Blame and danger: An essay on preventive detention

Boston University Law Review

Volumn 76, Issue 1-2, 1996, Pages 113-154

  Morse, Stephen J ^a  

^a NONE

Author keywords

[No Author keywords available]

Indexed keywords

EID: 0030529814     PISSN: 00068047     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Review

References (194)

1. Ferdinand Wakeman Hubbell Professor of Law and Professor of Psychology and Law in Psychiatry, University of Pennsylvania. I should like to thank Michael Davis, James Jacobs, Howard Lesnick, John Monahan, Dan Polsby, Kevin Reitz, and Pete Wales for sharing their very helpful thoughts with me. John Monahan deserves special thanks for teaching me so much about behavioral prediction and other matters. I have also been stimulated by thoughtful exploration of the issues by Michael Corrado, Paul Robinson, Ferdinand Schoeman, and Christopher Slobogin. See Michael Corrado, Punishment and the Wild Beast of Prey: The Problem of Preventive Detention, 86 J. CRIM. L. & CRIMINOLOGY (forthcoming 1996); Paul H. Robinson, Foreword to The Criminal-Civil Distinction and Dangerous Blameless Offenders, 83 J. CRIM. L. & CRIMINOLOGY 693 (1993); Ferdinand D. Schoeman, On Incapacitating the Dangerous, 16 AM. PHIL. Q. 27 (1979); Christopher Slobogin, Dangerousness as a Criterion in the Criminal Process, in LAW, MENTAL HEALTH, AND MENTAL DISORDER (Bruce Sales & Daniel Shuman eds., forthcoming 1996). Finally, thanks to Randy Barnett for his thoughtful and thought-provoking commentary. Randy E. Barnett, Getting Even: Restitution, Preventive Detention, and the Tort/Crime Distinction, 76 B.U. L. REV. 157 (1996). It is boilerplate but true that all errors are mine.
2. Ferdinand Wakeman Hubbell Professor of Law and Professor of Psychology and Law in Psychiatry, University of Pennsylvania. I should like to thank Michael Davis, James Jacobs, Howard Lesnick, John Monahan, Dan Polsby, Kevin Reitz, and Pete Wales for sharing their very helpful thoughts with me. John Monahan deserves special thanks for teaching me so much about behavioral prediction and other matters. I have also been stimulated by thoughtful exploration of the issues by Michael Corrado, Paul Robinson, Ferdinand Schoeman, and Christopher Slobogin. See Michael Corrado, Punishment and the Wild Beast of Prey: The Problem of Preventive Detention, 86 J. CRIM. L. & CRIMINOLOGY (forthcoming 1996); Paul H. Robinson, Foreword to The Criminal-Civil Distinction and Dangerous Blameless Offenders, 83 J. CRIM. L. & CRIMINOLOGY 693 (1993); Ferdinand D. Schoeman, On Incapacitating the Dangerous, 16 AM. PHIL. Q. 27 (1979); Christopher Slobogin, Dangerousness as a Criterion in the Criminal Process, in LAW, MENTAL HEALTH, AND MENTAL DISORDER (Bruce Sales & Daniel Shuman eds., forthcoming 1996). Finally, thanks to Randy Barnett for his thoughtful and thought-provoking commentary. Randy E. Barnett, Getting Even: Restitution, Preventive Detention, and the Tort/Crime Distinction, 76 B.U. L. REV. 157 (1996). It is boilerplate but true that all errors are mine.
3. Ferdinand Wakeman Hubbell Professor of Law and Professor of Psychology and Law in Psychiatry, University of Pennsylvania. I should like to thank Michael Davis, James Jacobs, Howard Lesnick, John Monahan, Dan Polsby, Kevin Reitz, and Pete Wales for sharing their very helpful thoughts with me. John Monahan deserves special thanks for teaching me so much about behavioral prediction and other matters. I have also been stimulated by thoughtful exploration of the issues by Michael Corrado, Paul Robinson, Ferdinand Schoeman, and Christopher Slobogin. See Michael Corrado, Punishment and the Wild Beast of Prey: The Problem of Preventive Detention, 86 J. CRIM. L. & CRIMINOLOGY (forthcoming 1996); Paul H. Robinson, Foreword to The Criminal-Civil Distinction and Dangerous Blameless Offenders, 83 J. CRIM. L. & CRIMINOLOGY 693 (1993); Ferdinand D. Schoeman, On Incapacitating the Dangerous, 16 AM. PHIL. Q. 27 (1979); Christopher Slobogin, Dangerousness as a Criterion in the Criminal Process, in LAW, MENTAL HEALTH, AND MENTAL DISORDER (Bruce Sales & Daniel Shuman eds., forthcoming 1996). Finally, thanks to Randy Barnett for his thoughtful and thought-provoking commentary. Randy E. Barnett, Getting Even: Restitution, Preventive Detention, and the Tort/Crime Distinction, 76 B.U. L. REV. 157 (1996). It is boilerplate but true that all errors are mine.
4. Ferdinand Wakeman Hubbell Professor of Law and Professor of Psychology and Law in Psychiatry, University of Pennsylvania. I should like to thank Michael Davis, James Jacobs, Howard Lesnick, John Monahan, Dan Polsby, Kevin Reitz, and Pete Wales for sharing their very helpful thoughts with me. John Monahan deserves special thanks for teaching me so much about behavioral prediction and other matters. I have also been stimulated by thoughtful exploration of the issues by Michael Corrado, Paul Robinson, Ferdinand Schoeman, and Christopher Slobogin. See Michael Corrado, Punishment and the Wild Beast of Prey: The Problem of Preventive Detention, 86 J. CRIM. L. & CRIMINOLOGY (forthcoming 1996); Paul H. Robinson, Foreword to The Criminal-Civil Distinction and Dangerous Blameless Offenders, 83 J. CRIM. L. & CRIMINOLOGY 693 (1993); Ferdinand D. Schoeman, On Incapacitating the Dangerous, 16 AM. PHIL. Q. 27 (1979); Christopher Slobogin, Dangerousness as a Criterion in the Criminal Process, in LAW, MENTAL HEALTH, AND MENTAL DISORDER (Bruce Sales & Daniel Shuman eds., forthcoming 1996). Finally, thanks to Randy Barnett for his thoughtful and thought-provoking commentary. Randy E. Barnett, Getting Even: Restitution, Preventive Detention, and the Tort/Crime Distinction, 76 B.U. L. REV. 157 (1996). It is boilerplate but true that all errors are mine.
5. Ferdinand Wakeman Hubbell Professor of Law and Professor of Psychology and Law in Psychiatry, University of Pennsylvania. I should like to thank Michael Davis, James Jacobs, Howard Lesnick, John Monahan, Dan Polsby, Kevin Reitz, and Pete Wales for sharing their very helpful thoughts with me. John Monahan deserves special thanks for teaching me so much about behavioral prediction and other matters. I have also been stimulated by thoughtful exploration of the issues by Michael Corrado, Paul Robinson, Ferdinand Schoeman, and Christopher Slobogin. See Michael Corrado, Punishment and the Wild Beast of Prey: The Problem of Preventive Detention, 86 J. CRIM. L. & CRIMINOLOGY (forthcoming 1996); Paul H. Robinson, Foreword to The Criminal-Civil Distinction and Dangerous Blameless Offenders, 83 J. CRIM. L. & CRIMINOLOGY 693 (1993); Ferdinand D. Schoeman, On Incapacitating the Dangerous, 16 AM. PHIL. Q. 27 (1979); Christopher Slobogin, Dangerousness as a Criterion in the Criminal Process, in LAW, MENTAL HEALTH, AND MENTAL DISORDER (Bruce Sales & Daniel Shuman eds., forthcoming 1996). Finally, thanks to Randy Barnett for his thoughtful and thought-provoking commentary. Randy E. Barnett, Getting Even: Restitution, Preventive Detention, and the Tort/Crime Distinction, 76 B.U. L. REV. 157 (1996). It is boilerplate but true that all errors are mine.
6. Today, of course, few if any wardens would bother because the exercise would be considered futile. We are more cynical now, and with good reason.
7. See, e.g., Schoeman, supra note 1, at 27, 32-35 (focussing on moral issues regarding preventive detention, after assuming that the inadequacy of available predictive techniques is remediable).
8. See DANIEL C. DENNETT, ELBOW ROOM: THE VARIETIES OF FREE WILL WORTH WANTING 5-18 (1984) (discussing the misleading nature of "bugbear" hypotheticals in discussions of determinism and freedom); see also D.H.M. Brooks, The Method of Thought Experiment, 25 METAPHILOSOPHY 71, 82 (1994) (arguing that only "natural possibility" should constrain thought experiments that go beyond conceptual analysis).
9. See DANIEL C. DENNETT, ELBOW ROOM: THE VARIETIES OF FREE WILL WORTH WANTING 5-18 (1984) (discussing the misleading nature of "bugbear" hypotheticals in discussions of determinism and freedom); see also D.H.M. Brooks, The Method of Thought Experiment, 25 METAPHILOSOPHY 71, 82 (1994) (arguing that only "natural possibility" should constrain thought experiments that go beyond conceptual analysis).
10. Cf. PAUL M. CHURCHLAND, THE ENGINE OF REASON, THE SEAT OF THE SOUL: A PHILOSOPHICAL JOURNEY INTO THE BRAIN 309-14 (1995) (speculating on the legal and moral implications of a thoroughly materialist view of behavior and the technology that would ensue);
11. Andrew E. Lelling, Eliminative Materialism, Neuroscience and the Criminal Law, 141 U. PA. L. REV. 1471, 1526-39, 1560-64 (1993) (addressing the consequences for criminal law and criminal responsibility if neuroscience demonstrates that eliminative materialism is true).
12. I use the term "crazy" interchangeably with the term "mentally disordered," with no disrespect towards people with behavioral problems. For legal purposes, I prefer the term "crazy" to "mental disorder" and other technical terms because it is more descriptive of the phenomena in question and has fewer connotations about disease processes that beg important questions about responsibility.
13. See Schoeman, supra note 1, at 32 (setting forth the individual autonomy argument).
14. See, e.g., Corrado, supra note 1 (manuscript at 11) ("In [preventively detaining, the state] must take into account that detention not only prevents the future criminal activity, it prevents all sorts of unrelated lawful activity the criminal might have engaged in.").
15. Many statutes and some commentators would characterize the criterion as lack of "competence," but the terminology is not crucial. The important point is that the agent is not rational: both "nonresponsibility" and "incompetence" capture this idea.
16. See, e.g., CAL. WELF. & INST. CODE § 5250 (West Supp. 1995) (permitting involuntary civil commitment if a person is a danger to others or to himself or herself as a result of mental disorder).
17. See, e.g., 18 U.S.C.A. § 3142(e) (West Supp. 1995) (authorizing pretrial detention upon a finding that no condition or combination of conditions will reasonably assure the appearance of the person as required, the safety of any other person, and the safety of the community). The Supreme Court upheld the constitutionality of this statute in United States v. Salerno, 481 U.S. 739 (1987).
18. Compare Stephen J. Morse, A Preference for Liberty: The Case Against Involuntary Commitment of the Mentally Disordered, in CAROL A.B. WARREN, THE COURT OF LAST RESORT: MENTAL ILLNESS AND THE LAW 69, 97 (1982) (suggesting that the argument that crazy people lack effective liberty, especially compared to noncrazy people, is vastly overstated) with
19. WARREN, supra, at 202-03 (1982) (arguing that socioeconomic conditions deprive most citizens, including crazy people, of effective liberty).
20. A false positive is an erroneous prediction that an event will occur. In the present context, it refers to cases in which a charged defendant is denied bail because we predict that he would offend if released, but in fact he would not. Behavioral prediction need not be cast in the binary fashion indicated, however. There are many nonbinary methods for characterizing risk that are far more precise and convey greater information. Indeed, there have been major advances in conceptualizing, investigating, and communicating predictions in the last decade. See, e.g., William Gardner et al., A Comparison of Actuarial Methods for Identifying Repetitively Violent Patients with Mental Illnesses, 20 LAW & HUM. BEHAV. 35 (1996); John Monahan & Henry J. Steadman, Violent Storms and Violent People: How Meteorology Can Inform Risk Communication in Mental Health Law, 51 AM. PSYCHOLOGIST (forthcoming 1996); Douglas Mossman, Assessing Predictions of Violence: Being Accurate About Accuracy, 62 J. CONSULTING & CLINICAL PSYCHOL. 783 (1994); Edward P. Mulvey & Charles W. Lidz, Conditional Prediction: A Model for Research on Dangerousness to Others in a New Era, 18 INT'L J.L. & PSYCHIATRY 129 (1995); Marnie E. Rice & Grant T. Harris, Violent Recidivism: Assessing Predictive Validity, 63 J. CONSULTING & CLINICAL PSYCHOL. 737 (1995). Despite the undoubted advantages of the newer methods for investigating and expressing risk, there is nonetheless still virtue in reporting rates of false positives and negatives. No matter how the risk data for an individual are obtained and expressed, a preventive detention decision and its accuracy ex post are ultimately binary: The potential detainee either will or will not be preventively detained and, if undetained, will or will not do the feared harm. The decision to detain or not to detain may be justified ex ante because, for example, the person's score on an actuarial method, which precisely identifies the risk probability the person presents, exceeds or is lower than the level of risk that our society has determined justifies detention. It may seem strange to declare a case a "false positive" or a "false negative" when the risk assessment is expressed probabilistically and no specific prediction has been made. Nevertheless, the decision has binary consequences for potential detainees and society at large. A person who is detained and would not ultimately have behaved harmfully has unnecessarily been deprived of liberty. A person who is not detained and harms another has deprived the victim directly and society indirectly of liberty. Focusing on these outcomes in this way helps us keep in mind the human, moral costs of predictive enterprises.
21. A false positive is an erroneous prediction that an event will occur. In the present context, it refers to cases in which a charged defendant is denied bail because we predict that he would offend if released, but in fact he would not. Behavioral prediction need not be cast in the binary fashion indicated, however. There are many nonbinary methods for characterizing risk that are far more precise and convey greater information. Indeed, there have been major advances in conceptualizing, investigating, and communicating predictions in the last decade. See, e.g., William Gardner et al., A Comparison of Actuarial Methods for Identifying Repetitively Violent Patients with Mental Illnesses, 20 LAW & HUM. BEHAV. 35 (1996); John Monahan & Henry J. Steadman, Violent Storms and Violent People: How Meteorology Can Inform Risk Communication in Mental Health Law, 51 AM. PSYCHOLOGIST (forthcoming 1996); Douglas Mossman, Assessing Predictions of Violence: Being Accurate About Accuracy, 62 J. CONSULTING & CLINICAL PSYCHOL. 783 (1994); Edward P. Mulvey & Charles W. Lidz, Conditional Prediction: A Model for Research on Dangerousness to Others in a New Era, 18 INT'L J.L. & PSYCHIATRY 129 (1995); Marnie E. Rice & Grant T. Harris, Violent Recidivism: Assessing Predictive Validity, 63 J. CONSULTING & CLINICAL PSYCHOL. 737 (1995). Despite the undoubted advantages of the newer methods for investigating and expressing risk, there is nonetheless still virtue in reporting rates of false positives and negatives. No matter how the risk data for an individual are obtained and expressed, a preventive detention decision and its accuracy ex post are ultimately binary: The potential detainee either will or will not be preventively detained and, if undetained, will or will not do the feared harm. The decision to detain or not to detain may be justified ex ante because, for example, the person's score on an actuarial method, which precisely identifies the risk probability the person presents, exceeds or is lower than the level of risk that our society has determined justifies detention. It may seem strange to declare a case a "false positive" or a "false negative" when the risk assessment is expressed probabilistically and no specific prediction has been made. Nevertheless, the decision has binary consequences for potential detainees and society at large. A person who is detained and would not ultimately have behaved harmfully has unnecessarily been deprived of liberty. A person who is not detained and harms another has deprived the victim directly and society indirectly of liberty. Focusing on these outcomes in this way helps us keep in mind the human, moral costs of predictive enterprises.
22. A false positive is an erroneous prediction that an event will occur. In the present context, it refers to cases in which a charged defendant is denied bail because we predict that he would offend if released, but in fact he would not. Behavioral prediction need not be cast in the binary fashion indicated, however. There are many nonbinary methods for characterizing risk that are far more precise and convey greater information. Indeed, there have been major advances in conceptualizing, investigating, and communicating predictions in the last decade. See, e.g., William Gardner et al., A Comparison of Actuarial Methods for Identifying Repetitively Violent Patients with Mental Illnesses, 20 LAW & HUM. BEHAV. 35 (1996); John Monahan & Henry J. Steadman, Violent Storms and Violent People: How Meteorology Can Inform Risk Communication in Mental Health Law, 51 AM. PSYCHOLOGIST (forthcoming 1996); Douglas Mossman, Assessing Predictions of Violence: Being Accurate About Accuracy, 62 J. CONSULTING & CLINICAL PSYCHOL. 783 (1994); Edward P. Mulvey & Charles W. Lidz, Conditional Prediction: A Model for Research on Dangerousness to Others in a New Era, 18 INT'L J.L. & PSYCHIATRY 129 (1995); Marnie E. Rice & Grant T. Harris, Violent Recidivism: Assessing Predictive Validity, 63 J. CONSULTING & CLINICAL PSYCHOL. 737 (1995). Despite the undoubted advantages of the newer methods for investigating and expressing risk, there is nonetheless still virtue in reporting rates of false positives and negatives. No matter how the risk data for an individual are obtained and expressed, a preventive detention decision and its accuracy ex post are ultimately binary: The potential detainee either will or will not be preventively detained and, if undetained, will or will not do the feared harm. The decision to detain or not to detain may be justified ex ante because, for example, the person's score on an actuarial method, which precisely identifies the risk probability the person presents, exceeds or is lower than the level of risk that our society has determined justifies detention. It may seem strange to declare a case a "false positive" or a "false negative" when the risk assessment is expressed probabilistically and no specific prediction has been made. Nevertheless, the decision has binary consequences for potential detainees and society at large. A person who is detained and would not ultimately have behaved harmfully has unnecessarily been deprived of liberty. A person who is not detained and harms another has deprived the victim directly and society indirectly of liberty. Focusing on these outcomes in this way helps us keep in mind the human, moral costs of predictive enterprises.
23. A false positive is an erroneous prediction that an event will occur. In the present context, it refers to cases in which a charged defendant is denied bail because we predict that he would offend if released, but in fact he would not. Behavioral prediction need not be cast in the binary fashion indicated, however. There are many nonbinary methods for characterizing risk that are far more precise and convey greater information. Indeed, there have been major advances in conceptualizing, investigating, and communicating predictions in the last decade. See, e.g., William Gardner et al., A Comparison of Actuarial Methods for Identifying Repetitively Violent Patients with Mental Illnesses, 20 LAW & HUM. BEHAV. 35 (1996); John Monahan & Henry J. Steadman, Violent Storms and Violent People: How Meteorology Can Inform Risk Communication in Mental Health Law, 51 AM. PSYCHOLOGIST (forthcoming 1996); Douglas Mossman, Assessing Predictions of Violence: Being Accurate About Accuracy, 62 J. CONSULTING & CLINICAL PSYCHOL. 783 (1994); Edward P. Mulvey & Charles W. Lidz, Conditional Prediction: A Model for Research on Dangerousness to Others in a New Era, 18 INT'L J.L. & PSYCHIATRY 129 (1995); Marnie E. Rice & Grant T. Harris, Violent Recidivism: Assessing Predictive Validity, 63 J. CONSULTING & CLINICAL PSYCHOL. 737 (1995). Despite the undoubted advantages of the newer methods for investigating and expressing risk, there is nonetheless still virtue in reporting rates of false positives and negatives. No matter how the risk data for an individual are obtained and expressed, a preventive detention decision and its accuracy ex post are ultimately binary: The potential detainee either will or will not be preventively detained and, if undetained, will or will not do the feared harm. The decision to detain or not to detain may be justified ex ante because, for example, the person's score on an actuarial method, which precisely identifies the risk probability the person presents, exceeds or is lower than the level of risk that our society has determined justifies detention. It may seem strange to declare a case a "false positive" or a "false negative" when the risk assessment is expressed probabilistically and no specific prediction has been made. Nevertheless, the decision has binary consequences for potential detainees and society at large. A person who is detained and would not ultimately have behaved harmfully has unnecessarily been deprived of liberty. A person who is not detained and harms another has deprived the victim directly and society indirectly of liberty. Focusing on these outcomes in this way helps us keep in mind the human, moral costs of predictive enterprises.
24. A false positive is an erroneous prediction that an event will occur. In the present context, it refers to cases in which a charged defendant is denied bail because we predict that he would offend if released, but in fact he would not. Behavioral prediction need not be cast in the binary fashion indicated, however. There are many nonbinary methods for characterizing risk that are far more precise and convey greater information. Indeed, there have been major advances in conceptualizing, investigating, and communicating predictions in the last decade. See, e.g., William Gardner et al., A Comparison of Actuarial Methods for Identifying Repetitively Violent Patients with Mental Illnesses, 20 LAW & HUM. BEHAV. 35 (1996); John Monahan & Henry J. Steadman, Violent Storms and Violent People: How Meteorology Can Inform Risk Communication in Mental Health Law, 51 AM. PSYCHOLOGIST (forthcoming 1996); Douglas Mossman, Assessing Predictions of Violence: Being Accurate About Accuracy, 62 J. CONSULTING & CLINICAL PSYCHOL. 783 (1994); Edward P. Mulvey & Charles W. Lidz, Conditional Prediction: A Model for Research on Dangerousness to Others in a New Era, 18 INT'L J.L. & PSYCHIATRY 129 (1995); Marnie E. Rice & Grant T. Harris, Violent Recidivism: Assessing Predictive Validity, 63 J. CONSULTING & CLINICAL PSYCHOL. 737 (1995). Despite the undoubted advantages of the newer methods for investigating and expressing risk, there is nonetheless still virtue in reporting rates of false positives and negatives. No matter how the risk data for an individual are obtained and expressed, a preventive detention decision and its accuracy ex post are ultimately binary: The potential detainee either will or will not be preventively detained and, if undetained, will or will not do the feared harm. The decision to detain or not to detain may be justified ex ante because, for example, the person's score on an actuarial method, which precisely identifies the risk probability the person presents, exceeds or is lower than the level of risk that our society has determined justifies detention. It may seem strange to declare a case a "false positive" or a "false negative" when the risk assessment is expressed probabilistically and no specific prediction has been made. Nevertheless, the decision has binary consequences for potential detainees and society at large. A person who is detained and would not ultimately have behaved harmfully has unnecessarily been deprived of liberty. A person who is not detained and harms another has deprived the victim directly and society indirectly of liberty. Focusing on these outcomes in this way helps us keep in mind the human, moral costs of predictive enterprises.
25. See, e.g., Salerno, 481 U.S. at 747 (noting that the Bail Reform Act carefully limits the circumstances under which detention may be sought to the most serious crimes). Why a criminal charge should justify preventive detention generally is unclear. Preventive detention for potential danger related to the charged offense, such as to protect a threatened witness, is comprehensible. But if the charged defendant is simply dangerous generally, it is difficult to understand why preventive detention is more justifiable than for any other person posing an equal risk of similar harmdoing. The claim that the criminal charge is good evidence of dangerousness is an evidentiary argument rather than a principled reason to distinguish charged offenders. It is perfectly plausible, for example, that the con in the introductory example is more predictably dangerous than many people charged with serious offenses. Albert Alschuler argues that a probable cause determination of past misconduct does provide a predicate for prediction because it is some indication of culpability. As he recognizes, however, probable cause is not proof of guilt and it hardly seems sufficient to overcome the usual presumption against preventive detention. Albert W. Alschuler, Preventive Pretrial Detention and the Failure of Interest-Balancing Approaches to Due Process, 85 MICH. L. REV. 510, 532-34 (1986).
26. See, e.g., Salerno, 481 U.S. at 747 (noting that the Bail Reform Act carefully limits the circumstances under which detention may be sought to the most serious crimes). Why a criminal charge should justify preventive detention generally is unclear. Preventive detention for potential danger related to the charged offense, such as to protect a threatened witness, is comprehensible. But if the charged defendant is simply dangerous generally, it is difficult to understand why preventive detention is more justifiable than for any other person posing an equal risk of similar harmdoing. The claim that the criminal charge is good evidence of dangerousness is an evidentiary argument rather than a principled reason to distinguish charged offenders. It is perfectly plausible, for example, that the con in the introductory example is more predictably dangerous than many people charged with serious offenses. Albert Alschuler argues that a probable cause determination of past misconduct does provide a predicate for prediction because it is some indication of culpability. As he recognizes, however, probable cause is not proof of guilt and it hardly seems sufficient to overcome the usual presumption against preventive detention. Albert W. Alschuler, Preventive Pretrial Detention and the Failure of Interest-Balancing Approaches to Due Process, 85 MICH. L. REV. 510, 532-34 (1986).
27. See Allen v. Illinois, 478 U.S. 364, 377-79 (1986) (Stevens, J., dissenting) (describing the criminal character of Illinois's "sexually dangerous person" proceeding); cf. id. at 368-69 (majority opinion) (conceding that if the condition of confinement were clearly punitive, the "civil veil" might be pierced and criminal procedural protections would be required).
28. The term "traditional" is meant to exclude newer "sexual predator" commitments, which occur at the completion of an offender's prison term for a sex crime. For a discussion of these newer commitments, see infra Part III.C.
29. Mentally disordered sex offenders are seldom sufficiently mentally abnormal mentally abnormal to warrant either the insanity defense or traditional involuntary commitment. See discussion infra Part III.C.
30. See Allen, 478 U.S. at 375 (holding that proceedings under the Illinois Sexually Dangerous Persons Act are not proceedings in a "criminal case" within the meaning of the Fifth Amendment, and thus that Amendment's guarantee against compulsory self-incrimination does not apply).
31. See Jones v. United States, 463 U.S. 354, 367-69 (1983) (holding that: (1) a preponderance of the evidence standard for involuntary, indefinite commitment of insanity acquittees comports with due process, and (2) the length of an insanity acquittee's involuntary commitment depends upon his or her recovery, and thus the commitment may be of a much longer duration than an appropriate criminal sentence); see also Foucha v. Louisiana, 504 U.S. 71, 86 (1992) (holding that the continued involuntary detention of an insanity acquittee, no longer suffering from mental illness, on the basis that he cannot demonstrate that he is not dangerous to himself or to others, violates due process). This term of commitment must end when the patient either regains mental health or is no longer dangerous. See id. at 77 ("[T]he acquittee may be held as long as he is both mentally ill and dangerous, but no longer."). But see State v. Randall, 532 N.W.2d 94, 106-10 (Wis. 1995) (purporting to distinguish Foucha by holding that the continued commitment of an insanity acquittee who is no longer mentally ill does not violate due process if the acquittee is dangerous, there is a "medical justification" to continue the commitment, and the commitment does not exceed the maximum term of imprisonment that could have been imposed for the crime charged).
32. See Jones v. United States, 463 U.S. 354, 367-69 (1983) (holding that: (1) a preponderance of the evidence standard for involuntary, indefinite commitment of insanity acquittees comports with due process, and (2) the length of an insanity acquittee's involuntary commitment depends upon his or her recovery, and thus the commitment may be of a much longer duration than an appropriate criminal sentence); see also Foucha v. Louisiana, 504 U.S. 71, 86 (1992) (holding that the continued involuntary detention of an insanity acquittee, no longer suffering from mental illness, on the basis that he cannot demonstrate that he is not dangerous to himself or to others, violates due process). This term of commitment must end when the patient either regains mental health or is no longer dangerous. See id. at 77 ("[T]he acquittee may be held as long as he is both mentally ill and dangerous, but no longer."). But see State v. Randall, 532 N.W.2d 94, 106-10 (Wis. 1995) (purporting to distinguish Foucha by holding that the continued commitment of an insanity acquittee who is no longer mentally ill does not violate due process if the acquittee is dangerous, there is a "medical justification" to continue the commitment, and the commitment does not exceed the maximum term of imprisonment that could have been imposed for the crime charged).
33. See Jones v. United States, 463 U.S. 354, 367-69 (1983) (holding that: (1) a preponderance of the evidence standard for involuntary, indefinite commitment of insanity acquittees comports with due process, and (2) the length of an insanity acquittee's involuntary commitment depends upon his or her recovery, and thus the commitment may be of a much longer duration than an appropriate criminal sentence); see also Foucha v. Louisiana, 504 U.S. 71, 86 (1992) (holding that the continued involuntary detention of an insanity acquittee, no longer suffering from mental illness, on the basis that he cannot demonstrate that he is not dangerous to himself or to others, violates due process). This term of commitment must end when the patient either regains mental health or is no longer dangerous. See id. at 77 ("[T]he acquittee may be held as long as he is both mentally ill and dangerous, but no longer."). But see State v. Randall, 532 N.W.2d 94, 106-10 (Wis. 1995) (purporting to distinguish Foucha by holding that the continued commitment of an insanity acquittee who is no longer mentally ill does not violate due process if the acquittee is dangerous, there is a "medical justification" to continue the commitment, and the commitment does not exceed the maximum term of imprisonment that could have been imposed for the crime charged).
34. See Jones v. United States, 463 U.S. 354, 367-69 (1983) (holding that: (1) a preponderance of the evidence standard for involuntary, indefinite commitment of insanity acquittees comports with due process, and (2) the length of an insanity acquittee's involuntary commitment depends upon his or her recovery, and thus the commitment may be of a much longer duration than an appropriate criminal sentence); see also Foucha v. Louisiana, 504 U.S. 71, 86 (1992) (holding that the continued involuntary detention of an insanity acquittee, no longer suffering from mental illness, on the basis that he cannot demonstrate that he is not dangerous to himself or to others, violates due process). This term of commitment must end when the patient either regains mental health or is no longer dangerous. See id. at 77 ("[T]he acquittee may be held as long as he is both mentally ill and dangerous, but no longer."). But see State v. Randall, 532 N.W.2d 94, 106-10 (Wis. 1995) (purporting to distinguish Foucha by holding that the continued commitment of an insanity acquittee who is no longer mentally ill does not violate due process if the acquittee is dangerous, there is a "medical justification" to continue the commitment, and the commitment does not exceed the maximum term of imprisonment that could have been imposed for the crime charged).
35. See Robinson, supra note 1, at 706-08 (setting forth the traditional account against a criminal approach to preventive detention).
36. But see Lord Windlesham, Punishment and Prevention: The Inappropriate Prisoners, 1988 CRIM. L. REV. 140, 146-47 (arguing that the conditions in which civil detainees are confined are indistinguishable in most respects from criminal restrictions on liberty).
37. A legal finding that an agent is nonresponsible and dangerous is of course also stigmatizing, but this is not the stigma of moral culpability that in part creates the need for the greater procedural protections of the criminal justice system. See In re Winship, 397 U.S. 358, 363 (1970) (citing the stigma of criminal conviction as one justification for the reasonable doubt standard in criminal cases). But see Michael Davis, Arresting the White Death: Preventive Detention, Confinement for Treatment, and Medical Ethics, APA NEWSL., Spring 1995, at 92, 95 (arguing that the stigma in such cases is properly moral, because dangerousness is a type of reckless endangerment and is thus in itself wrongdoing).
38. A legal finding that an agent is nonresponsible and dangerous is of course also stigmatizing, but this is not the stigma of moral culpability that in part creates the need for the greater procedural protections of the criminal justice system. See In re Winship, 397 U.S. 358, 363 (1970) (citing the stigma of criminal conviction as one justification for the reasonable doubt standard in criminal cases). But see Michael Davis, Arresting the White Death: Preventive Detention, Confinement for Treatment, and Medical Ethics, APA NEWSL., Spring 1995, at 92, 95 (arguing that the stigma in such cases is properly moral, because dangerousness is a type of reckless endangerment and is thus in itself wrongdoing).
39. See, e.g., Joseph Goldstein & Jay Katz, Abolish the "Insanity Defense" - Why Not?, 72 YALE L.J. 853, 865-68 (1963) (arguing that the insanity defense is not a defense but rather a means to restrain some members of the group of defendants who would ordinarily be free of criminal liability).
40. In the remainder of this Paper, I shall refer to this possibility as "pure preventive detention," signalling that the commitment lacks a nonresponsibility or therapeutic rationale. Its purpose is solely to prevent predicted harm.
41. I have argued primarily on consequential grounds against any form of involuntary commitment for people with mental disorder. Stephen J. Morse, A Preference for Liberty: The Case Against Involuntary Commitment of the Mentally Disordered, 70 CAL. L. REV. 54 (1982). I still believe that argument is correct, but for the purpose of this discussion I will acknowledge both the acceptability and the inevitability of involuntary commitments for dangerousness.
42. See Bruce G. Link & Ann Stueve, Psychotic Symptoms and the Violent/Illegal Behavior of Mental Patients Compared to Community Controls, in VIOLENCE AND MENTAL DISORDER: DEVELOPMENTS IN RISK ASSESSMENT 137, 154 (John Monahan & Henry J. Steadman eds., 1994) (noting that the preponderance of recent research suggests that the mentally ill are "somewhat more likely to be violent" than those who are not mentally ill); Jeffrey W. Swanson, Mental Disorder, Substance Abuse, and Community Violence: An Epidemiological Approach, in VIOLENCE AND MENTAL DISORDER: DEVELOPMENTS IN RISK ASSESSMENT, supra, at 101, 119 (finding a weak but robust association between mental disorder and violence, an association that strengthens as the definitions of mental disorder and violence widen).
43. See Bruce G. Link & Ann Stueve, Psychotic Symptoms and the Violent/Illegal Behavior of Mental Patients Compared to Community Controls, in VIOLENCE AND MENTAL DISORDER: DEVELOPMENTS IN RISK ASSESSMENT 137, 154 (John Monahan & Henry J. Steadman eds., 1994) (noting that the preponderance of recent research suggests that the mentally ill are "somewhat more likely to be violent" than those who are not mentally ill); Jeffrey W. Swanson, Mental Disorder, Substance Abuse, and Community Violence: An Epidemiological Approach, in VIOLENCE AND MENTAL DISORDER: DEVELOPMENTS IN RISK ASSESSMENT, supra, at 101, 119 (finding a weak but robust association between mental disorder and violence, an association that strengthens as the definitions of mental disorder and violence widen).
44. See Link & Stueve, supra note 27, at 156 (concluding that certain psychotic symptoms that cause a person to feel threatened elevate rates of violent behavior in people with those symptoms).
45. See Swanson, supra note 27, at 119 (concluding that mentally disordered individuals who were also substance abusers are significantly more likely to be violent than those with mental disorder alone). Even in these cases, however, most such crazy people do not behave violently, and especially not seriously violently. But some do, and at rates higher than among noncrazy persons of similar background.
46. See Link & Stueve, supra note 27, at 149-53 (describing research results on the correlation between certain psychotic symptoms and violence).
47. In contrast, people with disorders who have a propensity to violent conduct that is unrelated to their disorder are not appropriate subjects for civil commitment because their violence is not irrationally motivated.
48. I have argued elsewhere that the ease of involuntary hospitalization may have the effect of discouraging the search for reasonable alternatives. See Morse, supra note 26, at 103. This concern has lessened in recent years, however, because cutbacks in funds allocated to state and community mental hospitals and clinics have both limited the availability of intensive inpatient treatment and enhanced the desirability of creating successful outpatient alternatives.
49. My use of the word "objective" here does not betoken allegiance to any form of metaphysical moral reality. It simply refers to the intersubjective agreement reasonably informed people might reach based on good reasons and good evidence, without which effective human interaction is impossible.
50. AMERICAN PSYCHIATRIC ASS'N, DIAGNOSTIC AND STATISTICAL MANUAL OF MENTAL DISORDERS 645-50 (4th ed. 1994) [hereinafter DSM-IV] (describing this disorder).
51. Id. at 175, 176-81 (substance dependence), 182-83 (substance abuse).
52. For example, the introduction to DSM-IV states: [E]ach of the mental disorders is conceptualized as a clinically significant behavioral or psychological syndrome or pattern that occurs in an individual and that is associated with present distress . . . or disability . . . or with a significantly increased risk of suffering death, pain, disability, or an important loss of freedom. In addition, this syndrome or pattern must not be merely an expectable and culturally sanctioned response to a particular event. . . . Whatever its original cause, it must currently be considered a manifestation of a behavioral, psychological, or biological dysfunction in the individual. Neither deviant behavior . . . nor conflicts that are primarily between the individual and society are mental disorders unless the deviance or conflict is a symptom of a dysfunction in the individual . . . . Id. at xxi-xxii. But see Jerome C. Wakefield, Disorder as Harmful Dysfunction: A Conceptual Critique of DSM-III-R's Definition of Mental Disorder, 99 PSYCHOL. REV. 232 (1992) (criticizing the definition of "disorder" in DSM-IV's predecessor, DSM-III-R, a definition that remains unchanged in DSM-IV).
53. Indeed, DSM-IV explicitly cautions against using its diagnostic criteria to resolve legal issues. See DSM-IV, supra note 34, at xxvii ("The clinical and scientific considerations involved in categorization of these conditions as mental disorders may not be wholly relevant to legal judgments . . . that take into account such issues as individual responsibility, disability determination, and competency.").
54. See PAUL S. APPELBAUM, ALMOST A REVOLUTION: MENTAL HEALTH LAW AND THE LIMITS OF CHANGE 210 (1994) (concluding that the consequences of mental health law reforms were much more limited than anticipated).
55. See Charles W. Lidz et al., The Accuracy of Predictions of Violence to Others, 269 JAMA 1007, 1009 (1993) (reporting: (1) 60% sensitivity (true positive proportion) for clinicians' predictions of violence, compared to an expected chance accuracy rate of 50%; and (2) no increase in accuracy with patients the clinicians judged to be especially dangerous). This study, which is one of the most methodologically sophisticated and persuasive to appear in the literature, demonstrates that clinicians are "relatively inaccurate predictors of violence." Id. at 1010. Clinicians' predictions of violence exceeded chance for male patients, but for female patients their predictions were not significantly better than chance. Id. Moreover, the sample in this study was particularly violent suggesting that the sensitivity of violence predictions would decrease in a random sample of emergency room patients. Id. at 1009. Finally, although the sample was probably unusually violent, only 14.4% of the violent occurrences involved serious violence. Id. The study also concluded that it was unclear whether using clinical judgment to predict violence was more accurate than simply using a history of violence Id. at 1010. For a second analysis of clinical predictive accuracy, see Mossman supra note 14, at 788-90 (reanalyzing 58 data sets from 44 disparate published studies which demonstrate that: (1) clinicians' predictions generally exceed chance, but that errors are inevitable; (2) past behavior is the best predictor of future violent conduct; and (3) nonclinicians furnished with data about past behavior may outperform clinicians who rely solely on data gained from a clinical interview); see also David B. Villeneuve & Vernon L. Quinsey, Predictors of General and Violent Recidivism Among Mentally Disordered Inmates, 22 CRIM. JUST. & BEHAV. 397, 406-09 (1995) (investigating a sample of high-risk, recidivist, mentally disorderd offenders for an average 92-month follow-up period after their release from a treatment facility, and finding: (1) 78.3% of the sample were rearrested, of which 49.6% were arrested for a violent offense; (2) an actuarial prediction instrument was a modest predictor of general recidivism, yielding a 31.7% relative improvement over chance; (3) only 16.7% of the initial rearrests were for a serious violent offense, a figure so low that accurate identification of this latter group using generally more accurate actuarial methods was precluded; and (4) the presence of psychosis decreased the rearrest risk for a violent offense). But see Robert Menzies & Christopher D. Webster, Construction and Validation of Risk Assessments in a Six-Year Follow-Up of Forensic Patients: A Tridimensional Analysis, 63 J. CONSULTING & CLINICAL PSYCHOL. 766, 775-76 (1995) (reporting on the basis of a careful, sophisticated study of a high-risk sample that: (1) neither actuarial nor clinical prediction is highly accurate and that little progress on the risk assessment of people with mental disorder has been made; (2) with few exceptions, the direct predictions of dangerousness "were almost universally invalid[;]" (3) even those risk assessors who were best at the task provided predictions that were "neither powerful nor pragmatically of any value[;]" and, (4) a handful of sociodemographic variables were far better predictors than clinical assessments and actuarial instruments). This study "demonstrates once more that the 'holy grail' of violence prediction is still far off." Id. at 775. In general, actuarial prediction is superior to clinical prediction. See generally Robyn M. Dawes et al., Clinical Versus Actuarial Judgment, 243 SCIENCE 1668 (1989). But highly accurate prediction of future violent conduct by any method, even among high risk groups, eludes us. Optimists about prediction sometimes overstate the accuracy rates found in more recent research, neglect the methodological flaws or cautions of newer work, lump all "violence" together, or fail to discuss adequately the trade-off between sensitivity and specificity. See, e.g., Alexander D. Brooks, The Constitutionality and Morality of Civilly Committing Violent Sexual Predators, 15 U. PUGET SOUND L. REV. 709, 735-54 (properly criticizing flaws in the older research, but making some optimist's errors).
56. See Charles W. Lidz et al., The Accuracy of Predictions of Violence to Others, 269 JAMA 1007, 1009 (1993) (reporting: (1) 60% sensitivity (true positive proportion) for clinicians' predictions of violence, compared to an expected chance accuracy rate of 50%; and (2) no increase in accuracy with patients the clinicians judged to be especially dangerous). This study, which is one of the most methodologically sophisticated and persuasive to appear in the literature, demonstrates that clinicians are "relatively inaccurate predictors of violence." Id. at 1010. Clinicians' predictions of violence exceeded chance for male patients, but for female patients their predictions were not significantly better than chance. Id. Moreover, the sample in this study was particularly violent suggesting that the sensitivity of violence predictions would decrease in a random sample of emergency room patients. Id. at 1009. Finally, although the sample was probably unusually violent, only 14.4% of the violent occurrences involved serious violence. Id. The study also concluded that it was unclear whether using clinical judgment to predict violence was more accurate than simply using a history of violence Id. at 1010. For a second analysis of clinical predictive accuracy, see Mossman supra note 14, at 788-90 (reanalyzing 58 data sets from 44 disparate published studies which demonstrate that: (1) clinicians' predictions generally exceed chance, but that errors are inevitable; (2) past behavior is the best predictor of future violent conduct; and (3) nonclinicians furnished with data about past behavior may outperform clinicians who rely solely on data gained from a clinical interview); see also David B. Villeneuve & Vernon L. Quinsey, Predictors of General and Violent Recidivism Among Mentally Disordered Inmates, 22 CRIM. JUST. & BEHAV. 397, 406-09 (1995) (investigating a sample of high-risk, recidivist, mentally disorderd offenders for an average 92-month follow-up period after their release from a treatment facility, and finding: (1) 78.3% of the sample were rearrested, of which 49.6% were arrested for a violent offense; (2) an actuarial prediction instrument was a modest predictor of general recidivism, yielding a 31.7% relative improvement over chance; (3) only 16.7% of the initial rearrests were for a serious violent offense, a figure so low that accurate identification of this latter group using generally more accurate actuarial methods was precluded; and (4) the presence of psychosis decreased the rearrest risk for a violent offense). But see Robert Menzies & Christopher D. Webster, Construction and Validation of Risk Assessments in a Six-Year Follow-Up of Forensic Patients: A Tridimensional Analysis, 63 J. CONSULTING & CLINICAL PSYCHOL. 766, 775-76 (1995) (reporting on the basis of a careful, sophisticated study of a high-risk sample that: (1) neither actuarial nor clinical prediction is highly accurate and that little progress on the risk assessment of people with mental disorder has been made; (2) with few exceptions, the direct predictions of dangerousness "were almost universally invalid[;]" (3) even those risk assessors who were best at the task provided predictions that were "neither powerful nor pragmatically of any value[;]" and, (4) a handful of sociodemographic variables were far better predictors than clinical assessments and actuarial instruments). This study "demonstrates once more that the 'holy grail' of violence prediction is still far off." Id. at 775. In general, actuarial prediction is superior to clinical prediction. See generally Robyn M. Dawes et al., Clinical Versus Actuarial Judgment, 243 SCIENCE 1668 (1989). But highly accurate prediction of future violent conduct by any method, even among high risk groups, eludes us. Optimists about prediction sometimes overstate the accuracy rates found in more recent research, neglect the methodological flaws or cautions of newer work, lump all "violence" together, or fail to discuss adequately the trade-off between sensitivity and specificity. See, e.g., Alexander D. Brooks, The Constitutionality and Morality of Civilly Committing Violent Sexual Predators, 15 U. PUGET SOUND L. REV. 709, 735-54 (properly criticizing flaws in the older research, but making some optimist's errors).
57. See Charles W. Lidz et al., The Accuracy of Predictions of Violence to Others, 269 JAMA 1007, 1009 (1993) (reporting: (1) 60% sensitivity (true positive proportion) for clinicians' predictions of violence, compared to an expected chance accuracy rate of 50%; and (2) no increase in accuracy with patients the clinicians judged to be especially dangerous). This study, which is one of the most methodologically sophisticated and persuasive to appear in the literature, demonstrates that clinicians are "relatively inaccurate predictors of violence." Id. at 1010. Clinicians' predictions of violence exceeded chance for male patients, but for female patients their predictions were not significantly better than chance. Id. Moreover, the sample in this study was particularly violent suggesting that the sensitivity of violence predictions would decrease in a random sample of emergency room patients. Id. at 1009. Finally, although the sample was probably unusually violent, only 14.4% of the violent occurrences involved serious violence. Id. The study also concluded that it was unclear whether using clinical judgment to predict violence was more accurate than simply using a history of violence Id. at 1010. For a second analysis of clinical predictive accuracy, see Mossman supra note 14, at 788-90 (reanalyzing 58 data sets from 44 disparate published studies which demonstrate that: (1) clinicians' predictions generally exceed chance, but that errors are inevitable; (2) past behavior is the best predictor of future violent conduct; and (3) nonclinicians furnished with data about past behavior may outperform clinicians who rely solely on data gained from a clinical interview); see also David B. Villeneuve & Vernon L. Quinsey, Predictors of General and Violent Recidivism Among Mentally Disordered Inmates, 22 CRIM. JUST. & BEHAV. 397, 406-09 (1995) (investigating a sample of high-risk, recidivist, mentally disorderd offenders for an average 92-month follow-up period after their release from a treatment facility, and finding: (1) 78.3% of the sample were rearrested, of which 49.6% were arrested for a violent offense; (2) an actuarial prediction instrument was a modest predictor of general recidivism, yielding a 31.7% relative improvement over chance; (3) only 16.7% of the initial rearrests were for a serious violent offense, a figure so low that accurate identification of this latter group using generally more accurate actuarial methods was precluded; and (4) the presence of psychosis decreased the rearrest risk for a violent offense). But see Robert Menzies & Christopher D. Webster, Construction and Validation of Risk Assessments in a Six-Year Follow-Up of Forensic Patients: A Tridimensional Analysis, 63 J. CONSULTING & CLINICAL PSYCHOL. 766, 775-76 (1995) (reporting on the basis of a careful, sophisticated study of a high-risk sample that: (1) neither actuarial nor clinical prediction is highly accurate and that little progress on the risk assessment of people with mental disorder has been made; (2) with few exceptions, the direct predictions of dangerousness "were almost universally invalid[;]" (3) even those risk assessors who were best at the task provided predictions that were "neither powerful nor pragmatically of any value[;]" and, (4) a handful of sociodemographic variables were far better predictors than clinical assessments and actuarial instruments). This study "demonstrates once more that the 'holy grail' of violence prediction is still far off." Id. at 775. In general, actuarial prediction is superior to clinical prediction. See generally Robyn M. Dawes et al., Clinical Versus Actuarial Judgment, 243 SCIENCE 1668 (1989). But highly accurate prediction of future violent conduct by any method, even among high risk groups, eludes us. Optimists about prediction sometimes overstate the accuracy rates found in more recent research, neglect the methodological flaws or cautions of newer work, lump all "violence" together, or fail to discuss adequately the trade-off between sensitivity and specificity. See, e.g., Alexander D. Brooks, The Constitutionality and Morality of Civilly Committing Violent Sexual Predators, 15 U. PUGET SOUND L. REV. 709, 735-54 (properly criticizing flaws in the older research, but making some optimist's errors).
58. See Charles W. Lidz et al., The Accuracy of Predictions of Violence to Others, 269 JAMA 1007, 1009 (1993) (reporting: (1) 60% sensitivity (true positive proportion) for clinicians' predictions of violence, compared to an expected chance accuracy rate of 50%; and (2) no increase in accuracy with patients the clinicians judged to be especially dangerous). This study, which is one of the most methodologically sophisticated and persuasive to appear in the literature, demonstrates that clinicians are "relatively inaccurate predictors of violence." Id. at 1010. Clinicians' predictions of violence exceeded chance for male patients, but for female patients their predictions were not significantly better than chance. Id. Moreover, the sample in this study was particularly violent suggesting that the sensitivity of violence predictions would decrease in a random sample of emergency room patients. Id. at 1009. Finally, although the sample was probably unusually violent, only 14.4% of the violent occurrences involved serious violence. Id. The study also concluded that it was unclear whether using clinical judgment to predict violence was more accurate than simply using a history of violence Id. at 1010. For a second analysis of clinical predictive accuracy, see Mossman supra note 14, at 788-90 (reanalyzing 58 data sets from 44 disparate published studies which demonstrate that: (1) clinicians' predictions generally exceed chance, but that errors are inevitable; (2) past behavior is the best predictor of future violent conduct; and (3) nonclinicians furnished with data about past behavior may outperform clinicians who rely solely on data gained from a clinical interview); see also David B. Villeneuve & Vernon L. Quinsey, Predictors of General and Violent Recidivism Among Mentally Disordered Inmates, 22 CRIM. JUST. & BEHAV. 397, 406-09 (1995) (investigating a sample of high-risk, recidivist, mentally disorderd offenders for an average 92-month follow-up period after their release from a treatment facility, and finding: (1) 78.3% of the sample were rearrested, of which 49.6% were arrested for a violent offense; (2) an actuarial prediction instrument was a modest predictor of general recidivism, yielding a 31.7% relative improvement over chance; (3) only 16.7% of the initial rearrests were for a serious violent offense, a figure so low that accurate identification of this latter group using generally more accurate actuarial methods was precluded; and (4) the presence of psychosis decreased the rearrest risk for a violent offense). But see Robert Menzies & Christopher D. Webster, Construction and Validation of Risk Assessments in a Six-Year Follow-Up of Forensic Patients: A Tridimensional Analysis, 63 J. CONSULTING & CLINICAL PSYCHOL. 766, 775-76 (1995) (reporting on the basis of a careful, sophisticated study of a high-risk sample that: (1) neither actuarial nor clinical prediction is highly accurate and that little progress on the risk assessment of people with mental disorder has been made; (2) with few exceptions, the direct predictions of dangerousness "were almost universally invalid[;]" (3) even those risk assessors who were best at the task provided predictions that were "neither powerful nor pragmatically of any value[;]" and, (4) a handful of sociodemographic variables were far better predictors than clinical assessments and actuarial instruments). This study "demonstrates once more that the 'holy grail' of violence prediction is still far off." Id. at 775. In general, actuarial prediction is superior to clinical prediction. See generally Robyn M. Dawes et al., Clinical Versus Actuarial Judgment, 243 SCIENCE 1668 (1989). But highly accurate prediction of future violent conduct by any method, even among high risk groups, eludes us. Optimists about prediction sometimes overstate the accuracy rates found in more recent research, neglect the methodological flaws or cautions of newer work, lump all "violence" together, or fail to discuss adequately the trade-off between sensitivity and specificity. See, e.g., Alexander D. Brooks, The Constitutionality and Morality of Civilly Committing Violent Sexual Predators, 15 U. PUGET SOUND L. REV. 709, 735-54 (properly criticizing flaws in the older research, but making some optimist's errors).
59. See Charles W. Lidz et al., The Accuracy of Predictions of Violence to Others, 269 JAMA 1007, 1009 (1993) (reporting: (1) 60% sensitivity (true positive proportion) for clinicians' predictions of violence, compared to an expected chance accuracy rate of 50%; and (2) no increase in accuracy with patients the clinicians judged to be especially dangerous). This study, which is one of the most methodologically sophisticated and persuasive to appear in the literature, demonstrates that clinicians are "relatively inaccurate predictors of violence." Id. at 1010. Clinicians' predictions of violence exceeded chance for male patients, but for female patients their predictions were not significantly better than chance. Id. Moreover, the sample in this study was particularly violent suggesting that the sensitivity of violence predictions would decrease in a random sample of emergency room patients. Id. at 1009. Finally, although the sample was probably unusually violent, only 14.4% of the violent occurrences involved serious violence. Id. The study also concluded that it was unclear whether using clinical judgment to predict violence was more accurate than simply using a history of violence Id. at 1010. For a second analysis of clinical predictive accuracy, see Mossman supra note 14, at 788-90 (reanalyzing 58 data sets from 44 disparate published studies which demonstrate
60. See Charles W. Lidz et al., The Accuracy of Predictions of Violence to Others, 269 JAMA 1007, 1009 (1993) (reporting: (1) 60% sensitivity (true positive proportion) for clinicians' predictions of violence, compared to an expected chance accuracy rate of 50%; and (2) no increase in accuracy with patients the clinicians judged to be especially dangerous). This study, which is one of the most methodologically sophisticated and persuasive to appear in the literature, demonstrates that clinicians are "relatively inaccurate predictors of violence." Id. at 1010. Clinicians' predictions of violence exceeded chance for male patients, but for female patients their predictions were not significantly better than chance. Id. Moreover, the sample in this study was particularly violent suggesting that the sensitivity of violence predictions would decrease in a random sample of emergency room patients. Id. at 1009. Finally, although the sample was probably unusually violent, only 14.4% of the violent occurrences involved serious violence. Id. The study also concluded that it was unclear whether using clinical judgment to predict violence was more accurate than simply using a history of violence Id. at 1010. For a second analysis of clinical predictive accuracy, see Mossman supra note 14, at 788-90 (reanalyzing 58 data sets from 44 disparate published studies which demonstrate that: (1) clinicians' predictions generally exceed chance, but that errors are inevitable; (2) past behavior is the best predictor of future violent conduct; and (3) nonclinicians furnished with data about past behavior may outperform clinicians who rely solely on data gained from a clinical interview); see also David B. Villeneuve & Vernon L. Quinsey, Predictors of General and Violent Recidivism Among Mentally Disordered Inmates, 22 CRIM. JUST. & BEHAV. 397, 406-09 (1995) (investigating a sample of high-risk, recidivist, mentally disorderd offenders for an average 92-month follow-up period after their release from a treatment facility, and finding: (1) 78.3% of the sample were rearrested, of which 49.6% were arrested for a violent offense; (2) an actuarial prediction instrument was a modest predictor of general recidivism, yielding a 31.7% relative improvement over chance; (3) only 16.7% of the initial rearrests were for a serious violent offense, a figure so low that accurate identification of this latter group using generally more accurate actuarial methods was precluded; and (4) the presence of psychosis decreased the rearrest risk for a violent offense). But see Robert Menzies & Christopher D. Webster, Construction and Validation of Risk Assessments in a Six-Year Follow-Up of Forensic Patients: A Tridimensional Analysis, 63 J. CONSULTING & CLINICAL PSYCHOL. 766, 775-76 (1995) (reporting on the basis of a careful, sophisticated study of a high-risk sample that: (1) neither actuarial nor clinical prediction is highly accurate and that little progress on the risk assessment of people with mental disorder has been made; (2) with few exceptions, the direct predictions of dangerousness "were almost universally invalid[;]" (3) even those risk assessors who were best at the task provided predictions that were "neither powerful nor pragmatically of any value[;]" and, (4) a handful of sociodemographic variables were far better predictors than clinical assessments and actuarial instruments). This study "demonstrates once more that the 'holy grail' of violence prediction is still far off." Id. at 775. In general, actuarial prediction is superior to clinical prediction. See generally Robyn M. Dawes et al., Clinical Versus Actuarial Judgment, 243 SCIENCE 1668 (1989). But highly accurate prediction of future violent conduct by any method, even among high risk groups, eludes us. Optimists about prediction sometimes overstate the accuracy rates found in more recent research, neglect the methodological flaws or cautions of newer work, lump all "violence" together, or fail to discuss adequately the trade-off between sensitivity and specificity. See, e.g., Alexander D. Brooks, The Constitutionality and Morality of Civilly Committing Violent Sexual Predators, 15 U. PUGET SOUND L. REV. 709, 735-54 (properly criticizing flaws in the older research, but making some optimist's errors).
61. Cf. C.M.V. Clarkson, Law Commission Report on Offenses Against the Person and General Principles: (1) Violence and the Law Commission, 1994 CRIM. L. REV. 324, 327-29, 331-32 (discussing the British Law Commission's assessment of the seriousness of various forms of assaultive behavior).
62. See Lidz et al., supra note 39, at 1009.
63. CHARLES A. KIESLER & AMY E. SIBULKIN, MENTAL HOSPITALIZATION: MYTHS AND FACTS ABOUT A NATIONAL CRISIS 152-80 (1987) (concluding, based on empirical evidence, that "alternative care is more effective and less costly than mental hospitalization"). The claim is not that everyone can be treated more efficiently outside the hospital. Rather, it is that inpatient hospitalization is necessary for too few people with mental disorders to justify the expense and intrusiveness of maintaining an extensive system of inpatient treatment. See also Burton A. Weisbrod, A Guide to Benefit-Cost Analysis, as Seen Through a Controlled Experiment in Treating the Mentally Ill, 7 J. HEALTH POL. POL'Y & L. 808, 835-36 (1983) (analyzing data demonstrating that alternative community care is more effective at the same cost as traditional hospitalization, especially for people with schizophrenia and nonschizophrenic psychoses, but that community care may not be as cost-benefit justified in monetary terms for people with personality disorders).
64. CHARLES A. KIESLER & AMY E. SIBULKIN, MENTAL HOSPITALIZATION: MYTHS AND FACTS ABOUT A NATIONAL CRISIS 152-80 (1987) (concluding, based on empirical evidence, that "alternative care is more effective and less costly than mental hospitalization"). The claim is not that everyone can be treated more efficiently outside the hospital. Rather, it is that inpatient hospitalization is necessary for too few people with mental disorders to justify the expense and intrusiveness of maintaining an extensive system of inpatient treatment. See also Burton A. Weisbrod, A Guide to Benefit-Cost Analysis, as Seen Through a Controlled Experiment in Treating the Mentally Ill, 7 J. HEALTH POL. POL'Y & L. 808, 835-36 (1983) (analyzing data demonstrating that alternative community care is more effective at the same cost as traditional hospitalization, especially for people with schizophrenia and nonschizophrenic psychoses, but that community care may not be as cost-benefit justified in monetary terms for people with personality disorders).
65. See Mossman, supra note 14, at 789-90 (reporting that past behavior is the best predictor of future violence).
66. In the case of seriously crazy people, whose irrational practical reasoning leads to the intent to do harm, ameliorating the crazy thinking through proper medication should in fact reduce the risk of harmdoing. I know of no study that demonstrates this, but the conception of human action as rationalized by practical reason seems to entail it. For people who are not grossly irrational, however, the relation between their disorders and future violence is often tenuous at best. Treatment that would permit early release of such people is likely to be expensive, time consuming, and ultimately ineffective.
67. See Peter Conrad, Medicalization and Social Control, 18 ANN. REV. SOC. 209, 223-24 (Judith Blake & John Hagan eds., 1992) (describing one consequence of medicalization as the individualization of a collective social problem).
68. See HENRY J. STEADMAN ET AL., BEFORE AND AFTER HINCKLEY: EVALUATING INSANITY DEFENSE REFORM 27-28 (1993) (finding in a study of four states that the insanity plea was raised in just under 1% of all felony cases, the acquittal rates varied from 0.12 to 0.41 per 100 felony indictments, and the success of the plea as a percentage of pleas raised ranged from 7% to 46%, with an overall success rate of 23%).
69. Compare a harmdoer excused because she was acting under duress. In virtually all cases, there is no reason to believe that the coercion applied in the current situation is a general feature of her life. Continued restraint of the harmdoer to prevent future occurrences of coercion is unnecessary. Automatism might seem to be another doctrine that excuses defendants for the same reason that they are dangerous. Properly understood, however, many cases of automatism negate the act requirement of the prima facie case and thus are not excuses. In some jurisdictions, however, the defendant must plead legal insanity if automatism is produced by mental disease or defect. I believe that dissociative states, which often support automatism claims, should be treated as a potential excusing condition, rather than as denial of the act. Stephen J. Morse, Culpability and Control, 142 U. PA. L. REV. 1587, 1641-52 (1994) [hereinafter Morse, Culpability and Control]. Current law is often to the contrary. See MODEL PENAL CODE § 2.01(2)(b)-(c) (1985) (excluding a bodily movement during unconsciousness or sleep and conduct under hypnosis or posthypnotic suggestion from the definition of a voluntary act).
70. Compare a harmdoer excused because she was acting under duress. In virtually all cases, there is no reason to believe that the coercion applied in the current situation is a general feature of her life. Continued restraint of the harmdoer to prevent future occurrences of coercion is unnecessary. Automatism might seem to be another doctrine that excuses defendants for the same reason that they are dangerous. Properly understood, however, many cases of automatism negate the act requirement of the prima facie case and thus are not excuses. In some jurisdictions, however, the defendant must plead legal insanity if automatism is produced by mental disease or defect. I believe that dissociative states, which often support automatism claims, should be treated as a potential excusing condition, rather than as denial of the act. Stephen J. Morse, Culpability and Control, 142 U. PA. L. REV. 1587, 1641-52 (1994) [hereinafter Morse, Culpability and Control]. Current law is often to the contrary. See MODEL PENAL CODE § 2.01(2)(b)-(c) (1985) (excluding a bodily movement during unconsciousness or sleep and conduct under hypnosis or posthypnotic suggestion from the definition of a voluntary act).
71. 18 U.S.C.A. § 17(a) (West Supp. 1995) ("It is an affirmative defense . . . that, at the time of the commission of the acts constituting the offense, the defendant, as a result of a severe mental disease or defect, was unable to appreciate the nature and quality or the wrongfulness of his acts.").
72. MODEL PENAL CODE § 4.01(1) (1985) ("A person is not responsible for criminal conduct if at the time of such conduct as a result of mental disease or defect he lacks substantial capacity . . . to appreciate the criminality [wrongfulness] of his conduct
73. The most influential is, of course, the Model Penal Code's second prong. Id. ("A person is not responsible for criminal conduct if at the time of such conduct as a result of mental disease or defect he lacks substantial capacity . . . to conform his conduct to the requirements of law.").
74. United States v. Hinckley, 525 F. Supp. 1342 (D.D.C. 1981), clarified, 529 F. Supp. 520 (D.D.C), aff'd, 672 F.2d 115 (D.C. Cir. 1982).
75. See, e.g., Robinson, supra note 1, at 703 n.35 ("[T]o exclude the possibility of exculpation on [the ground of control dysfunction] is to create the possibility for convicting blameless offenders."). Professor Robinson also suggests that narrowing the insanity defense to exclude a control test may have exceeded the community's expectations of justice. He bases this view on very interesting empirical research that he and psychologist John Darley performed. They presented subjects with vignettes about criminal conduct to test whether the criminal law was congruent with the community's expectations of justice. The results from the part of the study that used a vignette of an allegedly mentally disordered agent indicated that the subjects thought a control test was appropriate. PAUL H. ROBINSON & JOHN M. DARLEY, JUSTICE, LIABILITY, AND BLAME: COMMUNITY VIEWS AND THE CRIMINAL LAW 128-39, 262-65 (1995). I believe that the general "vignette" methodology has serious problems and fear that Robinson and Darley's specific vignettes perhaps "stacked the deck" unfairly in favor of a control test. Even if their result is valid, however, its normative implication is unclear. The question for a desert theorist is whether some otherwise rational defendants are in fact unable to control their conduct. Many commentators deny this on conceptual and empirical grounds. If they are right - admittedly a big "if" - the following question arises: Should the public's unjustified belief justify the adoption of an unjustifiable defense? Robinson and Darley are surely correct that the legitimacy of the criminal law is undermined if it is too inconsistent with the public's sense of justice. Nevertheless, it is doubtful that the law should promulgate a rationally indefensible rule simply because the public demands it: The law must sometimes lead as well as follow. Finally, one wonders how much the public cares about a control test for legal insanity, compared to other possible inconsistencies. For example, suppose the law ignored the strong public sentiment that results matter and, following the lead of commentators such as Sanford Kadish, did not distinguish between attempted crimes and completed crimes on the theory that results do not affect the agent's blameworthiness or dangerousness. See Sanford H. Kadish, Foreword to The Criminal Law and the Luck of the Draw, 84 J. CRIM. L. & CRIMINOLOGY 679 (1994) (arguing that differential punishment for attempted crimes and completed crimes cannot be justified in terms of the crime preventive purposes of the criminal law or in terms of "any convincing principle of justice"). I conjecture that the public would care far more about this issue.
76. See, e.g., Robinson, supra note 1, at 703 n.35 ("[T]o exclude the possibility of exculpation on [the ground of control dysfunction] is to create the possibility for convicting blameless offenders."). Professor Robinson also suggests that narrowing the insanity defense to exclude a control test may have exceeded the community's expectations of justice. He bases this view on very interesting empirical research that he and psychologist John Darley performed. They presented subjects with vignettes about criminal conduct to test whether the criminal law was congruent with the community's expectations of justice. The results from the part of the study that used a vignette of an allegedly mentally disordered agent indicated that the subjects thought a control test was appropriate. PAUL H. ROBINSON & JOHN M. DARLEY, JUSTICE, LIABILITY, AND BLAME: COMMUNITY VIEWS AND THE CRIMINAL LAW 128-39, 262-65 (1995). I believe that the general "vignette" methodology has serious problems and fear that Robinson and Darley's specific vignettes perhaps "stacked the deck" unfairly in favor of a control test. Even if their result is valid, however, its normative implication is unclear. The question for a desert theorist is whether some otherwise rational defendants are in fact unable to control their conduct. Many commentators deny this on conceptual and empirical grounds. If they are right - admittedly a big "if" - the following question arises: Should the public's unjustified belief justify the adoption of an unjustifiable defense? Robinson and Darley are surely correct that the legitimacy of the criminal law is undermined if it is too inconsistent with the public's sense of justice. Nevertheless, it is doubtful that the law should promulgate a rationally indefensible rule simply because the public demands it: The law must sometimes lead as well as follow. Finally, one wonders how much the public cares about a control test for legal insanity, compared to other possible inconsistencies. For example, suppose the law ignored the strong public sentiment that results matter and, following the lead of commentators such as Sanford Kadish, did not distinguish between attempted crimes and completed crimes on the theory that results do not affect the agent's blameworthiness or dangerousness. See Sanford H. Kadish, Foreword to The Criminal Law and the Luck of the Draw, 84 J. CRIM. L. & CRIMINOLOGY 679 (1994) (arguing that differential punishment for attempted crimes and completed crimes cannot be justified in terms of the crime preventive purposes of the criminal law or in terms of "any convincing principle of justice"). I conjecture that the public would care far more about this issue.
77. See, e.g., Robinson, supra note 1, at 703 n.35 ("[T]o exclude the possibility of exculpation on [the ground of control dysfunction] is to create the possibility for convicting blameless offenders."). Professor Robinson also suggests that narrowing the insanity defense to exclude a control test may have exceeded the community's expectations of justice. He bases this view on very interesting empirical research that he and psychologist John Darley performed. They presented subjects with vignettes about criminal conduct to test whether the criminal law was congruent with the community's expectations of justice. The results from the part of the study that used a vignette of an allegedly mentally disordered agent indicated that the subjects thought a control test was appropriate. PAUL H. ROBINSON & JOHN M. DARLEY, JUSTICE, LIABILITY, AND BLAME: COMMUNITY VIEWS AND THE CRIMINAL LAW 128-39, 262-65 (1995). I believe that the general "vignette" methodology has serious problems and fear that Robinson and Darley's specific vignettes perhaps "stacked the deck" unfairly in favor of a control test. Even if their result is valid, however, its normative implication is unclear. The question for a desert theorist is whether some otherwise rational defendants are in fact unable to control their conduct. Many commentators deny this on conceptual and empirical grounds. If they are right - admittedly a big "if" - the following question arises: Should the public's unjustified belief justify the adoption of an unjustifiable defense? Robinson and Darley are surely correct that the legitimacy of the criminal law is undermined if it is too inconsistent with the public's sense of justice. Nevertheless, it is doubtful that the law should promulgate a rationally indefensible rule simply because the public demands it: The law must sometimes lead as well as follow. Finally, one wonders how much the public cares about a control test for legal insanity, compared to other possible inconsistencies. For example, suppose the law ignored the strong public sentiment that results matter and, following the lead of commentators such as Sanford Kadish, did not distinguish between attempted crimes and completed crimes on the theory that results do not affect the agent's blameworthiness or dangerousness. See Sanford H. Kadish, Foreword to The Criminal Law and the Luck of the Draw, 84 J. CRIM. L. & CRIMINOLOGY 679 (1994) (arguing that differential punishment for attempted crimes and completed crimes cannot be justified in terms of the crime preventive purposes of the criminal law or in terms of "any convincing principle of justice"). I conjecture that the public would care far more about this issue.
78. See Morse, Culpability and Control, supra note 47, at 1622-34 (setting forth in full the argument that most control difficulties are better understood as rationality problems);
79. Stephen J. Morse, Excusing the Crazy: The Insanity Defense Reconsidered, 58 S. CAL. L. REV. 777, 813 (1985) [hereinafter Morse, Excusing the Crazy] (arguing that the law should treat as instances of irrationality many cases that are now conceived of as control problems). I also argue that we lack the clinical and scientific means to measure control problems, even if they are conceptually coherent See id. at 817 ("If or to what degree a person's desire or impulse to act was controllable is not determinable: there is no scientific test to judge whether an impulse was irresistible or simply not resisted."). Thus, on consequential grounds, too, a control test is undesirable.
80. Although I believe that the insanity defense should not be broadened, I also strongly oppose adoption of the "guilty but mentally ill" verdict. This verdict is unnecessary to provide treatment to prisoners fairly convicted. More important, it can operate as an unfair "compromise" in cases in which jurors believe the defendant is legally insane, but reject appropriate acquittal by reason of insanity because they fear that acquittal will lead to early release. See Morse, Excusing the Crazy, supra note 53, at 803-04 (setting forth in full the argument against the "guilty but mentally ill" verdict).
81. See Michael S. Moore, Causation and the Excuses, 73 CAL. L. REV. 1091, 1112-13 (1985) (demonstrating that if causation were itself an excuse, then under a determinist theory, all actions would be excused); Stephen J. Morse, Psychology, Determinism and Legal Responsibility, in NEBRASKA SYMPOSIUM ON MOTIVATION: THE LAW AS A BEHAVIORAL INSTRUMENT 35, 48-50 (Gary B. Melton ed., 1986) (refuting the causal determinist theory that causation itself is an excuse).
82. See Michael S. Moore, Causation and the Excuses, 73 CAL. L. REV. 1091, 1112-13 (1985) (demonstrating that if causation were itself an excuse, then under a determinist theory, all actions would be excused); Stephen J. Morse, Psychology, Determinism and Legal Responsibility, in NEBRASKA SYMPOSIUM ON MOTIVATION: THE LAW AS A BEHAVIORAL INSTRUMENT 35, 48-50 (Gary B. Melton ed., 1986) (refuting the causal determinist theory that causation itself is an excuse).
83. I discuss these issues in greater detail elsewhere. E.g., Stephen J. Morse, The "New Syndrome Excuse Syndrome", CRIM. JUST. ETHICS, Winter/Spring 1995, at 3.
84. See LENORE E. WALKER, THE BATTERED WOMAN 31-35 (1979) (discussing common characteristics of battered women). For critiques of the battered woman syndrome defense, see Anne M. Coughlin, Excusing Women, 82 CAL. L. REV. 1 (1994) (arguing that by denying that women are capable of conforming to the law, the battered woman syndrome defense also denies that women have the same capacity for self-governance that is attributed to men); Robert F. Schopp et al., Battered Woman Syndrome, Expert Testimony, and the Distinction Between Justification and Excuse, 1994 U. ILL. L. REV. 45, 58-59, 63-64 (doubting the validity of the syndrome and whether it is an accurate description of the victims of battering relationships and suggesting that learned helplessness is not consistent with the behavior of battered women who do kill); David L. Faigman, Note, The Battered Woman Syndrome and Self-Defense: A Legal and Empirical Dissent, 72 VA. L. REV. 619 (1986) (examining methodological flaws of battered woman syndrome research and analyzing weaknesses in Walker's "cycle theory of violence").
85. See LENORE E. WALKER, THE BATTERED WOMAN 31-35 (1979) (discussing common characteristics of battered women). For critiques of the battered woman syndrome defense, see Anne M. Coughlin, Excusing Women, 82 CAL. L. REV. 1 (1994) (arguing that by denying that women are capable of conforming to the law, the battered woman syndrome defense also denies that women have the same capacity for self-governance that is attributed to men); Robert F. Schopp et al., Battered Woman Syndrome, Expert Testimony, and the Distinction Between Justification and Excuse, 1994 U. ILL. L. REV. 45, 58-59, 63-64 (doubting the validity of the syndrome and whether it is an accurate description of the victims of battering relationships and suggesting that learned helplessness is not consistent with the behavior of battered women who do kill); David L. Faigman, Note, The Battered Woman Syndrome and Self-Defense: A Legal and Empirical Dissent, 72 VA. L. REV. 619 (1986) (examining methodological flaws of battered woman syndrome research and analyzing weaknesses in Walker's "cycle theory of violence").
86. See LENORE E. WALKER, THE BATTERED WOMAN 31-35 (1979) (discussing common characteristics of battered women). For critiques of the battered woman syndrome defense, see Anne M. Coughlin, Excusing Women, 82 CAL. L. REV. 1 (1994) (arguing that by denying that women are capable of conforming to the law, the battered woman syndrome defense also denies that women have the same capacity for self-governance that is attributed to men); Robert F. Schopp et al., Battered Woman Syndrome, Expert Testimony, and the Distinction Between Justification and Excuse, 1994 U. ILL. L. REV. 45, 58-59, 63-64 (doubting the validity of the syndrome and whether it is an accurate description of the victims of battering relationships and suggesting that learned helplessness is not consistent with the behavior of battered women who do kill); David L. Faigman, Note, The Battered Woman Syndrome and Self-Defense: A Legal and Empirical Dissent, 72 VA. L. REV. 619 (1986) (examining methodological flaws of battered woman syndrome research and analyzing weaknesses in Walker's "cycle theory of violence").
87. See LENORE E. WALKER, THE BATTERED WOMAN 31-35 (1979) (discussing common characteristics of battered women). For critiques of the battered woman syndrome defense, see Anne M. Coughlin, Excusing Women, 82 CAL. L. REV. 1 (1994) (arguing that by denying that women are capable of conforming to the law, the battered woman syndrome defense also denies that women have the same capacity for self-governance that is attributed to men); Robert F. Schopp et al., Battered Woman Syndrome, Expert Testimony, and the Distinction Between Justification and Excuse, 1994 U. ILL. L. REV. 45, 58-59, 63-64 (doubting the validity of the syndrome and whether it is an accurate description of the victims of battering relationships and suggesting that learned helplessness is not consistent with the behavior of battered women who do kill); David L. Faigman, Note, The Battered Woman Syndrome and Self-Defense: A Legal and Empirical Dissent, 72 VA. L. REV. 619 (1986) (examining methodological flaws of battered woman syndrome research and analyzing weaknesses in Walker's "cycle theory of violence").
88. Holly Maguigan, Battered Women and Self-Defense: Myths and Misconceptions in Current Reform Proposals, 140 U. PA. L. REV. 379, 391-97 (1991) (reporting in her analysis of appellate decisions that 75% of situations involved confrontations).
89. Some sufferers may be sufficiently irrational to warrant an excuse, but such cases are rare.
90. See Coughlin, supra note 57, at 70-87 (arguing that routine excusing of women who kill their batterers undesirably treats women as less than full moral agents).
91. See Morse, Culpability and Control, supra note 47, at 1652-54 (exploring the issue of a history of deprivation, emotional or otherwise, as an independent ground for a control excuse). Among other reasons, most defendants with a "rotten social background" were rational when they offended and were not coerced to commit their crimes. And causation is not an excuse. See supra note 55.
92. See generally Estelle B. Freedman, "Uncontrolled Desires": The Response to the Sexual Psychopath, 1920-1960, 74 J. AM. HIST. 83 (1987) (presenting an overview of the response of the public, politicians, and psychiatrists to sexual psychopaths during the middle of this century); John F. Galliher & Cheryl Tyree, Edwin Sutherland's Research on the Origins of Sexual Psychopath Laws: An Early Case Study of the Medicalization of Deviance, 33 SOC. PROBS. 100 (1985) (criticizing an influential account of the development of sexual psychopath laws).
93. See generally Estelle B. Freedman, "Uncontrolled Desires": The Response to the Sexual Psychopath, 1920-1960, 74 J. AM. HIST. 83 (1987) (presenting an overview of the response of the public, politicians, and psychiatrists to sexual psychopaths during the middle of this century); John F. Galliher & Cheryl Tyree, Edwin Sutherland's Research on the Origins of Sexual Psychopath Laws: An Early Case Study of the Medicalization of Deviance, 33 SOC. PROBS. 100 (1985) (criticizing an influential account of the development of sexual psychopath laws).
94. WASH. REV. CODE ANN. §§ 71.09.010-.120 (West 1992 & Supp. 1995). In 1993, the Supreme Court of Washington upheld the constitutionality of this statute. In re Young, 857 P.2d 989, 1018 (Wash. 1993). However, in August 1995, a federal district court held that the statute violated the substantive due process component of the Fourteenth Amendment, the Ex Post Facto Clause of Article I, § 10, and the Double Jeopardy Clause of the Fifth Amendment. Young v. Weston, 898 F. Supp. 744, 754 (W.D. Wash. 1995). For interesting accounts of the passage of Washington's Sexually Violent Predator statute, see John Q. La Fond, Washington's Sexually Violent Predator Law: A Deliberate Misuse of the Therapeutic State for Social Control, 15 U. PUGET SOUND L. REV. 655, 670-84 (1992); Stuart Scheingold et al., The Politics of Sexual Psychopathy: Washington State's Sexual Predator Legislation, 15 U. PUGET SOUND L. REV. 809, 816-20 (1992).
95. WASH. REV. CODE ANN. §§ 71.09.010-.120 (West 1992 & Supp. 1995). In 1993, the Supreme Court of Washington upheld the constitutionality of this statute. In re Young, 857 P.2d 989, 1018 (Wash. 1993). However, in August 1995, a federal district court held that the statute violated the substantive due process component of the Fourteenth Amendment, the Ex Post Facto Clause of Article I, § 10, and the Double Jeopardy Clause of the Fifth Amendment. Young v. Weston, 898 F. Supp. 744, 754 (W.D. Wash. 1995). For interesting accounts of the passage of Washington's Sexually Violent Predator statute, see John Q. La Fond, Washington's Sexually Violent Predator Law: A Deliberate Misuse of the Therapeutic State for Social Control, 15 U. PUGET SOUND L. REV. 655, 670-84 (1992); Stuart Scheingold et al., The Politics of Sexual Psychopathy: Washington State's Sexual Predator Legislation, 15 U. PUGET SOUND L. REV. 809, 816-20 (1992).
96. Brooks, supra note 39, at 718-19 (discussing the civil/criminal dichotomy and noting that the Supreme Court has approved preventive detention under certain circumstances);
97. Robinson, supra note 1, at 715 (supporting Washington's Sexually Violent Predator statute, with its system of periodic review).
98. But see John Q. La Fond, Washington's Sexually Violent Predators Statute: Law or Lottery? A Response to Professor Brooks, 15 U. PUGET SOUND L. REV. 755 (1992) (criticizing Brooks's general position, and specifically denying Brooks's claim that it is better to preventively detain two individuals, one of whom we are confident will commit another crime of sexual violence in the future, rather than releasing them both). Because the substantive criteria for most forms of commitment are so vague and broad, the procedural protections have less force than they otherwise might. See supra text accompanying notes 71-74.
99. It is possible, of course, that any abnormality arose after the conviction, but this is unlikely. I shall make the simplifying assumption that the offender's mental state is the same at the time of the triggering conviction and of the commitment.
100. This is also true of traditional MDSO commitments that were triggered by a criminal conviction and imposed in lieu of sentence. Of course, a mere sex offense charge initiated some traditional MDSO commitments. There is no evidence however, that a judge or jury would have found these MDSOs legally insane if they had been tried.
101. Janet Warren et al., The Sexually Sadistic Serial Killer 14, 19 (1995) (unpublished manuscript, on file with author) (reviewing the literature and analyzing twenty cases of sexually sadistic serial killers selected from files obtained by the National Center of the Analysis of Violent Crime of the FBI).
102. Although the "sexual predator" has previously been convicted of actual harmdoing, his current commitment, premised on his propensity to commit future harms, is unrelated to his current blameworthiness. It is this latter feature that distinguishes this form of commitment as civil and differentiates it from criminal confinement based on a preventive incapacitation rationale. Incapacitative criminal confinement makes current desert a necessary precondition for punishment. See FRANKLIN E. ZIMRING & GORDON HAWKINS, INCAPACITATION: PENAL CONFINEMENT AND THE RESTRAINT OF CRIME 73 (1995) (arguing for the importance of maintaining desert as a necessary, albeit not sufficient, justification for punishment); see also Norval Morris & Marc Miller Predictions of Dangerousness, 6 CRIME & JUST.: AN ANN. REV. OF RES. 1, 35 (Michael Tonry & Norval Morris eds., 1985) (positing as a limitation on the proper use of predictions of dangerousness in the criminal law that "[p]unishment should not be imposed, nor the term of punishment extended, by virtue of a prediction of dangerousness, beyond that which would be justified as a deserved punishment independently of that prediction").
103. Although the "sexual predator" has previously been convicted of actual harmdoing, his current commitment, premised on his propensity to commit future harms, is unrelated to his current blameworthiness. It is this latter feature that distinguishes this form of commitment as civil and differentiates it from criminal confinement based on a preventive incapacitation rationale. Incapacitative criminal confinement makes current desert a necessary precondition for punishment. See FRANKLIN E. ZIMRING & GORDON HAWKINS, INCAPACITATION: PENAL CONFINEMENT AND THE RESTRAINT OF CRIME 73 (1995) (arguing for the importance of maintaining desert as a necessary, albeit not sufficient, justification for punishment); see also Norval Morris & Marc Miller Predictions of Dangerousness, 6 CRIME & JUST.: AN ANN. REV. OF RES. 1, 35 (Michael Tonry & Norval Morris eds., 1985) (positing as a limitation on the proper use of predictions of dangerousness in the criminal law that "[p]unishment should not be imposed, nor the term of punishment extended, by virtue of a prediction of dangerousness, beyond that which would be justified as a deserved punishment independently of that prediction").
104. WASH. REV. CODE ANN. § 71.09.020(1) (West Supp. 1995).
105. Id. § 71.09.020(2).
106. DSM-IV, supra note 34, at 629-73.
107. Brooks, supra note 39, at 730 (friend); La Fond, supra note 64, at 762-63 (foe).
108. Brooks, supra note 39, at 730 (friend); La Fond, supra note 64, at 762-63 (foe).
109. Alexander Brooks cogently makes this point. Brooks, supra note 39, at 730.
110. In addition, the definition implies that some criminal sexual acts might not be a "menace," but if not, why are they criminalized?
111. See supra note 55 and accompanying text.
112. See ROBERT NOZICK, THE NATURE OF RATIONALITY 139-40 (1993) ("At present, we have no adequate theory of the substantive rationality of goals or desires . . . .").
113. See supra text accompanying notes 65-67; cf. Ronet Bachman et al., The Rationality of Sexual Offending: Testing a Deterrence/Rational Choice Conception of Sexual Assault 26 LAW & SOC'Y REV. 343 (1992) (finding in a study of conditions that would affect the likelihood of committing sexual assault that both perceived risk of formal sanction and moral evaluation of the act had a significant restraining effect).
114. See supra text accompanying notes 65-67; cf. Ronet Bachman et al., The Rationality of Sexual Offending: Testing a Deterrence/Rational Choice Conception of Sexual Assault 26 LAW & SOC'Y REV. 343 (1992) (finding in a study of conditions that would affect the likelihood of committing sexual assault that both perceived risk of formal sanction and moral evaluation of the act had a significant restraining effect).
115. See MORSE'S MENTAL DIAGNOSTIC MANUAL § 1.3 (on file in the author's head).
116. See DSM-IV, supra note 34, at 530.
117. See, e.g., Brooks, supra note 39, at 730-32 (arguing that the Washington legislature chose to use the term "mental abnormality" in the Sexually Violent Predator statute to bring within the scope of that statute individuals with impaired volitional controls). Lack of control appears to be the dominant operative excusing condition that Brooks employs, but the empirical foundation is almost nonexistent and the theory is not conceptually supported.
118. Lita Furby et al., Sex Offender Recidivism: A Review, 105 PSYCHOL. BULL. 3, 4, 27 (1989).
119. Gordon C.N. Hall, Sexual Offender Recidivism Revisited: A Meta-Analysis of Recent Treatment Studies, 63 J. CONSULTING & CLINICAL PSYCHOL. 802, 806 (1995). Nonsexual recidivism was not reported.
120. See Robert A. Prentky et al., Predictive Validity of Lifestyle Impulsivity for Rapists, 22 CRIM. JUST. & BEHAV. 106 (1995) (reporting that high impulsivity had high predictive validity for sexual reoffending among previously convicted rapists, as well as for reoffending in general, as this group committed an even greater number of nonsexual offenses).
121. Furby et al., supra note 81, at 27 (reviewing the few studies that do organize their samples by category type and calling for more sophistication in identifying offender categories in recidivism research).
122. See Steven D. Gottfredson & Don M. Gottfredson, Behavioral Prediction and the Problem of Incapacitation, 32 CRIMINOLOGY 441, 468 (1994) (reporting that in their study of recidivism, subsequent arrest offenses were more likely to be trivial than serious by more than a three-to-one ratio).
123. See Scheingold et al., supra note 63, at 812 (reviewing recidivism statistics).
124. See Gordon C.N. Hall, Prediction of Sexual Aggression, 10 CLINICAL PSYCHOL. REV. 229, 239 (1990) (noting that some predictive techniques hold promise, but prediction of sexual aggression is still in an early stage of development, and concluding that (1) actuarial methods are superior to clinical prediction, and (2) past sexual aggression against adults is the best predictor of future sexual aggression against adults); Vernon L. Quinsey et al., Predicting Sexual Offenses, in ASSESSING DANGEROUSNESS: VIOLENCE BY SEXUAL OFFENDERS, BATTERERS, AND CHILD ABUSERS 114, 131-32 (Jacquelyn C. Campbell ed., 1995) (agreeing with the overall assessment of the Furby review, supra note 81, but claiming on the basis of one study that newer methods may enhance the accuracy of predictions of sexual recidivism in some groups).
125. See Gordon C.N. Hall, Prediction of Sexual Aggression, 10 CLINICAL PSYCHOL. REV. 229, 239 (1990) (noting that some predictive techniques hold promise, but prediction of sexual aggression is still in an early stage of development, and concluding that (1) actuarial methods are superior to clinical prediction, and (2) past sexual aggression against adults is the best predictor of future sexual aggression against adults); Vernon L. Quinsey et al., Predicting Sexual Offenses, in ASSESSING DANGEROUSNESS: VIOLENCE BY SEXUAL OFFENDERS, BATTERERS, AND CHILD ABUSERS 114, 131-32 (Jacquelyn C. Campbell ed., 1995) (agreeing with the overall assessment of the Furby review, supra note 81, but claiming on the basis of one study that newer methods may enhance the accuracy of predictions of sexual recidivism in some groups).
126. See MICHAEL R. GOTTFREDSON & TRAVIS HIRSCHI, A GENERAL THEORY OF CRIME 91-94 (1990) (reviewing "overwhelming" evidence of offender versatility rather than specialization in crimes committed).
127. WASH. REV. CODE ANN. § 71.09.020 (West Supp. 1995).
128. Id. § 9.94A.030(32).
129. See Furby et al., supra note 81, at 24-25 (reporting in a review of sex offender recidivism research that eight of nine studies of untreated offenders show relatively low recidivism rates, while two-thirds of treated offender studies show relatively high recidivism rates, but noting that these results may in part derive from the likelihood that treated offenders were monitored more closely during these studies than untreated offenders, increasing the probability of being caught when reoffending).
130. Hall, supra note 82, at 805-08.
131. The sexual predator unit in Washington can preventively detain no more than 36 men. La Fond, supra note 63, at 701. Given the recidivism rates of sex offenders, very little public safety is gained by this tiny program. To expand it considerably would be extremely expensive and would risk increasing the number of false positive commitments.
132. The average prison time per violent crime tripled between 1975 and 1989. MICHAEL TONRY, MALIGN NEGLECT: RACE, CRIME, AND PUNISHMENT IN AMERICA 17 (1995). Indeed, lengthier incarceration of greater numbers of offenders has resulted in almost a tripling of the prison population in the last fifteen years. Id. at 40. For a history of the recent dominance of the now regnant incapacitative rationale for confinement, see ZIMRING & HAWKINS, supra note 68, at 3-17 (arguing that incapacitative theory is not empirically supported and became popular through a passive process of elimination rather than after active academic and political debate).
133. The average prison time per violent crime tripled between 1975 and 1989. MICHAEL TONRY, MALIGN NEGLECT: RACE, CRIME, AND PUNISHMENT IN AMERICA 17 (1995). Indeed, lengthier incarceration of greater numbers of offenders has resulted in almost a tripling of the prison population in the last fifteen years. Id. at 40. For a history of the recent dominance of the now regnant incapacitative rationale for confinement, see ZIMRING & HAWKINS, supra note 68, at 3-17 (arguing that incapacitative theory is not empirically supported and became popular through a passive process of elimination rather than after active academic and political debate).
134. There is no question about the current constitutionality of almost any term of years a legislature sees fit to impose for virtually any offense. Even the most draconian terms of years appear to be within the legislative prerogative. See Harmelin v. Michigan, 501 U.S. 957, 994-96 (1991) (holding that a mandatory life sentence without parole for the possession of 672 grams of cocaine may be cruel, but it is not unusual under the Eighth Amendment because legislatures have adopted such severe, mandatory penalties in various forms throughout our nation's history).
135. What counts as a serious offense is debatable. Paul Robinson and I, among many others believe that many offenses are punishable by terms of years that are disproportionate because the notion of desert that supports them is too harsh or because their length reflects a pure preventive detention component that is unrelated to desert for the current offense. See Robinson, supra note 1, at 714-16 (arguing that a purge of dangerousness considerations from criminal sentencing would significantly enhance the moral credibility of that system);
136. see also TONRY, supra note 94, at 19-24, 196-201 (noting the disproportionality of sentencing in the United States, arguing that such harsh sentencing has had little effect on crime rates, and advocating a reversal in the trend of increases in penalties).
137. See, e.g., David Dolinko, Three Mistakes of Retributivism, 9 UCLA L. REV. 1623, 1636-42 (1992) (arguing that retributivism cannot proscribe what punishment any particular crime deserves, what the penalty for the least serious offense in a ranked list of offenses should be, and by how much to increase the penalty for each successive crime on such a list).
138. See BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, REPORT TO THE NATION ON CRIME AND JUSTICE 16-17 (2d ed. 1988) (reporting results of the National Survey of Crime Severity, showing that a stratified national sample of the U.S. population demonstrates substantial nationwide agreement about the severity rankings of various crimes). But see David A. Parton et al., Measuring Crime Seriousness: Lessons from the National Survey of Crime Severity, 31 BRIT. J. CRIMINOLOGY 72 (1991) (discussing theoretical and methodological limitations of the National Survey of Crime Severity).
139. See BUREAU OF JUSTICE STATISTICS, U.S. DEP'T OF JUSTICE, REPORT TO THE NATION ON CRIME AND JUSTICE 16-17 (2d ed. 1988) (reporting results of the National Survey of Crime Severity, showing that a stratified national sample of the U.S. population demonstrates substantial nationwide agreement about the severity rankings of various crimes). But see David A. Parton et al., Measuring Crime Seriousness: Lessons from the National Survey of Crime Severity, 31 BRIT. J. CRIMINOLOGY 72 (1991) (discussing theoretical and methodological limitations of the National Survey of Crime Severity).
140. See, e.g., Michael Davis, How to Make the Punishment Fit the Crime, in TO MAKE THE PUNISHMENT FIT THE CRIME: ESSAYS IN THE THEORY OF CRIMINAL JUSTICE 77-83 (1992).
141. See generally Daniel S. Nagin & David P. Farrington, The Onset and Persistence of Offending, 30 CRIMINOLOGY 501 (1992) (finding an inverse relationship between the age of onset of criminal behavior and persistence of offending attributable to time-stable individual differences); Daniel S. Nagin & David P. Farrington, The Stability of Criminal Potential from Childhood to Adulthood, 30 CRIMINOLOGY 235 (1992) (concluding based on a long panel data set, that the positive association between past and future criminal behavior is attributable largely to stable, unmeasured individual differences); Daniel S. Nagin & Raymond Paternoster, On the Relationship of Past to Future Participation in Delinquency, 29 CRIMINOLOGY 163 (1991) (concluding based on a short panel data set that the principal explanation for the positive relationship between past and future delinquency is state dependence, the effect of past criminal experience reducing inhibitions against engaging in future delinquent acts); see also Christy A. Visher et al., Predicting the Recidivism of Serious Youthful Offenders Using Survival Models, 29 CRIMINOLOGY 329 (1991) (identifying variables associated with recidivism and assigning "statistically reasonable" risk functions to individuals).
142. See generally Daniel S. Nagin & David P. Farrington, The Onset and Persistence of Offending, 30 CRIMINOLOGY 501 (1992) (finding an inverse relationship between the age of onset of criminal behavior and persistence of offending attributable to time-stable individual differences); Daniel S. Nagin & David P. Farrington, The Stability of Criminal Potential from Childhood to Adulthood, 30 CRIMINOLOGY 235 (1992) (concluding based on a long panel data set, that the positive association between past and future criminal behavior is attributable largely to stable, unmeasured individual differences); Daniel S. Nagin & Raymond Paternoster, On the Relationship of Past to Future Participation in Delinquency, 29 CRIMINOLOGY 163 (1991) (concluding based on a short panel data set that the principal explanation for the positive relationship between past and future delinquency is state dependence, the effect of past criminal experience reducing inhibitions against engaging in future delinquent acts); see also Christy A. Visher et al., Predicting the Recidivism of Serious Youthful Offenders Using Survival Models, 29 CRIMINOLOGY 329 (1991) (identifying variables associated with recidivism and assigning "statistically reasonable" risk functions to individuals).
143. See generally Daniel S. Nagin & David P. Farrington, The Onset and Persistence of Offending, 30 CRIMINOLOGY 501 (1992) (finding an inverse relationship between the age of onset of criminal behavior and persistence of offending attributable to time-stable individual differences); Daniel S. Nagin & David P. Farrington, The Stability of Criminal Potential from Childhood to Adulthood, 30 CRIMINOLOGY 235 (1992) (concluding based on a long panel data set, that the positive association between past and future criminal behavior is attributable largely to stable, unmeasured individual differences); Daniel S. Nagin & Raymond Paternoster, On the Relationship of Past to Future Participation in Delinquency, 29 CRIMINOLOGY 163 (1991) (concluding based on a short panel data set that the principal explanation for the positive relationship between past and future delinquency is state dependence, the effect of past criminal experience reducing inhibitions against engaging in future delinquent acts); see also Christy A. Visher et al., Predicting the Recidivism of Serious Youthful Offenders Using Survival Models, 29 CRIMINOLOGY 329 (1991) (identifying variables associated with recidivism and assigning "statistically reasonable" risk functions to individuals).
144. See generally Daniel S. Nagin & David P. Farrington, The Onset and Persistence of Offending, 30 CRIMINOLOGY 501 (1992) (finding an inverse relationship between the age of onset of criminal behavior and persistence of offending attributable to time-stable individual differences); Daniel S. Nagin & David P. Farrington, The Stability of Criminal Potential from Childhood to Adulthood, 30 CRIMINOLOGY 235 (1992) (concluding based on a long panel data set, that the positive association between past and future criminal behavior is attributable largely to stable, unmeasured individual differences); Daniel S. Nagin & Raymond Paternoster, On the Relationship of Past to Future Participation in Delinquency, 29 CRIMINOLOGY 163 (1991) (concluding based on a short panel data set that the principal explanation for the positive relationship between past and future delinquency is state dependence, the effect of past criminal experience reducing inhibitions against engaging in future delinquent acts); see also Christy A. Visher et al., Predicting the Recidivism of Serious Youthful Offenders Using Survival Models, 29 CRIMINOLOGY 329 (1991) (identifying variables associated with recidivism and assigning "statistically reasonable" risk functions to individuals).
145. See Stanford v. Kentucky, 492 U.S. 361, 380 (1989) (holding that the imposition of capital punishment in such a case is not violative of the Eighth Amendment).
146. Indeed, this is already happening. Between 1988 and 1992, the number of juvenile cases of crimes against the person judicially waved to criminal court doubled. HOWARD N. SNYDER & MELISSA SICKMUND, U.S. DEP'T OF JUSTICE, JUVENILE OFFENDERS AND VICTIMS: A NATIONAL REPORT 154 (1995).
147. See Elizabeth S. Scott et al., Evaluating Adolescent Decision Making in Legal Contexts, 19 LAW & HUM. BEHAV. 221, 238 (1995) (noting only a single comparative study of this type).
148. Id. at 229-35 (reviewing research comparing conformity and compliance in relation to peers and parents, attitudes toward risk, and temporal perspectives of adolescents and adults).
149. For those who believe - as I do nor - that an offender's prior record should properly be a factor in the culpability assessment for the present offense, disclosure of the juvenile record would also result in lengthier sentences because many younger offenders have records of serious offenses committed while they were juveniles. Among those who believe that prior record is relevant, many would deny that a juvenile record qualifies, because they believe that juveniles are not fully responsible for their deeds.
150. See Gottfredson & Gottfredson, supra note 85, at 468 (reporting that 22.4% of their sample, followed for more than 25 years, never reoffended, and subsequent offenses were trivial rather than serious by more than a three-to-one ratio); see also James Q. Wilson, Crime and Public Policy, in CRIME 489, 492-93 (James Q. Wilson & Joan Petersilia eds., 1995) (reporting consistent findings that about 6% of boys of a given age commit about half of all the serious crime committed by boys of that age, suggesting that the group of potential reoffenders consists of only a small number of individuals).
151. See Gottfredson & Gottfredson, supra note 85, at 468 (reporting that 22.4% of their sample, followed for more than 25 years, never reoffended, and subsequent offenses were trivial rather than serious by more than a three-to-one ratio); see also James Q. Wilson, Crime and Public Policy, in CRIME 489, 492-93 (James Q. Wilson & Joan Petersilia eds., 1995) (reporting consistent findings that about 6% of boys of a given age commit about half of all the serious crime committed by boys of that age, suggesting that the group of potential reoffenders consists of only a small number of individuals).
152. See generally NORVAL MORRIS, MADNESS AND THE CRIMINAL LAW 148-60, 168-72, 183-87, 196-206 (1982); Morris & Miller, supra note 68, at 37-38.
153. See generally NORVAL MORRIS, MADNESS AND THE CRIMINAL LAW 148-60, 168-72, 183-87, 196-206 (1982); Morris & Miller, supra note 68, at 37-38.
154. For a more complete critique, see Stephen J. Morse, Justice, Mercy, and Craziness, 36 STAN. L. REV. 1485, 1493-1503 (1984) (exploring the theoretical and practical problems besetting Professor Morris's proposal).
155. See, e.g., Rummel v. Estelle, 445 U.S. 263, 284-85 (1980) (holding that issues of the mechanics of recidivist statutes and the degree of punishment under them are within the discretion of the punishing jurisdiction); see also Harmelin v. Michigan, 501 U.S. 957, 994-96 (1991) (holding that a mandatory life
156. See, e.g., Rummel v. Estelle, 445 U.S. 263, 284-85 (1980) (holding that issues of the mechanics of recidivist statutes and the degree of punishment under them are within the discretion of the punishing jurisdiction); see also Harmelin v. Michigan, 501 U.S. 957, 994-96 (1991) (holding that a mandatory life sentence for the possession of 650 grams or more of certain controlled substances is not violative of the Eighth Amendment).
157. See, e.g., ZIMRING & HAWKINS, supra note 68, at 3-17 (exploring the rise in incapacitation as today's principal justification for imprisonment in America). See generally Leonard J. Long, Rethinking Selective Incapacitation: More At Stake Than Controlling Violent Crime, 62 UMKC L. REV. 107 (1993) (arguing that general and selective incapacitation are not legitimate exercises of governmental authority in a reasonably affluent, democratic society); Edward P. Richards, The Jurisprudence of Prevention: The Right of Societal Self-Defense Against Dangerous Individuals, 16 HASTINGS CONST. L.Q. 329 (1989) (arguing that courts and legislatures should engage in explicit risk analysis with respect to potentially dangerous people, balancing the risk of harm against the person's liberty interest); David Wood, Dangerous Offenders, and the Morality of Protective Sentencing, 1988 CRIM. L. REV. 424 (arguing that criminal protective sentencing is unjustifiable, but that the civil detention of people identified as dangerous may be justified under certain conditions); Markus D. Dubber, Note, The Unprincipled Punishment of Repeat Offenders: A Critique of California's Habitual Criminal Statute, 43 STAN. L. REV. 193 (1990) (arguing that California's Habitual Criminal statute is inconsistent with the punishment theories of retribution, rehabilitation, deterrence, and incapacitation, and urging California courts to interpret the statute as narrowly as possible).
158. See, e.g., ZIMRING & HAWKINS, supra note 68, at 3-17 (exploring the rise in incapacitation as today's principal justification for imprisonment in America). See generally Leonard J. Long, Rethinking Selective Incapacitation: More At Stake Than Controlling Violent Crime, 62 UMKC L. REV. 107 (1993) (arguing that general and selective incapacitation are not legitimate exercises of governmental authority in a reasonably affluent, democratic society); Edward P. Richards, The Jurisprudence of Prevention: The Right of Societal Self-Defense Against Dangerous Individuals, 16 HASTINGS CONST. L.Q. 329 (1989) (arguing that courts and legislatures should engage in explicit risk analysis with respect to potentially dangerous people, balancing the risk of harm against the person's liberty interest); David Wood, Dangerous Offenders, and the Morality of Protective Sentencing, 1988 CRIM. L. REV. 424 (arguing that criminal protective sentencing is unjustifiable, but that the civil detention of people identified as dangerous may be justified under certain conditions); Markus D. Dubber, Note, The Unprincipled Punishment of Repeat Offenders: A Critique of California's Habitual Criminal Statute, 43 STAN. L. REV. 193 (1990) (arguing that California's Habitual Criminal statute is inconsistent with the punishment theories of retribution, rehabilitation, deterrence, and incapacitation, and urging California courts to interpret the statute as narrowly as possible).
159. See, e.g., ZIMRING & HAWKINS, supra note 68, at 3-17 (exploring the rise in incapacitation as today's principal justification for imprisonment in America). See generally Leonard J. Long, Rethinking Selective Incapacitation: More At Stake Than Controlling Violent Crime, 62 UMKC L. REV. 107 (1993) (arguing that general and selective incapacitation are not legitimate exercises of governmental authority in a reasonably affluent, democratic society); Edward P. Richards, The Jurisprudence of Prevention: The Right of Societal Self-Defense Against Dangerous Individuals, 16 HASTINGS CONST. L.Q. 329 (1989) (arguing that courts and legislatures should engage in explicit risk analysis with respect to potentially dangerous people, balancing the risk of harm against the person's liberty interest); David Wood, Dangerous Offenders, and the Morality of Protective Sentencing, 1988 CRIM. L. REV. 424 (arguing that criminal protective sentencing is unjustifiable, but that the civil detention of people identified as dangerous may be justified under certain conditions); Markus D. Dubber, Note, The Unprincipled Punishment of Repeat Offenders: A Critique of California's Habitual Criminal Statute, 43 STAN. L. REV. 193 (1990) (arguing that California's Habitual Criminal statute is inconsistent with the punishment theories of retribution, rehabilitation, deterrence, and incapacitation, and urging California courts to interpret the statute as narrowly as possible).
160. See, e.g., ZIMRING & HAWKINS, supra note 68, at 3-17 (exploring the rise in incapacitation as today's principal justification for imprisonment in America). See generally Leonard J. Long, Rethinking Selective Incapacitation: More At Stake Than Controlling Violent Crime, 62 UMKC L. REV. 107 (1993) (arguing that general and selective incapacitation are not legitimate exercises of governmental authority in a reasonably affluent, democratic society); Edward P. Richards, The Jurisprudence of Prevention: The Right of Societal Self-Defense Against Dangerous Individuals, 16 HASTINGS CONST. L.Q. 329 (1989) (arguing that courts and legislatures should engage in explicit risk analysis with respect to potentially dangerous people, balancing the risk of harm against the person's liberty interest); David Wood, Dangerous Offenders, and the Morality of Protective Sentencing, 1988 CRIM. L. REV. 424 (arguing that criminal protective sentencing is unjustifiable, but that the civil detention of people identified as dangerous may be justified under certain conditions); Markus D. Dubber, Note, The Unprincipled Punishment of Repeat Offenders: A Critique of California's Habitual Criminal Statute, 43 STAN. L. REV. 193 (1990) (arguing that California's Habitual Criminal statute is inconsistent with the punishment theories of retribution, rehabilitation, deterrence, and incapacitation, and urging California courts to interpret the statute as narrowly as possible).
161. See, e.g., ZIMRING & HAWKINS, supra note 68, at 3-17 (exploring the rise in incapacitation as today's principal justification for imprisonment in America). See generally Leonard J. Long, Rethinking Selective Incapacitation: More At Stake Than Controlling Violent Crime, 62 UMKC L. REV. 107 (1993) (arguing that general and selective incapacitation are not legitimate exercises of governmental authority in a reasonably affluent, democratic society); Edward P. Richards, The Jurisprudence of Prevention: The Right of Societal Self-Defense Against Dangerous Individuals, 16 HASTINGS CONST. L.Q. 329 (1989) (arguing that courts and legislatures should engage in explicit risk analysis with respect to potentially dangerous people, balancing the risk of harm against the person's liberty interest); David Wood, Dangerous Offenders, and the Morality of Protective Sentencing, 1988 CRIM. L. REV. 424 (arguing that criminal protective sentencing is unjustifiable, but that the civil detention of people identified as dangerous may be justified under certain conditions); Markus D. Dubber, Note, The Unprincipled Punishment of Repeat Offenders: A Critique of California's Habitual Criminal Statute, 43 STAN. L. REV. 193 (1990) (arguing that California's Habitual Criminal statute is inconsistent with the punishment theories of retribution, rehabilitation, deterrence, and incapacitation, and urging California courts to interpret the statute as narrowly as possible).
162. See Morse, supra note 108, at 1494-1502, 1505-07.
163. Those adamantly opposed to act retributivism as a necessary and sufficient or as a necessary and limiting justification for punishment will not be convinced, but perhaps pause may be given.
164. Charging practices, plea bargaining, and other vagaries of the criminal justice system undermine this claim, but they also undermine more consequential claims about appropriate incapacitation and deterrence. The argument proves too much. It is plausible but ultimately unproductive to reject all arguments derived from the various justifications for punishment on the ground that the criminal justice system is too flawed to permit any coherent theoretical argumentation about it.
165. See C.S. Nino, A Consensual Theory of Punishment, 12 PHIL. & PUB. AFF. 289, 293-300 (1983) (presenting a justification for punishment based on the theory that offenders consent to forgo their immunity from punishment by committing criminal acts). Larry Alexander suggests that consent has draconian implications, in that it can justify any punishment, no matter how disproportionate. Larry Alexander, Consent, Punishment, and Proportionality, 15 PHIL. & PUB. AFF. 178, 179-82 (1986). But see C.S. Nino, Does Consent Override Proportionality?, 15 PHIL. & PUB. AFF. 183 (1986) (responding to Alexander's critique by setting forth and discussing some of the presuppositions of his own thesis).
166. See C.S. Nino, A Consensual Theory of Punishment, 12 PHIL. & PUB. AFF. 289, 293-300 (1983) (presenting a justification for punishment based on the theory that offenders consent to forgo their immunity from punishment by committing criminal acts). Larry Alexander suggests that consent has draconian implications, in that it can justify any punishment, no matter how disproportionate. Larry Alexander, Consent, Punishment, and Proportionality, 15 PHIL. & PUB. AFF. 178, 179-82 (1986). But see C.S. Nino, Does Consent Override Proportionality?, 15 PHIL. & PUB. AFF. 183 (1986) (responding to Alexander's critique by setting forth and discussing some of the presuppositions of his own thesis).
167. See C.S. Nino, A Consensual Theory of Punishment, 12 PHIL. & PUB. AFF. 289, 293-300 (1983) (presenting a justification for punishment based on the theory that offenders consent to forgo their immunity from punishment by committing criminal acts). Larry Alexander suggests that consent has draconian implications, in that it can justify any punishment, no matter how disproportionate. Larry Alexander, Consent, Punishment, and Proportionality, 15 PHIL. & PUB. AFF. 178, 179-82 (1986). But see C.S. Nino, Does Consent Override Proportionality?, 15 PHIL. & PUB. AFF. 183 (1986) (responding to Alexander's critique by setting forth and discussing some of the presuppositions of his own thesis).
168. Paul Robinson who prefers this solution, raises the compelling point that civil commitment has the attractive feature of periodic review, whereas lengthy sentences do not. Thus, civil commitment may protect the liberty of dangerous offenders more than criminal incarceration. Robinson, supra note 1, at 715. This is true under current legal arrangements, but if offenders deserve lengthy sentences, see supra text accompanying notes 94-96, the absence of periodic review does not endanger their liberty Furthermore, there is nothing to prevent a state from allowing periodic review of enhanced sentences, although doing so blurs the civil/criminal distinction by reintroducing the risk of future violent conduct as a criterion for determining the proper length of a sentence.
169. See supra Part III.C.
170. See supra Part III.A.
171. James Q. Wilson, What to Do About Crime, COMMENTARY, Sept. 1994, at 25, 28 (noting that the peak ages of criminality are between 16 and 18).
172. See United States v. Jackson, 835 F.2d 1195, 1199 (7th Cir. 1987) (Posner, J., concurring) (arguing that life imprisonment for a 35 year old four-time armed robber is too harsh because the likelihood that he would commit armed robbery if released at age 55 after a proposed 20 year sentence is very low).
173. Fox Butterfield, '3 Strikes' Law in California Is Clogging Courts and Jails, N.Y. TIMES, Mar. 23, 1995, at A1, B11 (reporting that whereas 94% of felony cases in California used to be handled by plea bargain, after passage of that state's habitual offender statute, only 14% of second-strike cases and 6% of third-strike cases were disposed of by plea bargain).
174. See Bruce Porter, Terror on an Eight-Hour Shift, N.Y. TIMES, Nov. 26, 1995, § 6 (Magazine), at 42, 44 (describing the increasing levels of rage and violence in prisons and attributing the increase to the lengthier terms prisoners now must serve).
175. Alex Brooks, Michael Corrado, and Ferdinand Schoeman agree. See Brooks, supra note 39, at 752-54 (arguing that such commitment is not only cost-benefit justified but also morally acceptable); Corrado, supra note 1 (manuscript at 11, 14) (balancing the lost freedom of future victims if the state does not detain a potentially dangerous person with that person's loss of liberty if detained); Schoeman, supra note 1, at 27, 32 (analogizing preventive detention to the cost-benefit decision of quarantine situations). Corrado and Schoeman also suggest that in appropriate cases the state owes compensation to people purely preventively detained. See Corrado, supra note 1, at 10, 25-28 (describing compensation as a means to protect the detainee's right to his lost freedom and as a check upon state action); Schoeman, supra note 1, at 31 (identifying compensation as an issue to be considered in a civil preventive detention system). But see Davis, supra note 23, at 95-96 & n.27 (arguing that such incarceration is criminal, not civil, and that the state does not owe the detainee compensation). I test this suggestion in Part VI, infra.
176. Alex Brooks, Michael Corrado, and Ferdinand Schoeman agree. See Brooks, supra note 39, at 752-54 (arguing that such commitment is not only cost-benefit justified but also morally acceptable); Corrado, supra note 1 (manuscript at 11, 14) (balancing the lost freedom of future victims if the state does not detain a potentially dangerous person with that person's loss of liberty if detained); Schoeman, supra note 1, at 27, 32 (analogizing preventive detention to the cost-benefit decision of quarantine situations). Corrado and Schoeman also suggest that in appropriate cases the state owes compensation to people purely preventively detained. See Corrado, supra note 1, at 10, 25-28 (describing compensation as a means to protect the detainee's right to his lost freedom and as a check upon state action); Schoeman, supra note 1, at 31 (identifying compensation as an issue to be considered in a civil preventive detention system). But see Davis, supra note 23, at 95-96 & n.27 (arguing that such incarceration is criminal, not civil, and that the state does not owe the detainee compensation). I test this suggestion in Part VI, infra.
177. Alex Brooks, Michael Corrado, and Ferdinand Schoeman agree. See Brooks, supra note 39, at 752-54 (arguing that such commitment is not only cost-benefit justified but also morally acceptable); Corrado, supra note 1 (manuscript at 11, 14) (balancing the lost freedom of future victims if the state does not detain a potentially dangerous person with that person's loss of liberty if detained); Schoeman, supra note 1, at 27, 32 (analogizing preventive detention to the cost-benefit decision of quarantine situations). Corrado and Schoeman also suggest that in appropriate cases the state owes compensation to people purely preventively detained. See Corrado, supra note 1, at 10, 25-28 (describing compensation as a means to protect the detainee's right to his lost freedom and as a check upon state action); Schoeman, supra note 1, at 31 (identifying compensation as an issue to be considered in a civil preventive detention system). But see Davis, supra note 23, at 95-96 & n.27 (arguing that such incarceration is criminal, not civil, and that the state does not owe the detainee compensation). I test this suggestion in Part VI, infra.
178. Alex Brooks, Michael Corrado, and Ferdinand Schoeman agree. See Brooks, supra note 39, at 752-54 (arguing that such commitment is not only cost-benefit justified but also morally acceptable); Corrado, supra note 1 (manuscript at 11, 14) (balancing the lost freedom of future victims if the state does not detain a potentially dangerous person with that person's loss of liberty if detained); Schoeman, supra note 1, at 27, 32 (analogizing preventive detention to the cost-benefit decision of quarantine situations). Corrado and Schoeman also suggest that in appropriate cases the state owes compensation to people purely preventively detained. See Corrado, supra note 1, at 10, 25-28 (describing compensation as a means to protect the detainee's right to his lost freedom and as a check upon state action); Schoeman, supra note 1, at 31 (identifying compensation as an issue to be considered in a civil preventive detention system). But see Davis, supra note 23, at 95-96 & n.27 (arguing that such incarceration is criminal, not civil, and that the state does not owe the detainee compensation). I test this suggestion in Part VI, infra.
179. Alex Brooks, Michael Corrado, and Ferdinand Schoeman agree. See Brooks, supra note 39, at 752-54 (arguing that such commitment is not only cost-benefit justified but also morally acceptable); Corrado, supra note 1 (manuscript at 11, 14) (balancing the lost freedom of future victims if the state does not detain a potentially dangerous person with that person's loss of liberty if detained); Schoeman, supra note 1, at 27, 32 (analogizing preventive detention to the cost-benefit decision of quarantine situations). Corrado and Schoeman also suggest that in appropriate cases the state owes compensation to people purely preventively detained. See Corrado, supra note 1, at 10, 25-28 (describing compensation as a means to protect the detainee's right to his lost freedom and as a check upon state action); Schoeman, supra note 1, at 31 (identifying compensation as an issue to be considered in a civil preventive detention system). But see Davis, supra note 23, at 95-96 & n.27 (arguing that such incarceration is criminal, not civil, and that the state does not owe the detainee compensation). I test this suggestion in Part VI, infra.
180. Alex Brooks, Michael Corrado, and Ferdinand Schoeman agree. See Brooks, supra note 39, at 752-54 (arguing that such commitment is not only cost-benefit justified but also morally acceptable); Corrado, supra note 1 (manuscript at 11, 14) (balancing the lost freedom of future victims if the state does not detain a potentially dangerous person with that person's loss of liberty if detained); Schoeman, supra note 1, at 27, 32 (analogizing preventive detention to the cost-benefit decision of quarantine situations). Corrado and Schoeman also suggest that in appropriate cases the state owes compensation to people purely preventively detained. See Corrado, supra note 1, at 10, 25-28 (describing compensation as a means to protect the detainee's right to his lost freedom and as a check upon state action); Schoeman, supra note 1, at 31 (identifying compensation as an issue to be considered in a civil preventive detention system). But see Davis, supra note 23, at 95-96 & n.27 (arguing that such incarceration is criminal, not civil, and that the state does not owe the detainee compensation). I test this suggestion in Part VI, infra.
181. Corrado, supra note 1, at 1 n.1.
182. An equally clear and unsettling related conclusion is that potentially dangerous people need not cause or threaten harm to justify pure preventive detention. Detaining people who have not given any direct indication that they might behave danger-ously would be justified if the predictive technology were accurate enough.
183. See Wilson, supra note 106, at 492 (noting that criminologists have identified characteristics of the group of youths who will commit serious crimes, but acknowledging that predictive technology is too imprecise to identify the particular members of this group).
184. Cf. MODEL PENAL CODE § 211.2 (1985) (Recklessly Endangering Another Person). I characterize the proposal as "purely heuristic" because it would be wildly intrusive and a nightmare to administer, and I would therefore not adopt it. It raises theoretically interesting issues, however, and provides an interesting comparison to schemes of civil preventive detention that try to accomplish similar goals of protecting society from potential harmdoers.
185. I have used this hypothetical in class for years, but learned after writing this Part that Michael Davis recently and independently has made a very similar proposal. Professor Davis's paper has since been published and should be compared with the present Part. See Davis, supra note 23. In brief, rather than treating the proposal as a heuristic device, Davis argues that criminal punishment is genuinely warranted and, indeed, is a proper response to such cases. He argues that Schoeman's justification for pure civil detention fails, even if all the technological and procedural problems were solved. As the last Part makes clear, I tend to agree with Schoeman. As the present Part suggests, however, expanding reckless endangerment is not a theoretically unthinkable alternative to pure preventive detention, although enforcing it would be nightmarish.
186. Consider the following further hypothetical. Imagine that an agent is consciously aware of a substantial risk that he has a deadly, highly communicable disease that is transmitted by normal social interaction. Transmission can be prevented, however, by simple means that allow the agent safely to engage in normal social interaction. Suppose he fails to take those steps and communicates the disease with conscious awareness of a substantial risk that he would communicate it. The victim dies. Is a conviction for involuntary manslaughter or murder unjustifiable? If not, is it unthinkable to consider the omission to take the necessary steps to prevent communication sufficient to warrant conviction and punishment for reckless endangerment? The criminal law already includes prohibitions against the active intentional or reckless transmission of some serious diseases. See, e.g., Weeks v. Scott, 55 F.3d 1059 (5th Cir. 1995) (affirming HIV-positive defendant's conviction for attempted murder by spitting at the victim with intent to kill); Simon Bronitt, Spreading Disease and the Criminal Law, 1994 CRIM. L. REV. 21 (exploring the difficulties in bringing conduct which contributes to the spread of disease within the scope of the criminal law).
187. Consider the following further hypothetical. Imagine that an agent is consciously aware of a substantial risk that he has a deadly, highly communicable disease that is transmitted by normal social interaction. Transmission can be prevented, however, by simple means that allow the agent safely to engage in normal social interaction. Suppose he fails to take those steps and communicates the disease with conscious awareness of a substantial risk that he would communicate it. The victim dies. Is a conviction for involuntary manslaughter or murder unjustifiable? If not, is it unthinkable to consider the omission to take the necessary steps to prevent communication sufficient to warrant conviction and punishment for reckless endangerment? The criminal law already includes prohibitions against the active intentional or reckless transmission of some serious diseases. See, e.g., Weeks v. Scott, 55 F.3d 1059 (5th Cir. 1995) (affirming HIV-positive defendant's conviction for attempted murder by spitting at the victim with intent to kill); Simon Bronitt, Spreading Disease and the Criminal Law, 1994 CRIM. L. REV. 21 (exploring the difficulties in bringing conduct which contributes to the spread of disease within the scope of the criminal law).
188. Cf. MODEL PENAL CODE § 2.02(7) (1985) (the Code's treatment of "willful blindness"). The section provides that knowledge of a particular fact is satisfied by knowledge of a high probability that the fact exists, unless the defendant actually believes that it does not exist.
189. In part this will be true because the state will have to prove that harm was substantially likely, a predictive enterprise subject to all the difficulties this Paper has already discussed. Still, the defendant's awareness that he or she was likely to cause harm - which also must be proved - surely increases the predictability of the harm.
190. Perhaps, contrary to my expectation, such a scheme would be widely (over)used.
191. See Larry Alexander, Mens Rea and Inchoate Crimes 2-8 (1995) (unpublished manuscript, on file with the author).
192. Kevin Reitz suggested in a personal communication that this proposal is more intrusive than the other preventive detention schemes that I reject because it requires the potentially dangerous person to turn himself in. The proposal is intrusive and I would not in fact adopt it. But the offender's culpability justifies the intrusion and in most cases the potential harmdoer can probably avoid both prison and preventive detention by seeking less intrusive means to prevent harmdoing.
193. JONATHAN SWIFT, A MODEST PROPOSAL (1729), reprinted in JONATHAN SWIFT: A SELECTION OF His WORKS 477 (Philip Pinkus ed., 1965).
194. JONATHAN SWIFT, A MODEST PROPOSAL (1729), reprinted in JONATHAN SWIFT: A SELECTION OF His WORKS 477 (Philip Pinkus ed., 1965).