NORVIN RICHARDS
CRIMINAL CHILDREN
During a six-month period in 1993, The Tuscaloosa News printed
the following stories from the national wire-services:
“Kids sentenced for trying to hurt teacher”
“Six fourth-graders held on drug charges”
“Boys raped woman, then went to play, police say”
None of these stories headlined the edition in which it appeared;
none even made the front page, for that matter. Evidently, events of
the sort they report are too commonplace for that, even though the
kids who tried to poison their teacher’s iced tea were only in the sixth
grade, the ones arrested for possession and intent to deliver cocaine
were as young as nine years old, and the boys who gang-raped
a woman and then stayed nearby to play basketball ranged from
thirteen to seventeen. It is still distressing to read about children
doing such things, but it is no longer astounding; it is no longer even
particularly difficult to believe.
That is largely because it happens so often. According to the
FBI Crime Index, in 1992 there were 112,409 arrests of persons
under 18 for violent crimes – for murder, forcible rape, robbery,
and aggravated assault, that is.1 More than 300 every day, to put
it differently. Children under fifteen accounted for 34,233 of these:
on an average day, 94 people that young were arrested for violent
crime.2 The number who committed such crimes but eluded arrest
is incalculable, of course.
1 Uniform Crime Reports for the United States, 1992, p. 277. Unfortunately,
this is scarcely a temporary aberration: “Between 1960 and 1981, arrests of juve-
niles for violent crime increased nearly 250%, more than double the figure for
adults during the same period. Although the juvenile court age group constitutes
less than 14% of the population, it accounts for nearly one-fourth of those arrested
for major violent crime such as homicide, rape, robbery, and felonious assault”
(Martin Gardner, ‘Punitive Juvenile Justice: Some Observations on a Recent
Trend’, International Journal of Law and Psychiatry, vol. 10, 1987, p. 140).
2 Uniform Crime Reports, p. 277.
Law and Philosophy 16: 63–89, 1997.
c
1997 Kluwer Academic Publishers. Printed in the Netherlands.
64 NORVIN RICHARDS
Given all this, it is hard to sustain a certain picture of children that
many of us have. According to that picture, a child is an innocent
whose misbehavior, apart from that of the truly exceptional “bad
seed,” is limited to minor mischief and the occasional great blunder.
Children do sometimes do wrong, on this view, but in a very different
way than adults do. The child’s wrongdoing never flows from an evil
heart, the idea is – it is never a deed done coldly and responsibly, but
something that simply happens, a kind of blameless error emerging
from the storm of forces to which the child is subject. The child who
does wrong must be corrected before he grows into a criminal, of
course, but the point is that he isn’t a criminal yet – he is only a child,
and one who needs our help.
This picture seems hopelessly naive and old-fashioned, when
one is confronted with the stories and the statistics. That is signif-
icant, since it is not merely a fond image in the minds of nostalgic
adults, but the conception upon which our juvenile court system was
founded and that still gives that system much of its form. This image
of children is the reason we don’t find adolescents to be criminal
but only delinquent, hold adjudications for them, rather than trials,
expect juvenile judges to function more as social workers than as
dispensers of justice, and so on.3
If the old picture is to be discarded, though, what is the new
one to be? Especially when it comes to children who act in ways
that would be criminal if they were adults: if such children are not
blameless innocents, what are they, exactly? One answer focuses
on adolescents, defined as those between 14 and 18 years old. It
contends that someone of this age who acts in a way that would be
criminal if he were an adult should be considered capable of having
acted culpably, but, if he does turn out to have done so, will have
acted less culpably than an adult who committed the same crime
would have. Adolescents can be criminals, on this view, but only
criminals of a junior kind. The reason for their junior status is not
that crimes are less serious or less harmful when committed by an
3 For a history of the juvenile system, see The Cycle of Juvenile Justice, by
Thomas J. Bernard (Oxford: Oxford University Press, 1992). Serious efforts to
move the system toward a penal model have been underway since at least In re
Gault 387 U.S. 1 (1967), with uneven results.
CRIMINAL CHILDREN 65
adolescent, but that adolescents are less blameworthy for committing
them.4
The lesser-culpability thesis has at least three versions, which I
shall begin by distinguishing. One of those versions is considerably
more plausible than the others, I will argue, but I will also argue
that not even this version withstands examination. As far as I can
see, there is no sound basis for systematically considering culpable
adolescents to be any less culpable than adults who commit the
same crimes. A closing section discusses briefly the implications of
declining to do so.
I
Charles Keene had been married to Billy Thompson’s sister, and
had treated her in ways that made Billy very angry. One night, Billy
and three older friends hunted Charles down. Later, “(t)he evidence
disclosed that the victim had been shot twice, and that his throat,
chest, and abdomen had been cut. He also had multiple bruises and
4
“To be just, punishment must be proportionate to the seriousness of the
offense committed. ‘Seriousness’ is determined by assessing the characteristic
harmfulness of the conduct and the degree of culpability of the offender. When
related to juvenile crime, these principles of justice require that juvenile offenders
be punished less severely than their adult counterparts : : : considerations of
: : : characteristic harmfulness seemingly do not themselves demand deviation
: : : Assessments of juvenile culpability do, however, strongly support a system of
scaled-down punishments for offenders dealt with through the juvenile system.
While the punitive model requires holding juveniles accountable for their offense,
the extent of their accountability should not be synonymous with that of a similarly
situated adults” (Gardner, p. 142).
Gardner’s version takes only those who would not be transferred for trial as
adults to be less culpable. This is a difficult class to specify, since a juvenile cannot
be transferred unless the District Attorney seeks to have this done – a decision
with political implications – and is then decided by reference to several variables.
(In Alabama, for example, the judge is required to take six factors into account.)
Versions of this position without this modification are put forward by Justice
Powell, writing for the majority in Eddings v. Oklahoma, 102 S. Ct. 869 (1982);
by Justice Stevens, writing for the plurality in Thompson v. Oklahoma, 108 S. Ct.
2687 (1988); by Andrew Walkover in ‘The Infancy Defense in the New Juvenile
Court’, UCLA Law Review 31 (1984); and by Barry C. Feld in both ‘The Decision
to Seek Criminal Charges: Just Deserts and the Waiver Decision’, Criminal Justice
Ethics, Summer/Fall 1984 and ‘The Juvenile Court Meets the Principle of Offense:
Punishment, Treatment, And The Difference It Makes’, Boston University Law
Review 68 (1988).
66 NORVIN RICHARDS
a broken leg. His body had been chained to a concrete block and
thrown into a river where it remained for almost four weeks.”5
Each of the four participants was tried separately, and each was
sentenced to death. Billy Thompson appealed his sentence, on the
ground that he was only fifteen years old when the murder was
committed. The Supreme Court ruled in Billy’s favor, finding that
his youth made executing him unconstitutional – that to do so
would violate the Eighth Amendment’s prohibition against cruel
and unusual punishment.
In part, the Court’s reasoning was that the death penalty would
necessarily exceed the deserts of anyone so young. Unlike his older
companions, a fifteen-year-old could not deserve so harsh a penalty;
as a fifteen-year-old, he must have acted less culpably. Here is Justice
Stevens to that effect, for the plurality:
Thus, the Court has already endorsed the proposition that less culpability should
attach to a crime committed by a juvenile than to a comparable crime committed
by an adult. The basis for this conclusion is too obvious to require extended
explanation. Inexperience, less education, and less intelligence make the teenager
less able to evaluate the consequences of his or her conduct while at the same
time he or she is much more apt to be motivated by mere emotion or peer pressure
than is an adult. The reasons why juveniles are not trusted with the privileges and
responsibilities of an adult also explain why their irresponsible conduct is not as
morally reprehensible as that of an adult.6
On one plausible reading, this sets forth a categorial assertion
about adolescents.7 Specifically, it asserts that no matter who the
adolescent happens to be, he or she must suffer from defects that
diminish responsibility: he or she must be inexperienced, unedu-
cated, less intelligent than an adult, and more vulnerable to surges
of emotion and to peer pressure. That is just the way we all are
when we are that age, and that is why no adolescent is as culpable
for “irresponsible behavior” as an adult would be who acted in the
same way.8 Although Justice Stevens believes that “(t)he basis for
5 Thompson v. Oklahoma, 108 S.Ct. at 2690.
6 Id. at 2698–2699.
7 Justice Scalia so reads it in his dissent. See especially 2711–2719.
8 Thompson v. Oklahoma, 108 S.Ct. at 2687, dissenting opinion. As Justice
Scalia reads them, the plurality has found a societal consensus that “no one so
much as a day under 16 can ever be mature and morally responsible enough
to deserve” the death penalty (2718, his emphasis). Justice Scalia contends the
consensus is only that such miscreants are rare.
CRIMINAL CHILDREN 67
this conclusion is too obvious to require extended explanation,” let
us consider how such an explanation would run.
Notice first that we could not establish the conclusion that adoles-
cents are always less culpable than adults by showing that every
adolescent act is performed under some mitigating circumstance or
other. That would establish only that adolescents are never fully
culpable, and never deserve as harsh a penalty as a perfect villain
would. This differs from their always deserving less than adults
who commit the same crime, since clearly not every adult felon is a
perfect villain either. It would remain possible for some adolescents
to be more culpable than some adults, even if we could show they
were always less than fully culpable.
The premise required instead to establish Justice Stevens’s
conclusion is that there is always greater mitigation for adolescent
misbehavior than there is for the corresponding adult misbehavior,
regardless who the two individuals are and what crime they commit.
We would need to show that adolescents are always more ignorant
of important matters, always under greater pressure, and so on. The
trouble is that this is far from obvious. It seems especially implau-
sible when the adolescent is about to graduate to adulthood and the
adult has only recently done so. The difference in age might be a
matter of days: why must the younger felon have acted with greater
mitigation than the older? We could just insist that his adolescence
guarantees this, but that would hardly be persuasive.
The more plausible picture appears to be the one drawn by Justice
Lewis Powell in an earlier case: “Minors who become embroiled
with the law range from the very young up to those on the brink
of majority. Some of the older minors become fully ‘street-wise,’
hardened criminals, deserving no greater consideration than that
properly accorded all persons suspected of crime.”9
In short, this first version of the lesser-culpability thesis is not very
convincing. It asserts that every adolescent must be less responsible
for his or her misdeeds than any adult. It is difficult to see why that
would be true, and easy to imagine apparent counterexamples.
Consider, then, a second version, derived from Justice O’Con-
nor’s reading of the Thompson opinion. She concurs with the plural-
9 Fare v. Michael C., 442 U.S. at 734, n. 4; 99 S.Ct. at 2576, n. 4; 61 L.Ed. 2d
197 (1979) (dissenting opinion).
68 NORVIN RICHARDS
ity, but she takes its position differently. As she reads him, Justice
Stevens was not asserting that every adolescent is necessarily less
culpable than any adult, but only “that adolescents are generally less
blameworthy than adults who commit the same crime.”10 Because
that is the way things generally are, her reasoning goes, it is the
best rule for us to follow. We should treat every adolescent as less
blameworthy not because he or she must be, but because the typical
adolescent is. Admittedly, this will give the exceptional adolescent
a break he or she does not deserve, but that is just an unavoidable
cost of employing the rule that fits most cases.
The natural question to ask in response is why we should employ
a rule at all, as opposed to determining each individual’s culpability
separately. There are many contexts in which we do employ rules,
of course, and in some of those we are concerned to give individuals
what they deserve. (Think, for example, of choosing which students
to admit to law school by reference to test scores and grades: those
are rules of thumb we use even though, of course, we want to admit
the most deserving applicants.)
In general, a defense of using rules rather than something more
individualized would rest on the following points.
(i) Employing rules could be the only practical method, if there
were a great many cases to be handled and it were necessary to
reach a decision promptly.
(ii) It could be that the generalizations on which the rules rested had
very few exceptions, so that we would make very few mistakes
if we employed them.
(iii) It could be that the mistakes we would make were unimportant.
(iv) It could be doubtful that anything more individualized would be
an improvement. (For example, the generalizations could be the
best basis we could have for predicting who would do well in
college or in law school, with individualized procedures simply
permitting hunches and biases to enter the process.)
Assuming that this is the general way in which to defend using
a rule to make decisions, how well could we defend using the rule
that every adolescent is to be punished less severely than his adult
counterpart? Certainly we could start by observing that there are a
10 Thompson v. Oklahoma, 108 S.Ct. at 2708 (concurring opinion), my italics.
CRIMINAL CHILDREN 69
great many cases to be dealt with, and that promptness is important.
After that, however, the defense would begin to break down.
First, although there are many cases to consider, it isn’t even
clear that using the rule would mean working through those cases a
great deal more quickly than considering desert individually, nor is
it clear that this is the only practical method by which to proceed. If
we were to employ the rule that adolescents always deserve less than
adults, we would still have to conduct a trial for each adolescent in
order to determine his or her guilt or innocence and, if the verdict
was “guilty,” decide which of the lesser adolescent penalties the
defendant deserved. The rule would save us only from considering
penalties so harsh that only an adult could deserve them, if there are
any – execution, perhaps. That would save time, but only in those
cases in which we might execute an adult.
In short, this is not a context in which we can defend using our rule
by saying we must do so in order to function. Nor is it easy to dismiss
the errors we would make in employing our rule as unimportant
slips that are not worth a great deal of trouble to avoid: they are
miscarriages of justice, after all. On the one hand, they are errors in
which we fail to give the wrongdoer what she deserved because we
underestimate how badly she acted. On the other, they are errors in
which we fail to treat equally culpable miscreants equally, since the
adolescent gets off more lightly than the adult who was (in fact) no
more culpable than she, and possibly even less so.
The other elements of the potential defense are at best claims
in need of support. It hasn’t been shown that there are only a few
exceptions to the generalization that adolescents act less culpably
than adults. Nor is it clear that considering adolescent culpability
case-by-case would be no more accurate than employing the rule:
certainly we think individualized consideration is more accurate
when the defendant is an adult, and there is no reason to think that
adolescence makes an individual’s culpability especially opaque to
us.
In sum, there are good reasons to doubt that we ought to employ
the rule guaranteeing every adolescent a milder punishment than his
adult counterpart. Essentially, it isn’t clear that this rule would be
especially useful or desirable. Admittedly, my argument to this con-
clusion has assumed that the burden is on those who want to abandon
70 NORVIN RICHARDS
an individualized approach to culpability: I have assumed that we
should consider cases individually unless there is good reason not
to do so. This might be resisted on the ground that to individualize
is to grant discretion, with discretion taken to be objectionable for a
variety of reasons.11 The individualized approach I am recommend-
ing for juveniles, however, is hardly the sort in which all decisions
are ad hoc and the judge is expected to deploy resources of personal
wisdom in the manner of King Solomon.12 I am recommending only
that we treat adolescents in the same way as we do adults, individ-
ualizing our consideration of them to the same extent we do that of
adults we try for the same offenses. The discretion afforded by doing
so has a variety of limits.13
More generally, I agree with Carl Schneider’s thesis that the
proper question is not whether to allow discretion or insist on rules,
but which mix of rules and discretion is best for the particular task
at hand. As he argues, “The less risk of bias, the greater the need to
ask whether substituting rules for bias, the greater the need to ask
whether substituting rules for discretion would be more costly than
11 For extended discussions of discretion, see Discretion and Welfare, edited by
Michael Adler and Stewart Asquith (London: Heinemann, 1981), Discretionary
Powers: A legal Study of Official Discretion, by Dennis Galligan (Oxford: The
Clarendon Press, 1986), and The Uses of Discretion, edited by Keith Hawkins
(Oxford: The Clarendon Press, 1992).
12 Carl Schneider dubs this “khadi-discretion”, after Max Weber’s concept of
khadi-justice. He borrows the following characterization from A. T. Kronman’s
Max Weber (Stanford: Stanford University Press, 1983): : : : cases are decided on
an individual basis and in accordance with an indiscriminate mixture of legal,
ethical, emotional and political considerations. Khadi-justice is irrational in the
sense that it is peculiarly ruleless: it makes no effort to base decisions on general
principles, but seeks, instead, to decide each case on its own merits and in light
of the unique considerations that distinguish it from every other case : : : The
characterization of khadi-justice as a substantive form of law-making highlights
another of its qualities, namely, its failure to distinguish in a principled fashion
between legal and extra-legal (ethical or political) grounds for decision. It is the
expansiveness of this form of adjudication-its willingness to take