Oxford: Oxford University Press, 214–37.
 –––, forthcoming, “ Criminal Law as an Institution: Rethinking Theoretical Approaches
to Criminalization”, in R.A. Duff, et al. (eds.), Criminalization: The Aims and Limits of
the Criminal Law, Oxford: Oxford University Press.
 Feinberg, J., 1984, Harm to Others, New York: Oxford University Press.
 –––, 1985, Offense to Others, New York: Oxford University Press.
 –––, 1986, Harm to Self, New York: Oxford University Press.
 –––, 1988, Harmless Wrongdoing, New York: Oxford University Press.
 Fletcher, G., 1978, Rethinking Criminal Law, Boston: Little, Brown.
 Gardner, J., 1998, “Crime: In Proportion and in Perspective”, in A. J. Ashworth & M.
Wasik (eds.),Fundamentals of Sentencing Theory, Oxford: Oxford University Press,
31–52.
 –––, 2007, Offences and Defences, Oxford: Oxford University Press.
 Glazebrook, P. R., 1978, “Situational Liability”, in Glazebrook (ed.), Reshaping the
Criminal Law, London: Stevens, 108–19.
 Goldberg, J., and B. Zipursky, 2010, “ Torts as Wrongs”, Texas Law Review, 88: 917–
86.

35
 Green, S. P., 1997, “Why it's a Crime to Tear the Tag off a Mattress: Over-
Criminalization and the Moral Content of Regulatory Offenses”, Emory Law Journal,
46: 1533–1615.
 –––, 2005, Lying, Cheating, and Stealing: A Moral Theory of White-Collar Crime,
Oxford: Oxford University Press.
 –––, 2012, Thirteen Ways to Steal a Bicycle: Theft Law in the Information Age,
Cambridge, Mass.: Harvard University Press.
 Harcourt, B.E., 1999, “The Collapse of the Harm Principle”, Journal of Criminal Law
and Criminology, 90: 109–92.
 Hart, H. L. A., 1968, Punishment and Responsibility, Oxford: Oxford University Press.
 –––, 1994, The Concept of Law, 2nd edition, Oxford: Oxford University Press.
 Hirst, M., 2003, Jurisdiction and the Ambit of the Criminal Law, Oxford: Oxford
University Press.
 Holtug, N., 2002, “The Harm Principle”, Ethical Theory and Moral Practice, 5: 357–
89.
 Horder, J., forthcoming, “Bureaucratic ‘Criminal’ Law: Too Much of a Bad Thing?”, in
R.A. Duff, et al. (eds.), Criminalization: The Aims and Limits of the Criminal Law,
Oxford: Oxford University Press. .
 Hughes, G., 1958, “Criminal Omissions”, Yale Law Journal, 67: 590–637.
 Hulsman, L., 1986, “Critical Criminology and the Concept of Crime”, Contemporary
Crises, 10: 63–80.
 Husak, D., 1987, Philosophy of Criminal Law, Totowa, NJ: Rowman & Littlefield.
 –––, 1995, “The Nature and Justifiability of Nonconsummate Offenses”, Arizona Law
Review, 37: 151–83.
 –––, 1998, “Does Criminal Liability Require an Act?” in R. A. Duff (ed.), Philosophy
and the Criminal Law, Cambridge: Cambridge University Press, 60–100.
 –––, 2004, “The Criminal Law as a Last Resort”, Oxford Journal of Legal Studies, 24:
207–36.
 –––, 2005, “Malum Prohibitum and Retributivism”, in R. A. Duff and S. P. Green
(eds.) Defining Crimes: Essays on the Special Part of the Criminal Law, Oxford:
Oxford University Press, 65–90.
 –––, 2007, Overcriminalization, Oxford: Oxford University Press.
 Jareborg, N., 2005, “Criminalization as Last Resort (Ultima Ratio)”, Ohio State Journal
of Criminal Law, 2: 521–34.
 Kelman, M., 1981, “Interpretive Construction in the Substantive Criminal
Law”, Stanford Law Review, 33: 591–673.

36
 Kleinig, J., 1978, “Crime and the Concept of Harm”, American Philosophical
Quarterly, 27: 32–42.
 Lacey, N., 2000, “‘Philosophical Foundations of the Common Law’: Social not
Metaphysical”, in J. Horder (ed.), Oxford Essays in Jurisprudence, 4th Series, Oxford:
Oxford University Press, 17–39.
 –––, 2001, “ In Search of the Responsible Subject: History, Philosophy and Criminal
Law Theory”,Modern Law Review, 64: 350–71.
 –––, 2009, “ Historicising Criminalisation: Conceptual and Empirical Issues”, Modern
Law Review, 72: 936–60.
 Lamond, G., 2007, “What is a Crime?”, Oxford Journal of Legal Studies, 27: 609–32.
 Lippke, R., 2011, The Ethics of Plea Bargaining, Oxford: Oxford University Press.
 Luban, D., 2004, “A Theory of Crimes Against Humanity”, Yale Journal of
International Law, 29: 85–168.
 –––, 2010, “Fairness to Rightness: Jurisdiction, Legality, and the Legitimacy of
International Criminal Law”, in Besson, S., and Tasioulas, J. (eds), Philosophy of
International Law, Oxford: Oxford University Press, 569–88.
 Marshall, S. E., and R.A. Duff, 1998, “Criminalization and Sharing Wrongs”, Canadian
Journal of Law & Jurisprudence, 11: 7–22.
 Matravers, M., 2011, “ Political Theory and the Criminal Law”, in Duff and Green
2011a, 67–82.
 May, L., 2005, Crimes Against Humanity, New York: Cambridge University Press.
 Mill, J. S., 1859, On Liberty, London: Parker.
 Moore, M. S., 1993, Act and Crime, Oxford: Oxford University Press.
 –––, 1997, Placing Blame: A Theory of Criminal Law, Oxford: Oxford University
Press.
 –––, 2009, “A Tale of Two Theories”, Criminal Justice Ethics, 28: 27–48.
 Murphy, J. G., and J. Coleman, 1990, The Philosophy of Law, 2nd edition, Boulder,
Colorado: Westview Press.
 Norrie, A. W., 2001, Crime, Reason and History, 2nd edition, London: Butterworths.
 Persak, N., 2007, Criminalising Harmful Conduct, Dordrecht: Springer.
 Pettit, P., 1997, “Republican Theory and Criminal Punishment”, Utilitas, 9: 59–79.
 –––, 2002, “Is Criminal Justice Politically Feasible?”, Buffalo Criminal Law Review, 5:
427–50.

37
 –––, forthcoming, “ Criminalization in Republican Theory”, in R.A. Duff, et al.
(eds.),Criminalization: The Aims and Limits of the Criminal Law, Oxford: Oxford
University Press.
 Posner, R. A., 1985, “An Economic Theory of the Criminal Law”, Columbia Law
Review, 85: 1193–1231.
 Ramsay, P., 2012, The Insecurity State: Vulnerable Autonomy and the Right to Security
in the Criminal Law, Oxford: Oxford University Press.
 Raz, J., 1986, The Morality of Freedom, Oxford: Oxford University Press.
 –––, 1987, “Autonomy, Toleration, and the Harm Principle”, in R. Gavison (ed.), Issues
in Contemporary Legal Philosophy, Oxford: Oxford University Press, 313–33.
 Renzo, M., 2013, “Responsibility and Answerability in International Criminal Law”, in
R.A. Duff, et al. (eds.), The Constitution of the Criminal Law, Oxford: Oxford
University Press, 209–36.
 –––, 2012, “Crimes Against Humanity and the Limits of International Criminal
Law”, Law and Philosophy, 31: 443–76.
 Rescher, N., 1972, Welfare: The Social Issue in Philosophical Perspective, Pittsburgh:
University of Pittsburgh Press.
 Ripstein, A., 1999, Equality, Responsibility and the Law, Cambridge: Cambridge
University Press.
 –––, 2006, “Beyond the Harm Principle”, Philosophy and Public Affairs, 34: 215–45.
 Robinson, P. H., 1997, Structure and Function in Criminal Law, Oxford: Oxford
University Press.
 Roxin, C., 2006, Strafrecht Allgemeiner Teil, 4th edition, Volume 1, Munich: CH Beck.
 Schonsheck, J., 1994, On Criminalization: An Essay in the Philosophy of the Criminal
Law, Dordrecht: Kluwer.
 Simester, A. P. (ed.), 2005, Appraising Strict Liability, Oxford, Oxford University
Press.
 Simester, A. P., J.R. Spencer, G.R. Sullivan, and G.J. Virgo, 2010, Simester and
Sullivan's Criminal Law: Theory and Doctrine, 4th edition, Oxford: Hart Publishing.
 Simester, A. P., and A. von Hirsch, 2006, Incivilities: Regulating Offensive Behaviour,
Oxford: Hart Publishing.
 –––, 2011, Crimes, Harms, and Wrongs: On the Principles of Criminalization, Oxford:
Hart Publishing.
 Steiker, C., 1997, “Punishment and Procedure: Punishment Theory and the Criminal-
Civil Procedural Divide”, Georgetown Law Journal, 85: 775–820.

38
 Stephen, J. F., 1873, Liberty, Equality, Fraternity, J. White (ed.), Cambridge:
Cambridge University Press, 1967.
 Stewart, H., 2010, “ The Limits of the Harm Principle”, Criminal Law and Philosophy,
4: 17–35.
 Stuntz, W., 2011, The Collapse of American Criminal Justice, Cambridge, MA:
Harvard University Press.
 Tadros, V., 2005, Criminal Responsibility, Oxford: Oxford University Press.
 –––, 2010, “ Criminalization and Regulation”, in R.A. Duff, et al. (eds.), The
Boundaries of the Criminal Law, Oxford: Oxford University Press, 163–90.
 Thorburn, M., 2011a, “ Criminal Law as Public Law”, in Duff and Green (eds.) 2011a,
21–43.
 –––, 2011b, “ Constitutionalism and the Limits of the Criminal Law”, in R. A. Duff, et
al. (eds.),The Structures of the Criminal Law, Oxford: Oxford University Press, 85–105.
 von Hirsch, A., with D. Garland, and A. Wakefield (eds.), 2000, Ethical and Social
Perspectives on Situational Crime Prevention, Oxford: Hart Publishing.
 Walker, N., 1980, Punishment, Danger and Stigma, Oxford: Blackwell.
 Weigend, T., 1988, “The Legal and Practical Problems Posed by the Difference
between Criminal Law and Administrative Penal Law”, Revue Internationale de Droit
Pénal, 59: 67–86.
 Wells, C., and O. Quick, 2010, Lacey, Wells and Quick, Reconstructing Criminal Law:
Text and Materials, 4th edition, Cambridge: Cambridge University Press.
 Williams, B., 1976, Morality, Cambridge: Cambridge University Press.
 Wohlers, W., with A. von Hirsch, and R. Hefendehl (eds.), 2003, Die
Rechtsgutstheorie, Baden-Baden: Nomos Verlagsgesellschaft.
 Wolfenden, J., 1957, Report of the Committee on Homosexual Offences and
Prostitution (The Wolfenden Report) (Cmnd 247), London: HMSO.
 Yaffe, G., 2010, Attempts in in the Philosophy of Action and the Criminal Law, Oxford:
Oxford University Press.
 Zipursky, B., 1998, “ Rights, Wrongs, and Recourse in the Law of Torts”, Vanderbilt
Law Review, 51: 1–100.

39