INTRODUCTION

The legal response to juvenile crime is undergoing revolutionary change, and its ultimate shape is uncertain. The traditional juvenile court, grounded in optimism about the potential for rehabilitation of young offenders, has long been the target of criticism, and even its defenders have been forced to acknowledge that it has failed to meet its objectives.1 Beginning in the late 1960s, when the Supreme Court introduced procedural regularity to delinquency proceedings in In re Gault,2 courts and legislatures began to slowly chip away at the foundations of the juvenile justice system.3 Recent developments have accelerated and intensified that process, as policy-makers at both the state and federal level respond to public fear and anger at what is perceived to be an epidemic of youth violence, including an alarming increase in juvenile homicide.4 Increasingly, critics of the traditional juvenile justice system argue that young offenders should be subject to the same punishment as adults for the harms they cause.5 The next step, which many are quite ready to take, is to abolish the separate juvenile justice system.

One way to think about the evolution of legal policies responding to youthful crime is in terms of the empirical account of adolescence that is expressed through these policies. From this perspective, the traditional (pre-Gault) juvenile court was shaped in important ways by a conception of errant youth as childlike, psychologically troubled, and malleable.6 On this view, the job of the court was not to punish, but to rehabilitate and protect its charges. With the reform movement of the 1970s and 1980s, a less idealized view of adolescence emerged, together with a growing skepticism about the potential for rehabilitation.7 Immature youth were seen as less culpable than adults, but not as blameless children. Lacking experience and judgment, young offenders needed lessons in accountability. A perusal of the current landscape of juvenile justice reform suggests a view of delinquent youth as appropriately subject to adult punishment and procedures and thus as indistinguishable in any important way from their adult counterparts.

Our aim in this essay is to examine these changing accounts through a developmental lens, with a purpose of bringing into the policy debate on juvenile justice reform the insights of development psychology. This perspective is useful in providing a scientific measure of the empirical assumptions, intuitions, and predictions about adolescence that have always played a large role in policy formation in this context. The framework challenges two assumptions underlying the contemporary punitivist reforms. The first is what might be called the "competence assumption:" that no important differences distinguish adolescents and adults charged with crimes. Modern developmental psychology provides substantial, if indirect, evidence that adolescent choices about involvement in crime and their decisions as defendants in the legal process reflect cognitive and psychosocial immaturity. This evidence challenges contemporary juvenile justice policies which discount the importance of conventional notions of criminal responsibility and constitutional requirements for fair proceedings (neither of which the critics challenge directly).

The second premise implicit in the recent reforms might be labeled the "utilitarian assumption." It qualifies the competence assumption in holding that, even if developmental differences do distinguish adolescent and adult offenders, the enormous social cost inflicted by young criminals requires that those differences be ignored in formulating a legal response to youth crime. This assumption is challenged by evidence derived from taxonomy of adolescent delinquent behavior,9 which explains much about the nature of adolescent criminal conduct and clarifies the importance of differentiating among young offenders. The modern punitivist reforms tend to treat adolescent offenders as though most are young career criminals-a premise that is true of only a small group of offenders whose delinquency in adolescence is part of a persistent pattern of antisocial behavior, often beginning in early childhood. The criminal activity of most adolescents, in contrast, reflects a relatively typical inclination to engage in antisocial behavior during this developmental stage-a tendency that desists with maturity.10 Thus, policies that focus solely on the harm caused by youthful offenders may not be the optimal means to achieve the instrumentalist goals of their proponents. These policies fail to calculate the long-term social costs of categorical punishment, particularly the costs incurred by diminishing the prospects for productive adulthood of those offenders whose delinquent behavior reflects transient developmental influences.

In combination, the developmental lessons produce a counterintuitive insight about the relationship among immaturity, culpability and criminal persistence, that has not yet been recognized in the formulation of juvenile justice policy. Youths who offend at a younger age (and who are thus less mature and less culpable) may be more likely to become adult career criminals than teens who first initiate even serious antisocial behavior in mid-adolescence or later. This observation poses a formidable challenge to the development of fair and effective policies responding to youth crime. It suggests that juvenile justice policy must attend to patterns of delinquent behavior and not simply to the severity of the offense, and it contributes to our conclusion that a developmental model of juvenile justice should incorporate both rehabilitative and retributive dimensions.

The following roadmap may be helpful. In Part I, we will briefly sketch the changing conceptions of adolescence that have been reflected in the evolution of juvenile justice policy over the past century. In Part II, we present a developmental framework. First, we describe the role of antisocial conduct in adolescent development, and sketch the taxonomy, which includes two rough categories of youthful criminal behavior: "adolescent-limited" and "life-course-persistent."11 We then offer a positive account of developmental factors that may influence decision-making in ways that distinguish adolescents from adults.12 Finally, we apply this framework, examining the impact of developmental factors on decisions to engage in criminal conduct and on decisions in the criminal process. In Part III, we explore the possible implications of developmental knowledge for criminal blameworthiness, and conclude that developmental psychology evidence supports a presumption of youthful diminished responsibility for younger and midadolescents. We then examine through a developmental lens the utilitarian argument that societal protection necessitates severe penalties for youthful offenders. This perspective clarifies that punitive policies may poorly serve the efficiency goals of their supporters. In Part IV, we examine the lessons of the developmental perspective for juvenile justice policy and suggest some directions for policy that is formulated in a developmental framework.

I. CHANGING PERSPECTIVES ON ADOLESCENCE IN JUVENILE JUSTICE REFORM

In this Part we describe accounts of adolescence as reflected in juvenile justice policy during three periods in the last century. Each period, in its way, was one of reform-a time in which existing legal policy toward youth crime was challenged by a new conception of adolescence and adolescent crime. The first was the period of Progressive reform early in the century; the second was the post-Gault period of the 1970s and 1980s; and the third is the contemporary period.

A. THE TRADITIONAL COURT

The creation at the turn of the century of a separate system of juvenile justice, committed to rehabilitation of young offenders, was a product of the social reform movement of that period.iS It also reflected the late l9th century understanding of the nature of crime and a new recognition of psychological differences between youths and adults, which was emerging from the "new" science of psychology.14

The focus on rehabilitation came in part out of a conception of criminal conduct as a symptom of an underlying condition that required treatment, rather than as bad conduct warranting punishment.15 This approach to crime influenced the sentencing reform of the late l9th century, generating indeterminate sentencing policies that dominated adult criminal corrections for more than a century.16 Juvenile offenders, because they were young and malleable, were believed to be ideally suited to a regime grounded in rehabilitation.17 In juveniles, the "condition" that required treatment was caused by poor parental guidance, care and supervision as well as social harms associated with poverty.18

Although the Progressive conception of youth included the modern understanding of adolescence as a separate developmental stage, and of adolescents as "not fully formed" persons,19 the Progressive reformers tended to describe young offenders in more childlike terms.20 As Judge Ben Lindsey, a prominent early reformer, put it, the criminal prosecution of youth was an "outrage against childhood."21 Lindsey argued that "our laws against crime were as inapplicable to children as they would be to idiots."22 In part, this characterization may have reflected a tendency to exaggerate the differences between adult and youthful offenders, by advocates seeking to underscore the appropriateness of a more lenient legal response to youth crime. Moreover, the reformers, in fact, may have been imagining younger offenders. The jurisdiction of juvenile courts in the early years ended at age fourteen or sixteen, and the real victory for reformers was that children between the ages of seven and fourteen were no longer tried as adults.23

Under the rehabilitative model of juvenile justice, the immaturity of young defendants was important, for several reasons.24 Juvenile offenders were assumed to have different capacities and needs from adults (in part due to youthful vulnerability and dependence) that warranted separate adjudicatory procedures and a differential correctional response.25 The reformers also believed that the criminal acts of youthful offenders reflected their immaturity; thus, juveniles were not criminally responsible and should not be subject to the same punishment as adults.26 Neither the retributive or deterrent purposes of the criminal law were appropriately served with this group.

The qualities that were presumed to make juveniles less responsible than adults were not generally articulated with any precision. Since retribution is generally justified on the ground that criminal acts represent the free and rational choice to "do the bad thing,"28 a rejection of retribution implicitly reflected a view that juveniles were developmentally unable to make this choice. Some observers pointed to the impulsiveness and malleability of youth as traits that were the basis of limited criminal responsibility.29 Impulsiveness presumably contributes to incapacity because it impedes the ability to weigh the consequences of behavior, while malleability might make juveniles vulnerable to bad influences, particularly from peers.30

The rehabilitative approach of the traditional juvenile court presumed that state intervention could have either negative or positive effects on youthful offenders, and it emphasized the importance of preserving the future prospects of young offenders.31 The Progressives, however, did not maintain that criminal conduct reflected developmental influences that would attenuate with age. Rather, the belief was that the delinquent youth was on a path to a criminal career, from which he could be diverted, through rehabilitation, or toward which he would proceed without appropriate intervention.32

B. REFORMS IN THE WAKE OF GAULT

In the years since In re Gault33 extended to juveniles in delinquency proceedings many procedural rights accorded to criminal defendants, both the procedures and purposes of the juvenile court have been radically reformed. In part, the changes grew out of mounting skepticism about the empirical premise that rehabilitation was effective with youthful offenders.34 The reforms also reflected a growing belief that juveniles are more like adults than the traditional model recognized, and that young offenders should be held accountable for their offenses.35 Focus on the seriousness of the offense rather than the needs of the youthful offender came to dominate the sentencing regime in many jurisdictions.

In the post-Gault period, policymakers grappled with the challenge of constructing a retributive system that recognized the youth and immaturity of juvenile offenders. The reformers of the 1970s and 1980s thought that youths were more likely to offend than adults, and attributed high crime rates among juveniles to psychological and biological factors associated with this developmental stage.36 Under the post-Gault model of juvenile justice, juveniles were held responsible for criminal acts because they had sufficient moral judgment and capacity for self-control to justify this response. Moreover, adolescents needed to learn to take responsibility for their choices, in preparation for adulthood.38 However, because of youthful immaturity, the criminal choices of juveniles were seen as less culpable and thus as deserving less punishment than were those of adults.39 Thus, dispositional duration could be linked to age, because age was presumed to be linked to criminal responsibility.40 Differential treatment of adults and youth was also justified as a means to preserve the child's future options, a goal that presumed that delinquency did not necessarily presage adult criminality.41

Although the influence of developmental factors on criminal conduct was mostly presumed rather than analyzed, several characteristics of adolescents were understood by post-Gault reformers both to contribute to criminal conduct and to render adolescents less blameworthy. As compared to adults, minors were assumed to be more impulsive, to have less capacity for self control, to lack experience, and to be more inclined to focus on immediate rather than long-term consequences of their choices.42 Observers also pointed to the importance of peer approval (and peer pressure), rebellion against parental authority and restrictions, and the inclination of adolescents to experiment and engage in risk-taking as developmentally-linked factors contributing to youth crime.43 In combination, these traits were taken to characterize adolescence itself, and therefore to justify a conclusion that adolescent criminal conduct generally was less culpable than that of adults.44

These reform initiatives contributed to enormous changes in legal policy toward juveniles charged with crime. In the postGault generation, many legislatures have enacted retributive and more determinate sentencing provisions, in which the purpose of public protection is explicit and the seriousness of the present offense and the juvenile's prior record have become central factors.45 Nevertheless, the statutory revisions until recently generally reflected the reformers' core prescription limiting retribution-that most juveniles, because of developmental immaturity, should be subject to less punishment than adult criminals.

C. MODERN REFORMS-OBSCURING THE DEVELOPMENTAL LINES

The constraint on retribution that weighed so heavily with the post-Gault reformers has become considerably less important in recent years. The emphasis today is on social controlon protecting society from the harms inflicted by young offenders-and the clear trend has been toward imposing penalties on adolescents (especially those who commit violent crimes) that approximate sanctions imposed on adults.46 These policies explicitly or implicitly present adolescent offenders as indistinguishable from adults, and reject the importance of youthful immaturity in the legal response to youth crime. As one advocate of "get tough" policies put it, juvenile offenders "are criminals who happen to be young, not children who happen to be criminal."47

Some supporters of this trend challenge the notion that tough young criminals are less culpable or less mature than their adult counterparts, and argue that the juvenile justice system coddles young criminals. On this view, the traditional account of immature youngsters getting into scrapes with the law simply does not describe the savvy young offenders committing serious crimes today.48 Minors who inflict serious harm are deemed to be engaging in adult-like criminal conduct, and thus are presumed to be sufficiently mature to be tried and punished as adults. According to one critic, "[s]ociety may wish to be lenient with first offenders, particularly for lesser crimes, but there is no reason that society should be more lenient with a sixteen year old offender than a thirty year old offender."49

Several legislative strategies have been employed to treat juvenile offenders more like their adult counterparts. Many statutes provide for increasingly broad authority to adjudicate juveniles charged with serious offenses in adult criminal court, either through judicial transfer or through legislative waiver.50 The age of transfer has been lowered in many jurisdictions, and a broader range of felonies can lead to adjudication as an adult.51 Moreover, categories of youthful offenders increasingly are statutorily defined as adults and excluded from juvenile court jurisdiction based on offense and age (rather than traits of the juvenile).52 Currently, in many states, juveniles who have barely reached adolescence can be tried as adults and incarcerated in adult prisons.53 Another response is to provide stiff statutory minimum sentences in juvenile court for certain of fenses, often resulting in adult incarceration.54 Some critics of the juvenile justice system would push this trend to its logical extreme and abolish separate juvenile courts altogether.

Criminal courts also appear to have adopted the view that youth is not an important factor in distinguishing juveniles from adults. Despite some predictions to the contrary, adult criminal courts do not seem to respond to juvenile offenders more leniently than to adult criminals.55 Studies have found that adolescents charged with serious offenses are convicted at about the same rate as adults and, if convicted, receive sentences of similar severity.56

This approach to juvenile criminal conduct seems to rest on an assumption of adolescent competence, implicitly holding that there are no psychological differences between adolescent and adult offenders that are important to criminal responsibility or to participation in an adult criminal proceeding. Usually the issue is not addressed, and the argument for tough sanctions focuses on the harm caused by youthful predators.57 However, those who would treat young offenders as adults appear to assume that almost all juveniles meet any standard for maturity that might be relevant.

Despite the seemingly inexorable quality of the trend toward imposing full criminal responsibility on juvenile defendants, the account of adolescence that is embedded in the punitivist reforms has generated discomfort and controversy.58 Many observers continue to emphasize the importance of youthful immaturity as an important limitation on criminal responsibility.59 Even some proponents of the abolition of the juvenile court, such as Barry Feld, argue that minors are less capable of sound judgment, because of impulsiveness and a reduced capacity to appreciate the consequences of their acts, and thus are less culpable than adult offenders.60 Feld concludes that the immaturity of juvenile offenders can be accommodated at sentencing.61 Thus, the intuitions about the immaturity of youth that drove the early Progressive founders of the court and the post-Gault reformers continue to be expressed today. Currently, however, these concerns tend to be overwhelmed in the debate by accounts of the social costs of a system that fails to restrain young offenders whose age alone distinguishes them from adult criminals.

II. YOUTHFUL OFFENDING IN A DEVELOPMENTAL FRAMEWORK

Developmental psychology offers a useful perspective from which to examine and evaluate the changing conceptions of adolescence that have been reflected in the legal responses to juvenile crime during this century. It also clarifies the role that delinquent behavior may play in adolescence. In this Part, we clarify that many youths engage in criminal activity during adolescence but do not persist into adulthood-a phenomenon that can be explained most satisfactorily in developmental and social terms.62 We then narrow the focus to examine developmental influences on decision-making that may distinguish the choices of adolescents from those of adults. Our interest here is in factors that may affect the understanding, reasoning andperhaps most importantly-the judgment of youths who engage in criminal conduct. Finally, we explore how these developmental influences may affect choices made in the criminal justice context-both choices associated with criminal conduct and choices made by youthful defendants in the criminal process.

A. ANTISOCIAL BEHAVIOR AS A PART OF ADOLESCENCE

Substantial evidence indicates that many adolescents become involved in criminal activity in their teens and desist by the time they reach young adulthood. Beginning in early adolescence, criminal behavior increases through age sixteen and declines sharply from age seventeen onward.63 Self-report studies indicate that most teenage males engage in some criminal conduct, leading criminologists to conclude that participation in delinquency is "a normal part of teen life."64 For most adolescent delinquents, desistance from antisocial behavior also seems to be a predictable component of the maturation process.65 Only a small group of young offenders will persist in a life of crime.66

A representative sample of adolescents involved in criminal activity will include a large group whose antisocial conduct is "adolescence-limited" and a much smaller group whose conduct is "life-course-persistent."67 Although some youths in the latter group initiate antisocial behavior in adolescence, many display a variety of problem behaviors, beginning early in life and persisting through adolescence into adulthood.68 Of those whose adolescent delinquent conduct is a continuation of earlier antisocial behavior, many, although certainly not all, will become career criminals. However, as Moffitt points out, most youths who engage in delinquent conduct have little notable history of antisocial conduct in childhood; nor will the conduct continue into adulthood.69 Involvement in criminal activity and other antisocial behavior begins in adolescence and tends to follow a "natural onset and recovery process."70

The developmental forces that contribute to the onset and desistance of delinquent adolescent behavior are not well understood. Moffitt offers a plausible etiological theory under which the tendency of adolescents to engage in antisocial behavior can be understood as linked to the gap experienced by contemporary youth between early biological maturity and late social maturity and independence.71 Moffitt argues that adolescents are striving for elusive autonomy from parental and adult authority in a context in which most privileges of adult status are withheld. Many adolescents may be inclined to mimic their antisocial peers, who appear to have attained adult status in many ways.72 Through antisocial conduct, the adolescent attenuates the ties of childhood and demonstrates that he can act independently. Under Moffit's theory, youthful antisocial risktaking acts are personal statements of independence by individuals who are precluded from yet assuming legitimate adult roles.73 Desistance in young adulthood is explained under the theory as the adaptive response to changed contingencies, as more legitimate adult roles become available. Delinquent behavior becomes costly rather than rewarding, as many young adults perceive that it threatens now-available conventional opportunities and may foreclose future goals. In short, they come to realize that they have something to lose.74

B. REASONING AND JUDGMENT IN ADOLESCENT DECISION-MAKING

In this Part, we shift the focus from a general explanatory account of youthful involvement in crime to an examination of particular developmental influences on individual decisionmaking that may shape the choices of youthful actors in ways that distinguish them from adults.75 Most familiar, and of particular importance in assessing the competence of younger adolescents to participate in criminal proceedings, are elements of cognitive development-reasoning and understanding. More salient to decisions about participation in criminal conduct are psychosocial factors such as peer influence, temporal perspective (a tendency to focus on short-term versus long-term consequences), and risk perception and preference. These psychosocial factors may affect decision-making in powerful ways that may distinguish juveniles from adults. We designate these psychosocial influences as "judgment" factors, and argue that immature judgment in adolescence may contribute to choices about involvement in crime.76 This framework is largely consistent with Moffitt's theory, but focuses on internal dynamic influences on adolescent decision-making that are associated with this developmental stage, whereas Moffitt's emphasis is on changing external contingencies.

1. Cognitive Capacity: The Process of Decision-making

It is generally recognized that decision-making capacities increase through childhood into adolescence and that, although there is great variability among individuals, preadolescents and younger teens differ substantially from adults in their abilities. Development occurs along several lines. The capacities to process information and to think hypothetically develop into adolescence, and cognitive performance improves generally due to knowledge gained in specific domains. Moreover, cognitive skills acquired earlier mature and develop into adolescence.

The question of how adolescents' capacities for understanding and reasoning in making decisions compare with that of adults has received much attention among policy analysts and children's rights advocates in recent years-although largely not in the context of juvenile justice policy. Proponents of broader self-determination rights for minors, drawing on child development theory and empirical research, have argued that, by about age fourteen, adolescents' cognitive decision-making abilities are similar to those of adults.78 This argument holds that adolescents are capable of making informed and competent decisions about medical treatment and other matters, and should have the legal authority to do so.9 The evidence for these claims is drawn in part from Piaget's stage theory of cognitive development and from several empirical studies of minors' ability to understand and reason about medical and abortion decisions,81 to understand treatment issues in psychotherapy,82 and to understand their Fifth Amendment rights and the meaning of Miranda waivers.83

Together, this scientific research and theory support the claim that adolescents are more competent decision-makers than has been presumed under paternalistic policies, but the scientific evidence for the claim that their cognitive decisionmaking capacity is comparable to that of adults is unclear. We and others have made this argument elsewhere,84 and thus we will sketch only the most salient points. First, Piaget's strict stage theory of cognitive development is no longer accepted among cognitive psychologists.85 Further, the studies that support the claim of competence are small and mostly involve middie class subjects of average intelligence.86 Only a handful compare the decision-making of minors with that of adults.87 Finally, the studies of adolescent decision-making have been conducted in laboratory settings in which the decision is hypothetical and "pre-framed" in the sense that all relevant information is provided to the subjects.88 This format yields little useful data about how decisions are made in informal unstructured settings (such as the street), in which decision-makers must rely on their own experience and knowledge in making choices. Moreover, research offers little evidence of how minors may function relative to adults in stressful situations in which decisions have salience to their lives.89

In sum, scientific authority indicates that, in general, the cognitive capacity for reasoning and understanding of preadolescents and many younger teens differs substantially in some regards from that of older teens and adults. Tentative authority also supports the conclusion that, by mid-adolescence, youthful capacities for reasoning and understanding approximate those of adults. Whether and how these capacities are employed, however, may be quite variable, and adolescent performance is not necessarily like that of adults in various contexts. Because the research was largely undertaken in structured settings, the findings may be more useful in shedding light on questions about competence to stand trial than on cognitive capacity as it affects choices relevant to criminal conduct.90

2. Judgment Factors in Decision-making

Psychosocial developmental factors may also influence decision-making by adolescents in ways that are relevant to competence to stand trial and criminal responsibility. Particularly salient in this context might be factors such as (1) conformity and compliance in relation to peers, (2) attitude toward and perception of risk, and (3) temporal perspective. If these factors influence decision-making, the impact is not on cognitive competence, narrowly defined,91 but rather on "judgment" (as the term is used in common parlance).92 The traditional presumption (in juvenile justice and in many other policy areas) that minors are not fully accountable and need legal protection rests in part on a view that their judgment is immature. In essence, the intuition is that developmentally-linked predispositions and responses systematically affect decision-making of adolescents in ways that may incline them to make choices that threaten harm to themselves and to others. Whereas cognitive competence affects the process of decision-making, immature judgment is reflected in outcomes, in that developmental factors influence values and preferences, which in turn shape the cost-benefit calculus. The influence of these factors (peer influence, attitude toward risk, and temporal perspective) will change as the individual matures and values and preferences change-resulting in different choices.

a. Peer Influence

It is widely assumed that peer influence plays an important role in adolescent crime,93 and evidence supports the claim that teens are more subject to this influence than are adults.94 Peer influence seems to operate through two means: social comparison and conformity. Through social comparison, adolescents measure their own behavior by comparing it to others.95 Social conformity to peers, which peaks at about age fourteen, influences adolescents to adapt their behavior and attitudes to that of their peers.96 Peer influence could affect adolescent decisionmaking in several ways. In some contexts, adolescents might make choices in response to direct peer pressure. More indirectly, adolescent desire for peer approval could affect the choices made, without any direct coercion.97 Finally, as Moffitt suggests, peers may provide models for behavior that adolescents believe will assist them to accomplish their own ends.98

b. Attitude Toward Risk

Research evidence also indicates that adolescents differ from adults in their attitude toward and perception of risk.99 It is well established that adolescents and young adults generally take more risks with health and safety than do older adults by engaging more frequently in behavior such as unprotected sex, drunk driving and criminal conduct.100 This inclination may result because adolescents are less aware of risks than are adults,101 because they calculate the probability of risks differently, or because they value them differently.102 In some contexts, adolescent risk preferences may be linked to other developmental factors. For example, adolescents may be more averse than adults to risking social ostracism.

c. Temporal Perspective

Differences between adults and adolescents in attitude toward risk are related to differences in temporal perspective. Adolescents seem to discount the future more than adults do, and to weigh more heavily short-term consequences of decisions-both risks and benefits-a response that in some circumstances can lead to risky behavior.103 This tendency may be linked to the greater uncertainty that young people have about their future, an uncertainty that makes short-term consequences seem more salient.104 It may also reflect the difference in experience between teens and adults. It may simply be harder for an adolescent than for an adult to contemplate the meaning of a consequence that will have an impact ten or fifteen years into the future.

In general, the fact that adolescents have less experience than adults seems likely to affect decision-making in tangible and intangible ways. Although the relative inexperience of adolescents has not been contested as a general proposition, the relevance of inexperience to decision-making and judgment is uncertain.

C. THE IMPACT OF DEVELOPMENTAL FACTORS ON DECISIONMAKING ABOUT OFFENDING

The research evidence on the impact of immaturity of reasoning and judgment on adolescent decision-making is sketchy, and, thus, assertions about the effects must be very tentative.105 With this caveat, it seems probable that developmental factors associated with adolescence could affect decision-making in several ways. First, adolescents may use information differently from adults. They may consider different or fewer options in thinking about their available choices or in identifying consequences when comparing alternatives.106 The extent and sources of the differences in use of information are unclear. It is plausible that dissimilar and more limited experience and knowledge, as well as attitudes toward risk, temporal perspective, and peer influence, are all implicated. These differences may be most evident in unstructured informal settings, where information is not provided, and individuals must make choices based on their own knowledge and experience.107 Thus, adolescents on the street, who are making choices that lead to criminal conduct, may be less able than adults to consider alternative options that could extricate them from a precarious situation. Secondly, substantial theoretical arguments hold that while older adolescents may have adult-like capacities for reasoning, they may not deploy those capacities as uniformly across different problem-solving situations as do adults,108 and they may do so less dependably in ambiguous or stressful situations.109 Finally, adolescents, for developmental reasons, could differ from adults in the subjective value that is assigned to perceived consequences in the process of making choices.110 Influenced by the developmental factors that we have described, adolescents may weigh costs and benefits differently (or view as a benefit what adults would count as a cost).111

These developmentally driven differences could be important to choices about participation in crime. Consider the following example. A youth hangs out with his buddies on the street. Someone suggests holding up a nearby convenience store. The boy's decision to go along with the plan may proceed in the following way. He has mixed feelings about the proposal, but doesn't think of ways to extricate himself-although perhaps a more mature person might develop a strategy. The possibility that one of his friends has a gun and the consequences of that may not occur to him. He goes along, mostly because he fears rejection by his friends, a consequence that he attaches to a decision not to participate-and that carries substantial negative weight. Also the excitement of the hold-up and the possibility of getting some money are attractive. These weigh more heavily in his decision than the cost of possible apprehension by the police, or the long-term costs to his future life of conviction of a serious crime.112

The example presents, in our view, a quite plausible account of the influence of the developmental factors that we have described on the decisions of adolescents to engage in criminal conduct.113 The choice, although it may reflect a lack of knowledge and experience, may not be irrational, because the youth is choosing the option that promotes subjective utility, given his values. The decision does, however, implicate immaturity of judgment-at least from a societal perspective-because it causes harm to a victim, and threatens harm to the youth himself.

If these influences on decision-making are developmental and not simply reflective of individual idiosyncratic preferences for risk-taking, they should abate with maturity. This prediction is consistent with the pattern of desistance from delinquent conduct in late adolescence or early adulthood that we have described.114 In general, it is reasonable to argue that the developmental factors of peer influence, temporal perspective, and risk perception and preference contribute to delinquent behavior and that their declining influence contributes to desistance. Although the factors contributing to desistance have not been adequately studied, researchers have linked desistance in late adolescence to a longer-term perspective and to changing patterns of peer relationships.115 As Moffitt postulates, young adults may cease to commit crimes because they come to understand that the decision to offend carries the risk of lost future opportunities.116 In other words, a cost-benefit calculus leads to a conclusion that choosing crime no longer maximizes subjective utility.

What is not clear is whether the source of the change is exogenous or endogenous. Moffitt seems to suggest that the calculus shifts because external contingencies change.117 A focus on psychosocial influences that contribute to immaturity of judgment suggests that desistance can be linked to developmental maturation. Based on the developmental research and theory that we have described, we are inclined to believe that much adolescent participation in crime is the result of interaction between developmental influences on decision-making and external contingencies that affect individuals during this stage. Very recent studies have begun to demonstrate the relationship between psychosocial developmental factors and the quality of youths' choices compared to those of adults.118 Further research is required, however, to determine the degree to which these developmental factors directly affect the decisions of youths participating in delinquent activity.

D. THE IMPACT OF DEVELOPMENTAL FACTORS ON DECISIONMAKING IN THE CRIMINAL PROCESS

Increasingly, under the legislative reforms of recent years, youths who are charged with crimes are tried as adults, or face severe punishment for their offenses even if they are tried in juvenile court.119 In this context, it becomes relevant to ask whether developmental factors are likely to influence the capacities of youthful defendants to participate in their defense. The importance of this inquiry is clear in light of the constitutional requirement that criminal defendants be competent to participate in the proceedings against them.120 Historically, questions of competence to stand trial have arisen in cases involving mentally ill and mentally retarded defendants. As younger defendants face adult criminal proceedings and punishment, cases involving trial incompetence due to immaturity are likely to increase.

Richard Bonnie has described two broad types of abilities associated with the defendants' legal competence to participate in criminal proceedings.121 The first is the capacity to assist counsel. This involves the defendant's ability to understand and appreciate the meaning of the legal procedure and her rights within that process, and the ability to assist counsel in developing a defense.122 The second concept is decisional competence, referring to defendants' capacities for reasoning and judgment needed to make decisions in the process, including decisions to waive important rights.123 In these areas of ability, developmental immaturity may impede the capacity of juvenile defendants to participate in criminal proceedings. As compared to the scant empirical data about youthful decisions to offend,TM substantial research has examined different dimensions of juveniles' capacities as they could affect their participation in trials. A recent comprehensive review of research suggests that delinquent adolescents are at risk of being less competent participants in their defense then are adults, and that this risk is especially great for youths under the age of fourteen.l25

Studies on youths' understanding of matters related to trials, such as the roles of participants and the trial process, have found that youths under the age of fourteen typically are deficient in their knowledge of the legal process and its basic purposes.126 In contrast, few differences in basic understanding of trial-related matters have been observed between adolescents fourteen to seventeen years of age and adults, when the populations studied were "average" adolescents. Similarly, a conventional grasp of the nature of legal rights typically has developed by mid-adolescence. Moreover, fundamental abilities of sensation, perception, and memory ordinarily have matured by early adolescence, suggesting that adolescents on average should be as capable as adults of providing accurate information to their attorneys from their experience.

Delinquent youths, however, are more likely than average to have disabilities-for example, emotional disturbances, learning and attention deficit disorders, or poorer intellectual capacities-that may contribute to delays in the development of capacities for understanding, communication, and the ability to attend to the trial process as it unfolds.128 Studies of delinquent youths' understanding of the trial process and capacity to assist counsel have found important deficiencies, often distinguishing these juveniles from adults and from "average" adolescents. Compared to adults, both delinquent and non-delinquent adolescents who have lower intelligence test scores, problematic educational histories, learning disabilities, and mental disorders have shown poorer comprehension of basic information about the legal process.129 Other evidence has suggested that delinquent youths' experience with courts, attorneys, and law enforcement officers does not reliably compensate for these tendencies toward poorer understanding of information related to the trial process and rights.130

Defendants must also be able to make decisions in the trial process, and thus, reasoning and judgment capacities are implicated, as well as understanding. The direct evidence is sketchy about how adolescents compare with adults in their capacities to reason about important legal decisions and in their valuation of the consequences of those decisions. As we have described, capacities associated with reasoning and problem solving are in a formative stage in the primary school years.131 One study, for example, found that youths between ten and thirteen-years-ofage were significantly less likely than older adolescents to think "strategically" about pleading decisions, when hypothetical conditions varied as to evidence of guilt and seriousness of accusations.132 Research finding younger adolescents to be less capable of imagining risky consequences during hypothetical problem solving,133 and to consider a more constricted number and range of consequences,134 also suggests that they may have difficulty considering the merits of plea agreements and making other decisions about their defense. As we have indicated, most studies have found few differences between older adolescents (ages fifteen to seventeen) and adults in formal decision-making functions.135 Again, however, most of these studies have involved non-delinquent youths, without documented disabilities, processing hypothetical rather than real decision problems (in nonstressful settings), focused on medical treatment rather than criminal or delinquency adjudication.136

Beyond the formal ability to understand and process information, youthful judgment may differ from that of adults in ways that could affect the ability to assist counsel and to make decisions. How defendants respond to attorneys' advice and weigh the consequences of their choices in the trial process may be affected by psychosocial factors such as peer and adult influence, temporal perspective, and risk preference and perception. Such differences might influence youths' judgments about the value of accepting plea bargains137 and of waiving important rights in the legal process. In one study, for example, delinquent youths, in considering the waiver of Miranda rights, focused more than adults on immediate consequences of waiver (release from custody) rather than the impact of the decision on later events in court.138 A more subtle issue is the effect of psychosocial factors on the attorney-client relationship, particularly on the inclination to trust the attorney and to value her advice, as compared, for example, to advice from peers.

In the trend toward treating more and younger juveniles charged with crimes as adults, little attention has been paid to whether youthful defendants can competently participate in the process. Yet it is uncontroversial that competence is a critical requirement for fair criminal proceedings. The research evidence suggests that many youths may be less competent than adults to assist counsel and make important decisions in their defense.

III. THE DEVELOPMENTAL FRAMEWORK AND CRIMINAL PUNISHMENT

A. CRIMINAL RESPONSIBILITY AND PSYCHOLOGICAL IMMATURITY

It seems likely that developmental influences relating to understanding and judgment affect adolescent choices to commit crime in ways that distinguish most young offenders from their adult counterparts. Although individual adult and juvenile offenders may vary, decision-making factors associated with age that affect decision-making provide a basis for differentiating between the two categories of offenders. Moreover, the fact that delinquent behavior desists for most adolescents as they approach adulthood strongly suggests that criminal conduct, for most youths, is associated with factors peculiar to adolescence. Thus, evidence from developmental psychology challenges the account of adolescence offered by the modern punitive reformers, who generally discount the idea that relevant differences exist between youthful and adult offenders. This evidence also suggests that the Progressives' account of youth involved in crime as childlike and blameless actors in need of treatment is skewed. In our view, when analyzed in the framework of conventional criminal excuse doctrine, the developmental evidence supports a presumption of diminished responsibility for adolescent offenders-but not a lack of responsibility.

Analysis of how the differences between youthful and adult offenders should count in formulating a legal response to juvenile crime can usefully begin by looking to the accepted underpinnings of adult criminal responsibility. The criminal law posits that the offender is a rational actor, autonomously choosing "to do the bad thing" on the basis of personal values and preferences.139 The legitimacy of punishment is undermined if criminal choices depart substantially from this autonomy model.140 If youthful choices to offend are based on diminished ability to make decisions, or if the choices (or the values that shape the choices) are strongly driven by transient developmental influences, then the presumption of free will and rational choice is weakened. Psychology, in providing evidence that developmental psychosocial factors may shape decision-making well into adolescence, lends support to the intuitive conclusion that immature offenders are less culpable than their adult counterparts.

It is important, however, not to overstate the practical importance of whatever differences may exist between adolescents and adults.141 Although much rhetoric about free will and blameworthy choice surrounds the issue, criminal responsibility defenses are very narrowly drawn under established doctrinal standards for evaluating the quality of criminal choice. Many adult defendants whose choices reflect serious impairment, or coercion, or other exogenous influences are nevertheless convicted and punished for their crimes. For example, although mental disability and disorder are not uncommon among criminals,142 the modern insanity defense is available only to defendants with severe cognitive impairment that renders them unable to understand or appreciate the nature and quality of their wrongful conduct.143 Poverty and parental failure to inculcate appropriate norms are likely to exert strong influence on criminal choices, and are outside the offender's control, and yet are deemed irrelevant to adult culpability. In short, despite the theoretical commitment to requirements of free will and rational choice, in practice, only a small group of offenders whose choices are extremely compromised are fully excused from criminal liability.

When analyzed in this doctrinal framework, developmental differences affecting decision-making are likely to be substantial enough to provide categorical excuse from responsibility only for very young juveniles, who are qualitatively different from adults in moral, cognitive, and social development. Certainly by mid-adolescence, the differences between adolescent and adult decision-making are more subtle than those that distinguish the adult offenders whose crimes are excused under current law from other adult offenders. A categorical presumption of adolescent nonresponsibility, such as that which was endorsed by the traditional juvenile justice system, is hard to defend on grounds of immaturity alone.144

On the other hand, the perspective on adolescent criminal conduct offered by developmental psychology also challenges the retributive arguments of modern advocates of punitive policies. A claim that juvenile offenders deserve equivalent punishment to that imposed on adults presumes that no substantial differences exist that undermine the legitimacy of imposing equal measures of retribution on the two groups.l45 On this point, the evidence disputes the conclusion that most delinquents are indistinguishable from adults in any way that is relevant to culpability, and supports the creation of two distinct culpability categories-although, of course, there will be outlyers in both groups. In short, the predispositions and behavioral characteristics that are associated with the developmental stage of adolescence support a policy of reduced culpability for this category of offenders.

The developmental evidence supports the argument of the post-Gault reformers of the 1970s and 1980s that a presumptive diminished responsibility standard be applied to juveniles. Moreover, other dimensions of the developmental picture strengthen the argument for reduced accountability. Because of inexperience and immature judgment, youths will make many mistakes during this period. Thus, as Franklin Zimring has argued, adolescence can usefully be conceptualized as a probationary period, in which young decision-makers learn to make responsible choices, without bearing the full costs of their mistakes.146 This approach argues for a legal response to crime that signals that choices have consequences, and that bad behavior is punished, but not as severely as is appropriate with older offenders (who have been given the opportunity to learn to make better choices) .147

Not surprisingly perhaps, some of these themes can be discerned in responses to adult offenders whose conduct suggests reduced culpability. The behavioral traits and inexperience that in general characterize youthful offenders as a group may be relevant for grading purposes at sentencing of individual adult criminals. The "immature" offender, whose acts were influenced by others, who is a first offender, or who "made a mistake" out of inexperience and can be expected to have "learned her lesson," may receive a reduced sentence, based on an official judgment that her crime is less culpable and deserves less punishment than does that of a seasoned criminal.148 In the case of adults, lesser punishment acknowledges an individual deficiency, a failure to attain an adult level of maturity and experience. For utilitarian reasons, however, such leniency will be cautiously exercised, lest the presumption of free will would completely collapse. Most adults are presumed to act on the basis of individual values and preferences. With minors, in contrast, criminal choices are presumed less to express individual preferences and more to reflect the behavioral influences characteristic of a transitory developmental stage that are generally shared with others in the age cohort. This difference supports drawing a line based on age, and subjecting adolescents to a categorical presumption of reduced responsibility.149

B. A SOCIAL COST PERSPECTIVE ON JUVENILE CRIME

Contrary to what we have called the "utilitarian assumption" underlying current policies, a developmental perspective suggests that the goal of crime reduction is also served by attending to developmental dimensions of criminal behavior. In the current climate of public anxiety about violent youth, many policymakers are persuaded that society must treat immature youthful offenders as a class more severely because of the magnitude of social harm caused by juvenile offenders.150 A developmental perspective reveals that, despite its superficial appeal, this position is short-sighted, even in terms of the reformers' objectives. On purely instrumentalist grounds, effective policy will attend to the developmental character of much youth crime, and to the divergent patterns of offending represented by the two groups described earlier-those youths whose criminal behavior is limited to adolescence, and those whose delinquency is part of a life-course persistent pattern.

1. The Instrumentalist Argument for Tough Sanctions

Political arguments for tough juvenile sanctions combine a retributive claim that juvenile and adult offenders deserve analogous punishment, with a utilitarian argument that such policies will reduce crime. Even if a strong claim that the criminal choices of adolescents and adults are the same is rejected, as we have suggested it should be, many are ready to discount the extent and relevance of developmental differences in assessing criminal responsibility.151 Although only the most fervent advocates of punitive policies are comfortable punishing children, regardless of age, like their adult counterparts, 152 many more conclude that the principles limiting criminal punishment are not unduly strained by ignoring the impact of immaturity on adolescent criminal choices.153

In the current climate, any modest developmental claim for leniency seems to be far outweighed by the importance of reducing the social threat of adolescent crime. In general, the criminal law balances autonomy-based constraints on retribution against the social cost of excusing conduct that inflicts harm. How the balance is struck depends in part on the magnitude of the perceived harm.154 Many observers believe that, although youthful immaturity may have served to justify a differential response to the criminal conduct of minors in a more peaceful time,155 today the stakes are too high to retain a system that sacrifices social protection.156 In response to the evidence that violent youth crime has escalated in recent years,157 the recent legal reforms have shifted the balance toward greater social protection.

From an instrumentalist perspective, the argument against counting youth as a mitigating factor in applying criminal sanctions seems to be even more compelling, because the developmental factors at issue are likely to contribute to the inclination to commit crimes. On this view, if adolescents tend to be riskpreferring actors who discount future consequences in favor of immediate gratification, they present a greater threat than do older offenders, and the social interest in constraint is more compelling.158 Through this lens, a system that responds leniently to its most dangerous offenders appears to be muddled and inefficient.

2. Utilitarian Policies Through a Developmental Lens

The assumption underlying the current trend holds that a rigorously punitive approach is the optimal means to accomplish the goal of reducing the social cost of youth crime. At one level the "cure" seems promising, given the objectives. Young offenders who are incarcerated in prison cannot be on the streets committing crimes. The developmental analysis in Part II, however, suggests that the utilitarian assumption is flawed and that punitive policies may not be the optimal means to limit the costs of juvenile crime.

A policy response designed to minimize social cost will recognize the substantial societal interest in facilitating desistance in delinquents whose crimes are adolescence-limited, and in preserving their future prospects. Most delinquent youths will grow into useful (or at least not criminal) citizens if they survive this stage without destroying their life chances. When and whether they emerge into productive adulthood is likely to be at least in part a function of the system's response to their adolescent criminal conduct.159 It seems likely, although it has not been demonstrated, that categorically imposing adult criminal penalties on adolescents will increase the likelihood that they will become career criminals, or at least that it may delay desistance.160 Moreover, even if desistance is not directly affected by a criminal sentence, the future educational, employment and social productivity of those youths whose crimes are adolescentlimited behavior is likely to be negatively affected, either directly or indirectly.161 Policy that attends to this developmental pattern would give these delinquent minors "room to reform," protecting them from the most severe consequences of choices based on immature judgment.162

A modest claim, based on our developmental analysis, is that the broad use of adult criminal penalties against youths may carry social costs that currently are being left out of the calculus, when such penalties are applied to normative adolescent offenders. The social benefits that proponents of tough penalties promise are not likely to be realized because two assumptions underlying the policies appear to be erroneous: that most youths who commit serious crimes are young career criminals; and that adult penalties will not generate severe iatrogenic effects that harm those who would otherwise outgrow their inclination to engage in criminal conduct.

As to the small group of delinquents whose conduct is likely to persist into adulthood, the picture is quite different. The research suggests that many adolescents in this group are likely to offend at an earlier age,163 and that their criminal conduct is frequent, chronic and more likely to be violent than is that of their delinquent peers.164 Thus, purely on social cost grounds, early and severe punitive sanctions may seem to be justified once a "differential diagnosis" identifies an offender as belonging to this group.165 However, such a response may be shortsighted, even aside from the formidable problem of false positive identification given our current level of knowledge,166 and the moral constraints on punishment of the youngest offenders.167 The enormous social cost inflicted by a lifelong career in crime-or by lifelong incarceration-argues for continuing to invest in developing early and comprehensive remedial interventions. Although knowledge currently is inadequate to the task, and simple prescriptions seem unlikely, a long-term strategy for dealing with this group that relies only on incarceration is not likely to be effective.

This conclusion does not signify that the optimal legal response to youth crime is "benign neglect." Although some sociologists have advocated a minimalist approach to intervention as the best means to avoid turning delinquents into career criminals,168 this is not a desirable response on several grounds. First, the prediction on which it is premised is simply wrong as applied to life-course-persistent offenders, the group threatening the greatest harm. As to the larger group of adolescent offenders, lessons in accountability are important. As Franklin Zimring has argued, adolescents need to learn from their foolish youthful choices so that they can successfully assume adult roles.169

IV. LESSONS FOR JUVENILE JUSTICE POLICY

The analysis to this point suggests that a developmental perspective may be usefully employed in formulating legal responses to juvenile crime. In this part, we will sketch a few policy lessons that we take from the analysis, and suggest directions for future innovation and research. Our aim is not to provide a detailed policy blueprint, but rather to suggest the contours of a system for responding to juvenile crime that is informed by developmental knowledge. We accept as a starting point that a central objective of any viable contemporary system is to reduce the social cost of youth crime, through means that conform to conventional limits on retribution and notions of procedural fairness. These goals function as constraints on the generation of policies under a developmental model of juvenile justice.

We take three core points from the developmental psychology evidence to be of potential importance to juvenile justice policy. First, on average, adolescents' decision-making abilities, especially related to judgment, have not matured to a level characteristic of adults. This gap is likely to be substantial with younger teens, and, for this group, is likely to include greater differences in understanding and reasoning as well as judgment. These differences are important both to culpability and to competence to stand trial. Second, younger adolescent offenders are likely to present a higher risk of becoming career criminals than are teens who initiate criminal conduct in midadolescence. And finally, delinquent behavior itself is shaped by developmental influences, and most adolescents desist as they age. The following sections suggest some ways in which an optimal legal response to youth crime will attend to these developmental considerations.

A. RESPONDING TO CRIMES OF YOUNGER ADOLESCENT OFFENDERS

1. Limiting Criminal Court Jurisdiction

The developmental evidence supports a conclusion that, as a general matter, younger teens are sufficiently different from adults in cognitive and psychosocial development that they should not be tried or punished in the adult criminal justice system. First, the argument for diminished responsibility is more compelling as applied to younger teens. Subjecting thirteenyear-old offenders to the same criminal punishment that is imposed on adults offends the principles that define the boundaries of criminal responsibility. Beyond this, the research indicates a substantial risk that many (perhaps most) younger adolescents may be substantially less competent to stand trial than are their adult counterparts.170 Thus, concerns about procedural fairness and about diminished culpability both point in the direction of a categorical exclusion of younger adolescents from adult criminal adjudication.171

The procedural fairness issue deserves further comment. The debate about juvenile justice reform has rarely focussed on this dimension of the trend toward trying ever-younger defendants as adults. Upon reflection, it presents formidable problems. The prohibition against trying incompetent defendants historically has been raised in cases involving mentally ill defendants, who, upon a finding of incompetence to stand trial, typically are hospitalized so that they may be restored to competence.172 It is quite unclear what should happen to a youthful defendant who is unable to participate adequately in his defense due to developmental immaturity. Should the criminal adjudication be suspended for some indefinite period of time until he matures? During the intervening period, should he be confined, even though he has not been convicted of a crime? If not, what is the alternative? The questions posed suggest the difficult issues that will arise if younger adolescents are routinely subject to adult criminal adjudication. It seems far wiser to minimize the problem by categorically excluding from adult criminal adjudication younger teens, whose immaturity is most likely to impair their competence to function as defendants in criminal courts.

2. Recognizing and Responding to High-Risk Juveniles

The conclusion that younger teens, because of their immaturity, should not be subject to criminal adjudication and punishment does not mean that a minimalist response to their offenses is indicated. As we have indicated, young adolescent offenders are at considerably greater risk of becoming chronic serious offenders than are those whose delinquency begins a few years later,173 because the criminal conduct of these younger teens is less likely to be part of a typical developmental process that will progress to a stage of natural recovery and desistance. Their behavior is more likely to reflect a nascent personality disorder, developing over time through the interaction of individual vulnerabilities and environmental factors.174 Currently, when younger teens first offend, we cannot discern those who will proceed to criminal careers. Further research is needed to distinguish these youths from young juveniles whose crimes reflect transient developmental influences.175 Nonetheless, preadolescent and early adolescent criminal conduct is a sufficiently important predictor so as to place these youths in a highrisk category that warrants serious attention.176

It is in responding to these young offenders that a revitalized concept of rehabilitation is important in a contemporary model of juvenile justice. Rehabilitative interventions with preadolescent and young adolescent offenders should be far more intensive than has been the traditional response. Indeed, in some sense, the traditional dispositional regime has it backwards, in its tendency to intervene minimally with younger of fenders and more assertively with older adolescents.177 Thus, a response that waits until these offenders are older and identified as chronic violent offenders misses the opportunity to remediate at a point when the possibility for rehabilitation is more promising than it will be later.178

It is fair to say that we have not yet developed effective rehabilitative interventions for these very young offenders, although a few delinquency programs have reported promising results in recent years.179 On the other hand, we also have not yet invested substantially in developing intensive and comprehensive interventions directed at the multiple problems that these youths face. Such an effort requires a serious long-term commitment of societal resources, and until it is undertaken, it cannot be said that "nothing works.180 Given the long term social cost of failure to make the investment-costs that include both the wasted lives of these youths and the harm they inflict on society-it seems well worth it.

In our view, these comprehensive interventions should explicitly be part of the response to juvenile crime, though the interventions may have educational, mental health, and social service components. Investment in intensive rehabilitative interventions that is closely linked to public protection from juvenile crime may be more palatable to legislators and to the public. Some proposals for juvenile justice reform have combined a "get tough" response to juvenile crime with recommendations for an investment in comprehensive services to younger children who demonstrate problem behaviors.181 It is surely true that a long-term solution requires such an investment. The problem is that, in the political process, the punitive policies survive and proposals for funding comprehensive services for younger children may never be implemented. A first step is to reinforce the link between social protection through crime reduction and intensive rehabilitation of very young offenders.

If comprehensive remediation efforts are unsuccessful, at some point society's interests in protection will dominate in determining the legal response. However, since rehabilitative intervention with persistently antisocial youths is unlikely to result in a "quick fix," this determination of failure should not be made prematurely. A commitment to rehabilitation reinforces the constraints on punishment of young offenders that we have described, making a powerful case against subjecting fourteenyear-old repeat offenders to adult criminal punishment.

B. INTERVENTIONS WITH ADOLESCENCE-LIMITED OFFENDERS

An important insight of the developmental analysis is that neither the rehabilitative model (with its treatment focus) nor the criminal justice model (with its single minded focus on punishment) provides the conceptual tools to respond effectively to a large category of adolescent offenders: those whose first offenses occur in the mid-adolescent years. In the contemporary context, the most important practical implication of adopting the developmental framework is the recognition that severe sanctions imposed on adolescents whose crimes reflect transient developmental influences are unlikely to serve the interest of either society or of offenders. In policy terms, this should translate into a presumption against adult criminal adjudication and sanctions for first offenses by these juveniles, even for serious crimes. Contrary to the assumption of current law, if reducing social cost is important, the case for this more "lenient" response holds as powerfully for older (and thus presumably more competent) youths as for younger adolescents. Indeed, as we have demonstrated, the mid-adolescent first offender with no prior history of problem behavior is less likely than his younger counterpart (especially with such a history) to represent a substantial threat to society later in adulthood.

Under a developmental model, delinquency interventions directed at adolescents whose crimes are likely to reflect transient developmental influences serve multiple purposes. First, a developmental model recognizes the importance of lessons in accountability; slaps on the wrist fail to serve this purpose. Second (and linked to the first), the objective of protecting society would not be discounted. The fact that many youthful offenders will desist in their criminal activity as they mature does not justify a license to offend during adolescence. Third, the systemic response would be tailored to protect rather than to damage the prospects for a productive future of adolescents whose desistance is probable.

The future opportunities of juvenile offenders can be protected in several ways, including some that were designed to serve this purpose under the traditional juvenile justice system. Thus, policies of maintaining the anonymity of juvenile defendants in the press and of giving accused juveniles the right to choose a closed hearing may limit the stigma of delinquency status and its lasting impact.l82 The sealing of juvenile justice records reduces the likelihood that the young adult who has desisted from crime will be haunted by the mistakes of his youth in employment and educational contexts.183 Dispositional programs that emphasize education and the acquisition of job skills will encourage future productivity and may facilitate the process of assuming adult roles.

The developmental model also suggests directions for future innovations in juvenile justice research and policy. First, criminological research to date has focused on the etiology of delinquency; little attention has been directed toward the process of "maturing out" of delinquency or of the mechanisms that contribute to desistance.184 The developmental framework clarifies the importance of seeking to understand this process and of beginning to develop interventions that may accelerate desistance-or at a minimum that do not delay the process.185

A very different type of policy initiative would draw on developmental knowledge about adolescent decision-making to structure incentives that could discourage participation in criminal activity. Currently, we can only hypothesize about the influences on decision-making in this context, based on knowledge of traits that shape adolescent choices in other settings. However, a better understanding of the process by which adolescents make choices to participate in crime could be useful in formulating prophylactic responses. Indeed, some current policies seem to be directed toward deterrence, based on intuitions about the influences affecting juvenile choices. For example, curfew regulations discourage youths from gathering together at night, partly in recognition of the role that peer influence may have on individual choice. A more sophisticated understanding of decision-making may contribute to more effective efforts to influence youthful choices about involvement in crime.

C. DIFFERENTIAL TREATMENT IN A SEPARATE SYSTEM

The most effective means to implement the lessons from developmental psychology is to maintain a system of adjudication and disposition that is separate from the adult criminal justice system. First, a juvenile court can better recognize and accommodate the reduced culpability and more limited trial competence of younger offenders. Moreover, a separate juvenile correctional system is more likely to utilize dispositional strategies, goals, and approaches that are grounded in developmental knowledge. Commitment to the development of intensive rehabilitative interventions for young offenders and to the protection of the future prospects of youths whose crimes are adolescent-limited offers the long term promise of lowering the social cost of youth crime more effectively than the blanket punitive policies that are currently in vogue. We think that it is unlikely that either juvenile offenders or society will be better served by a unified criminal justice system-even one that treats minors more leniently for sentencing purposes.186 The ability or inclination of the criminal justice system to tailor its response to juvenile crime so as to utilize the lessons of developmental psychology is questionable. The evidence suggests that political pressure functions as a one-way ratchet, in the direction of everstiffer penalties. Programs designed for adolescents and sentencing distinctions between adults and juveniles will be much harder to maintain in a unified system in which juveniles are otherwise treated as adults; it seems predictable that the lines between age groups will become blurred.

The argument for separate treatment of juvenile offenders becomes weaker when the offender is older and persists in serious criminal activity. Policies regarding the adjudication of chronic serious adolescent offenders in the criminal justice system are subject to considerable debate, which generally reflects political and ideological differences. The empirical evidence and theoretical insights from developmental psychology and criminology provide no answers about the contours of an optimal policy. For a broad range of juvenile offenders however, utilitarian and retributive arguments converge to support adjudication and disposition in a separate juvenile justice system. Such a system, grounded in developmental principles, not only can function more coherently and effectively to achieve the complex societal objectives at stake, but also stands as a powerful symbol that most young offenders are different from their adult counterparts.