DUFFMACRO 5/5/2003 4:46 PM
Virtue, Vice, and Criminal Liability:
Do We Want an Aristotelian Criminal Law?
R.A. Duff†
1. INTRODUCTION: THE YEARNING FOR GRAND THEORY
In criminal law theory, as in other kinds of theorizing,
there is a powerful temptation (or, less question-beggingly, a
powerful disposition) to search for a grand, unitary theory of
criminal liability: some unitary account of what we are or
should be liable for, of the conditions under which we are or
should be held liable, of the basic structure and elements of
liability, which applies across the whole range of criminal
offenses. Such theorizing is typically an exercise in censorial
rather than in purely analytical jurisprudence: although
theorists of a Herculean disposition might claim that their
account makes best sense of the values and aspirations that
can be discerned within our existing legal systems and
structures, their claim has in the end to be that this is how
the law ought to be, not that this is how the law actually is
in its contingent historical complexity. Such theories thus
offer us, as any good normative theory should offer us, a
standard against which we can assess and criticize existing
legal doctrine and practice.
The most ambitious theory would specify necessary
and sufficient conditions of criminal liability: for instance
that we are liable for and only for the choice to do or to risk
doing what the law defines as the actus reus of a crime.
Others, more modestly, claim only to specify a necessary
condition of liability: for instance that criminal liability
always requires, or must always be for, an “act” or a
† Professor, Department of Philosophy, University of Stirling. This paper
was completed during my tenure of a Leverhulme Research Fellowship; I am very
grateful to the Leverhulme Trust for this support, and to participants in the 2002
Analytic Legal Philosophy Conference for helpful comments and criticism.
DUFFMACRO 5/5/2003 4:46 PM
148
BUFFALO CRIMINAL LAW REVIEW [Vol. 6:147
“voluntary act.” (Slightly less ambitious theorists of these
kinds also allow that there could be justified exceptions to
such requirements.) Others eschew any ambition to specify
even normally sufficient and/or necessary conditions of
liability in favor of a paradigm-based account: they aim to
explicate the paradigm of liability (for instance, the
intentional commission of what the law defines as an actus
reus), and to explain any justified modes of liability that do
not fit that paradigm as being parasitic on or derivative
from it.
Now I do not wish to denigrate the aspirations to, and
of, such grand theorizing—indeed, in some moods and on
some days I share such aspirations. Even if such
theorizing is doomed to ultimate frustration, and even
though it can lead to dangerously procrustean attempts to
fit every aspect of criminal liability to one model, its
pursuit can reveal structures which, whilst not as unitary
or universal as the grand theorizer might want, are
nonetheless significant and interesting. This paper,
however, is rather motivated by a respect for the more
particularist, deflationary critiques of grand theory: for
arguments to the effect, not just that our existing systems
of criminal law do not match any such theory (to which the
answer is that the theory is normative rather than merely
descriptive), but that we should not expect any viable
human system of criminal law to match any such grand
theory—that we should look not for a single model of
criminal liability, but for a number of different models,
patterns and structures that interweave (and may conflict)
in various and complex ways.
This paper is particularly concerned with the recent
development of Aristotelian accounts of criminal liability,
some of which do seem to aspire to the status of grand,
unitary theory. I will argue that whilst Aristotelian virtue
theory might play a central role in our understanding of
moral life and moral goodness, it should play no such
central role in our understanding of, or in the law’s
determinations of, criminal liability.
DUFFMACRO 5/5/2003 4:46 PM
2002] VIRTUE, VICE AND CRIMINAL LIABILITY
149
Section 2 sketches the recent emergence of an
Aristotelian virtue-based approach, and its claim to deal
with the problems faced by Kantian “choice” theories and
Humean “character” theories. Section 3 highlights some
issues that any general theory faces, about the relationship
between the favored basis of liability (be it choice,
character, or vice) and the conduct which seems to provide
the focus both of the law’s definitions of offenses and of the
criminal trial. Section 4 brings out some central features of
Aristotle’s account of virtue and vice, to show why the
criminal law should not be concerned with virtue and vice
in their fully Aristotelian senses. Section 5 discusses the
main ways in which theorists have thought that some kind
of (roughly) Aristotelian conception of virtue or vice should
nonetheless figure in an account of criminal liability and its
conditions: the conclusion, in Section 6, is that whilst we
can find a useful role for some etiolated ideas of virtue and
vice in explaining some aspects or types of criminal
liability, the grander ambitions of Aristotelian theorists are
misguided: criminal liability is not generally grounded in
the manifestation of anything like Aristotelian vice.
2. CHOICE, CHARACTER AND VICE
Perhaps the most familiar kind of grand theory in
recent years has been the “choice” theory in its various
forms. Criminal liability, insist choice theorists, must be
determined by what we choose to do: by our choice to do or
to risk doing something that the law defines as a crime.1
Such a conception of criminal liability sits happily with
central aspects of a familiar kind of liberal individualism
that often finds its inspiration in Kant. It can claim to take
individual freedom seriously: by making our fate at the
1. See, e.g., H.L.A. Hart, Punishment and Responsibility: Essays in the
Philosophy of Law (1968); Andrew Ashworth, Belief, Intent and Criminal
Liability, in Oxford Essays in Jurisprudence (John Eekelaar & John Bell eds., 3d
ser. 1987); Joshua Dressler, Reflections on Excusing Wrongdoers: Moral Theory,
New Excuses, and the Model Penal Code, 19 Rutgers L. J. 671 (1988); Michael S.
Moore, Choice, Character, and Excuse, 7 Soc. Phil. & Pol’y 29 (1990).
DUFFMACRO 5/5/2003 4:46 PM
150
BUFFALO CRIMINAL LAW REVIEW [Vol. 6:147
law’s hands depend on our choices, it increases the extent
to which we can predict and so control that aspect of our
lives. It claims to respect our autonomy, by making the law
respond to us on the basis of our choices: for it is in our
exercise of the capacity for choice that we most clearly
display our rational agency. It can claim to treat us justly,
in holding us liable only for what we choose: for we can be
justly held responsible only for what lies under our rational
control, and it is in choosing that we exercise such rational
control as we have. It can claim to respect individual
privacy, and the neutrality that a liberal state should
supposedly exercise: for whilst on this view the criminal
law must of course look beyond overt conduct to the choices
by which that conduct is informed, it does not inquire (or
intrude) into the deeper aspects of our character or into our
conceptions of the good; it is interested only in the choices
by which we ourselves voluntarily engage as agents with
the public world.
Choice theories face some familiar criticisms, whose
general form is that theories which focus solely on “choice,”
however that notion is understood, omit too much that is
relevant to criminal liability. The broadest criticism of this
kind is that by “abstracting” individual agents from the
social contexts which help to determine their actions, such
theories distract us from a range of social factors that
should bear crucially on criminal liability—that the fiction
of the rational choosing individual which such theories
create is one that undermines the legitimacy of our
ascriptions of criminal liability to those socially situated
agents who actually appear in our courts.2 That kind of
criticism, however, will apply with as much (or as little)
force to the “character” and “vice” theories that I consider
later, since they too focus on individual agents who can,
they presume, be sufficiently abstracted from their social
contexts to be bearers of criminal responsibility; I will
2. See, e.g., Alan W. Norrie, Crime, Reason, and History: A Critical
Introduction to Criminal Law (1993); for one response, see R.A. Duff, Principle
and Contradiction in the Criminal Law, in Philosophy and the Criminal Law:
Principle and Critique (R.A. Duff ed., 1998).
DUFFMACRO 5/5/2003 4:46 PM
2002] VIRTUE, VICE AND CRIMINAL LIABILITY
151
therefore not attend to it further here. Two other
criticisms are more relevant to present purposes, since they
might seem to favor a character-based over a choice-based
theory. They are that choice is neither a necessary nor a
sufficient condition of criminal liability.
The objection that choice is not a necessary condition
of criminal liability does not rest simply on the claim that
criminally culpable action need not be preceded by any
process of deliberation: any plausible choice theory will
allow that criminal actions can be spontaneous, and that
the choice to do Φ need not be separable from the doing of
it. Nor does it rest simply on the claim that we can be
criminally liable for negligent conduct, which precisely does
not involve choosing to cause or to risk causing criminal
harm: for choice theorists who want to allow for such
liability can both legitimate it and show why negligence is
a lesser species of fault by talking about the choices (to
attend, to take care or precautions) that the agent should
and could have made but did not make.3 It rests, rather, on
the claim that there are “crimes of indifference” which are
more serious than crimes involving merely negligence, but
in which the agent’s reckless indifference is displayed
precisely in his failure to notice or advert to the harm that
he is likely or certain to do.4 For a choice theorist, the
agent’s unawareness of the relevant features of his action
(for instance, that the attack in which he is engaged might
be fatal, or that this woman might not be consenting to
sexual intercourse with him) precludes liability for
anything worse than a crime of negligence; to understand
the seriousness of his wrongdoing we must attend not just
to the choices he makes, but to the attitudes or concerns (or
3. See H.L.A. Hart, Negligence, Mens Rea, and Criminal Responsibility, in
Hart, supra note 1.
4. See, e.g., Samuel H. Pillsbury, Crimes of Indifference, 49 Rutgers L. Rev.
105 (1996) (though he wants to build choice—choices about what to attend to or to
care about, for instance—into the picture in a way that seems to me neither
plausible nor necessary); R.A. Duff, Intention, Agency, and Criminal Liability:
Philosophy of Action and the Criminal Law ch. 7 (1990).
DUFFMACRO 5/5/2003 4:46 PM
152
BUFFALO CRIMINAL LAW REVIEW [Vol. 6:147
lack of concern) that are manifest in his actions and in his
very failure to notice or attend to those relevant aspects.
The objection that choice is not a sufficient condition of
liability rests on the claim that, in order to explain how
someone who chooses (in any ordinary sense of the word) to
commit what the law defines as a crime can escape
conviction by offering some legally recognized excuse, we
cannot just appeal to the idea of a capacity, or of a fair
opportunity, to choose to refrain from the crime. To explain
why duress, for instance, should be an excuse (and when it
should not excuse), we must look not just at the agent’s
choice to give in to the threat, but at what his giving in
showed about his courage or cowardice, and about his
commitment or his lack of commitment to the interests or
values that he injured—at, in other words, features of his
character that lie behind the choices he made.5
These two (alleged) defects in choice theories help
explain the attractions of the Humean character-based
theories of liability that were typically offered as
alternatives.6 What matters, on a Humean theory, is not
(just) the agent’s overt conduct, nor (just) her choice to
engage in such conduct (even if she did so choose), but the
character traits that can be inferred from her conduct: she
should be criminally liable if and only if we can infer some
relevantly defective or undesirable character trait from her
conduct; excuses exculpate those who commit what is
defined as a criminal act if and because they cite some
factor that blocks the usual inference from criminal
conduct to defective character trait.
One general worry about character-based theories, for
traditional liberals, is whether we should allow the
criminal law to take such an interest in our characters;
5. See, e.g., Claire O. Finkelstein, Duress: A Philosophical Account of the
Defense in Law, 37 Ariz. L. Rev. 251 (1995); see also R.A. Duff, Choice, Character,
and Criminal Liability, 12 Law & Phil. 345, 350-61 (1993).
6. See, e.g., Michael D. Bayles, Character, Purpose, and Criminal
Responsibility, 1 Law & Phil. 5 (1982); George Vuoso, Note, Background,
Responsibility, and Excuse, 96 Yale L. J. 1661 (1987); Nicola Lacey, State
Punishment: Political Principles and Community Values ch. 3 (1988).
DUFFMACRO 5/5/2003 4:46 PM
2002] VIRTUE, VICE AND CRIMINAL LIABILITY
153
another is whether such character-based liability must
presuppose, if it is to be just, that we are responsible for
our characters—and if so, whether we are thus responsible.
I comment on these two worries later. There are, however,
two more particular problems with the Humean character-
based conception of liability, which point the way towards
the new, Aristotelian accounts.
First, it is not clear which kinds of character trait
should ground liability. Hume himself counted as virtues
those character traits which were either useful or
agreeable, to the person who possessed them or to others;7
which implies a broad class of vices, as character traits that
are either harmful or disagreeable, to their possessor or to
others. Modern Humean character theorists also
sometimes cast the net of liability rather widely: so Bayles
would hold mentally disordered offenders criminally liable
(although their disposal need not be punitive).8 Second, on
the Humean picture the relationship between character
and action is contingent and causal: traits are underlying
psychological conditions which cause actions; given
adequate empirical expertise, we can infer traits from
actions in the way that we generally infer causes from their
effects. But this makes the criminal law’s apparent focus
on actions very puzzling: it is now unclear either why
criminal action should be necessary for liability at all—
since we could surely have other persuasive evidence of
undesirable character traits than that; or why just one
criminal action should suffice for liability—since one would
usually expect more than one such piece of evidence as
grounds for an adequately confident inference to its cause.9
Humean theorists are not, of course, simply silenced
by such problems: but theorists who look to “character”
rather than “choice” as the basis of criminal liability have
recently turned from Hume towards Aristotle for their
7. David Hume, A Treatise of Human Nature III.iii.1, at 591 (L.A. Selby-
Bigge & P.H. Nidditch eds., 2d ed. 1978) (1739).
8. Bayles, supra note 6, at 17-18.
9. See Bayles, supra note 6, at 10; Lacey, supra note 6, at 75; George P.
Fletcher, Rethinking Criminal Law 800-01 (1978).
DUFFMACRO 5/5/2003 4:46 PM
154
BUFFALO CRIMINAL LAW REVIEW [Vol. 6:147
inspiration; and an Aristotelian account can claim to avoid
these two problems. For we can say, first, that the
character traits which should ground criminal liability are
those that constitute vices—i.e., in Aristotelian terms,
those that embody a distorted or ethically mistaken
conception of the good; and, second, that while there is
more to Aristotelian vice than wrong action, such action is
partly constitutive of, not just empirical evidence for, the
relevant character trait.10
This Aristotelian turn in criminal law theorizing is
part of a wider movement into “virtue jurisprudence,”11 and
reflects the revival of interest in Aristotelianism in
philosophical ethics. It also reflects reactions against the
neutralist liberal individualism that the traditional choice
theories expressed—reactions both by communitarians who
eschew liberal individualism and by perfectionist liberals
who eschew neutralism, all of whom seem prepared to
allow the state to take a closer interest in its citizens’ moral
characters than traditional liberals would allow.
Before I turn to the question of whether or how far
Aristotelian conceptions of virtue and vice can illuminate
the grounds of criminal liability, a more general question
must be tackled, about the role that the favored basis of
criminal liability—choice, or character, or vice—is to play
in grounding liability.
10. For examples of Aristotelian accounts of the grounds of criminal liability
(by no means all of which aspire to provide grand, unitary theories in such terms),
see Finkelstein, supra note 5; Kyron Huigens, Virtue and Inculpation, 108 Harv.
L. Rev. 1423 (1995), Virtue and Criminal Negligence, 1 Buff. Crim. L. Rev. 431
(1998); Dan M. Kahan & Martha C. Nussbaum, Two Conceptions of Emotion in
Criminal Law, 96 Colum. L. Rev. 269 (1996); John Gardner, The Gist of Excuses,
1 Buff. Crim. L. Rev. 575 (1998), On the General Part of the Criminal Law, in
Philosophy and the Criminal Law: Principle and Critique, supra note 2; Victor
Tadros, The Characters of Excuse, 21 Oxford J. of Legal Stud. 495 (2001).
11. See, e.g., Lawrence B. Solum, The Virtues and Vices of a Judge: An
Aristotelian Guide to Judicial Selection, 61 S. Cal. L. Rev. 1735 (1988).
DUFFMACRO 5/5/2003 4:46 PM
2002] VIRTUE, VICE AND CRIMINAL LIABILITY
155
3. INTENTIONAL OBJECTS OF LIABILITY, OR CONDITIONS OF
LIABILITY?
When I am blamed, or condemned, or held liable, there
is something for which I am blamed, condemned, or held
liable: which is to say that blame, condemnation and
holdings of liability require an intentional object on which
they are focused and towards which they are directed.
Within any practice of blame, condemnation, or holding
liable, there will also be conditions of liability—conditions
which must be satisfied if the blame, condemnation, or
holding liable is to be justified (or perhaps even intelligible)
but which are not themselves part of the object of the
blame, condemnation, or liability. It is, for instance, a
condition of being justly blamed, or condemned, or held
liable for a wrong I have committed that I am not mentally
disordered in a way that undermines my responsibility for
that action: but I am not blamed, or condemned, or held
liable for not being mentally disordered.12
When it is said that criminal liability is grounded in
choice, or character, or vice we must therefore ask whether
that “ground” is supposed to constitute the intentional
object of liability, or a condition of liability. Are we to be
held liable for our choices (to do or to risk doing what the
law defines as criminal), or defective character traits, or
vices? Or are we to be held liable for something else—for,
as the most obvious possibility, a criminal action—on
condition that we chose to do or to risk doing that action; or
on condition that that action flowed from, and thus justified
an inference to, a relevant undesirable character trait; or
on condition that that action flowed from a relevant vice?
12. This distinction is too often ignored by theorists: but see Jeremy Horder,
Criminal Culpability: The Possibility of a General Theory, 12 Law & Phil. 193,
204-06 (1993); Douglas Husak, Does Criminal Liability Require an Act?, in
Philosophy and the Criminal Law: Principle and Critique, supra note 2, at 67-73;
Tadros, supra note 10, at 501. The distinction is clearly a close relative of that
between “wrongdoing” and “attribution” (see Fletcher, supra note 9, at ch. 6.6-8),
but I cannot discuss their relationship here.
DUFFMACRO 5/5/2003 4:46 PM
156
BUFFALO CRIMINAL LAW REVIEW [Vol. 6:147
The possible views about how choice, character, or vice
can ground liability are actually more symmetrical, and
more complex, than the previous paragraph suggests. They
are more symmetrical because even theorists who focus on
choice, character, or vice as the intentional object of
liability typically accept the necessity for some kind of “act
requirement,” if only to provide suitably reliable, non-
intrusive proof of guilt: their view then is that we are liable
for our choices, character traits, or vices on condition that
they lead to or are revealed in criminal action—something
that constitutes the actus reus of an offense. They are also
more complex, because different conceptions of the
relationship between choice, character, or vice and action
make room for different accounts of the intentional object of
liability.
If we see that relationship as purely contingent and
causal, we will see a sharp distinction between the
intentional objects and the conditions of liability. Action
would then be a distinct condition of liability for criminal
choices, character traits, or vices: a condition most
plausibly justified by the need to base convictions on
reliable evidence that can be established without unduly
oppressive or intrusive investigative methods. Or choice,
character trait, or vice would be a distinct condition of
liability for criminal actions: a condition most plausibly
justified in terms of culpability or responsibility.
If we instead see the relationship as logical or
constitutive, matters are more complicated. The strongest
claim of logical connection would be that—either in virtue
of the very concept of choice, character, or vice, or as a
feature of the conception of choice, character, or vice that is
appropriate to the criminal law—the relevant choice,
character trait, or vice is actualized in and only in action of
the relevant kind. A person might intend a criminal action
without as yet having taken even a first step towards
carrying that intention out, but counts as “choosing” to
commit that crime only at the point when she begins to
translate intention into action. Or she might harbor
dishonest thoughts, inclinations, or intentions, but counts
Add New Comment