ISSN 1862-2941

WelcomeOnline-Issues2-20191-20191/2-20182-20171-20172-20161-20162-20151-20152-20141-20142-20131-20132-20121-20122-20111-20112-20101-20102-20091-20092-20081-20082-20071-20073-20062-20061-2006Guidelines to authorsAbout usLegal NoticePrivacy



Online-Issues » 1-2007 » Daniel F. Montaldi

The Logic of Sexually Violent Predator Status in the United States of America

Daniel F. Montaldi
Evaluation Team, Arizona Community Protection and Treatment Center

[Sexual Offender Treatment, Volume 2 (2007), Issue 1]


Sexually Violent Predator (SVP) laws have placed great legal weight on psychosexual evaluations of sex offenders by mental health experts.  The conclusions of these evaluations are used to civilly commit hundreds of offenders throughout the United States after the completion of their criminal sentences, possibly for life.  This paper examines the reasoning used by evaluators and attorneys for the state to justify the claim that someone is SVP.  I discuss serious flaws in this reasoning and show how argument for SVP status must proceed if the case for civil commitment is to be logically coherent and consistent with constitutional values. 

Key words:  sexually violent predator, forensic evaluation, risk assessment, paraphilia, hypersexuality, compulsive sexual behavior.

This paper examines the logic of Sexually Violent Predator (SVP) status as articulated by the U. S. Supreme Court in Kansas v. Hendricks, 521 U.S. 347 (1997), and Kansas v. Crane, 534 U.S. 407 (2002).1 I argue that the Court sets out a sensible principle of parsimony to guide SVP reasoning but obscures the importance of this principle by introducing an unjustified assumption about SVP-relevant mental disorders. This assumption opens the door to thinly reasoned psychosexual evaluations and professional opinions of little use to a finder of fact in an SVP case. An explanation of proper SVP reasoning is offered, as well as an introductory discussion of a clinical construct related to “volitional” forms of sexual impairment (hypersexuality).
Some readers will interpret this discussion as undermining the legitimacy of American SVP laws. Critics of SVP laws are not few and their arguments are powerful (e.g., Schopp, 2001). However, the argument presented here does not entail that sex offender civil commitment is illegitimate as such. It does restrict SVP eligibility quite sharply, such that a “narrow” category (Crane) becomes narrower still. As a society we are rightly concerned about individuals who have served their sentences yet remain sexually dangerous. But for good reason the Constitution greatly restricts state power to confine persons who have not committed (new) crimes, whatever their propensities (or our estimations thereof). Serious consideration of the alternative – history offers plenty of examples – is sobering.
Other readers might interpret this discussion as a defense of SVP laws, by way of modification. It is not. Rather, the argument takes this form: Given that SVP laws exist, there are better and worse ways to interpret their basic concepts and apply them to individual cases. If the interpretation and application of these laws is to be even minimally logical and faithful to the constitutional basis provided by the U. S. Supreme Court, then certain claims follow about how mental health experts should evaluate someone for SVP status, and how the state should argue its case in court. The discussion that follows does not accept Supreme Court judgment uncritically; indeed, the Court’s global assumption about paraphilias is rejected as a baseless throwback to 19th century stereotypes. However, the argument below does follow what is basically correct reasoning by the Court concerning SVP logic. Genuinely useful SVP assessment and sound case presentation are possible, although they are more difficult than current practice would suggest. In a nutshell, this paper is aimed at improving the intellectual integrity of SVP work, for as long as society wants it done.
Sex offender civil commitment laws and related Supreme Court opinions are not intelligible outside of wider Constitutional considerations that give meaning to SVP concepts and dictate the direction of reasoning toward SVP conclusions. In theory – practice aside – SVP status is consistent with what our Constitution outlines with regard to proper powers of state and individual rights. However, the state must walk a line that is tightrope thin. [Some degree of incoherency is unavoidable if SVP statutes remain as they are: Civil commitment in lieu of, or after mitigated criminal sentencing, is arguable logically. Civil commitment post-sentence, barring change in an offender’s mental status in prison, is not.] Unfortunately, the logic of SVP reasoning is mostly ignored or inverted by many clinical evaluators, attorneys, and judges.

Constitutional Foundations

Even without Supreme Court opinions, there is a basic SVP logic embedded within wider notions of personal autonomy, equality before the law, and protective paternalism, that is inescapable if SVP commitment is to avoid violating basic liberties and slipping down the slope into the “preventive state” (Janus, 2006). The bottom line is that the Court’s notion of “serious difficulty controlling behavior” (Crane) is necessarily the core component of SVP status. This is true out of logical necessity. “Mental disorder” (or “mental abnormality”) and “likely” (or in some states, “more likely than not”, “probable”, “highly probable”) sexual violence cannot be anything but derivative notions: The first is the psychological condition that creates “serious difficulty” and the second is the natural trend and expectable prospect for anyone both drawn to sexual crime and suffering from serious difficulty controlling sexual behavior. Without some kind of volition-related impairment, mental disorder is no more than a kind of motivation for crime and cannot, as such, distinguish the volitionally dangerous (a large group) from the dangerously impaired (a narrow group). The same is true about likelihood of future violent crime. Both the volitionally dangerous and the dangerously impaired are likely to do violent acts; otherwise they would not be dangerous. Of course, from reading the typical SVP evaluation one would think “likely” sexual violence is the only issue (and that our principal guide in assessing an individual’s dangerousness consists of observed reconviction rates in samples differentiated by criminal history but not psychological reasons for violent behavior). One would also get the impression that “mental disorder” is no more than a DSM label for virtually any sexual recidivism with “serious difficulty” as an alternative label in legal-speak.
If we proceed logically, and constitutionally, it follows that offenders not accused or suspected of a new crime are eligible for civil commitment only if they are importantly different from other offenders, such that the basic moral and/or empirical assumptions grounding our criminal justice system do not apply to them and special legal status is warranted. Otherwise, these offenders should be handled like any other offender where further confinement awaits further crime. In Foucha v. Louisiana, 504 U.S. 71 (1992) the Supreme Court is clear that criminal dangerousness per se is not a sufficient reason to justify involuntary hospitalization when there has not been (or suspicion of) a new crime. Punishing crime (retributive justice) and protecting society from dangerous criminals (deterrence, segregation) are the primary functions of the criminal justice system. Protecting citizens from the dangers of crime is not the primary function of civil and mental health law (Schopp, 2001). Criminal law can incorporate considerations of dangerousness along with guilt when someone is sentenced. Civil commitment law applies only to those dangerous persons whose risk owes to some form of incapacitation, or incompetence, such that their risk to self or others lies beyond their control. Clearly, no one, criminal or non-criminal, controls whether or not they ever have motivating feelings (dispositions, predispositions) inclining them toward bad acts. Thoughts and feelings are not against the law. What people choose is their response to inclination. Civil commitment makes sense only when the dangerous person cannot be said to choose either their motivations for harmful acts or their responses to these motivations.
Typical offenders are distinguished from atypical and commitment eligible offenders by whether or not they satisfy the basic assumptions of both deterrence logic and criminal responsibility. Deterrence is the most basic strategy through which the criminal justice system protects society from criminals. Deterrence policy presupposes, as necessary conditions, that persons have the ability to understand what is wrong and illegal (e.g., killing under most circumstances is wrong and prohibited), awareness of the empirical nature of one’s behavior (e.g., stabbing someone with a knife is not tickling their ribs), the capacity to appreciate legal contingencies (crime brings punishment; punishment is aversive), and the ability to make behavior-related decisions consistent with this understanding and appreciation. Unless persons can be assumed to possess at least these abilities, they cannot respond to deterrence-related incentives. A policy of deterrence would be unintelligible, not just ineffective.
Criminal responsibility is largely the same, although not quite identical. Deterrence logic is incoherent if a person lacks any one of the above-noted abilities. However, only certain deficits negate responsibility. Someone lacking in the ability to understand law or legal contingencies is not considered criminally responsible. Nor is someone truly unable to adjust his behavior to the law (although utter inability, or ‘irresistible impulse,’ is now considered by most courts to be impossible to determine and no longer part of legal insanity criteria in most if not all states). In contrast, someone unable to appreciate (emotionally) the punitive quality of prison, and/or unable to feel even minimal empathy or attachment to others, would still be responsible. Also responsible would be someone possessing limited, yet some ability to obey the law, despite adequate knowledge and appreciation. The responsibility of the latter person might be mitigated, depending on the extent of his limitations, but it would still exist. Not so for the first person, provided that he meets other conditions. Whether or not he cares about prison is irrelevant; he can avoid crime, and its legal consequences, if he so chooses, giving him a fair chance to avoid state-imposed loss of liberty. Likewise, neither deterrence logic nor accountability presupposes conscience, empathy, or respect for law. Their intelligibility rests only on the presumption that people possess some sense of self-interest, at least modest powers of deliberation, and basic awareness of the fact that society has rules and institutions for their enforcement.2
Commitment-eligible persons are individuals who fail to satisfy the necessary, not just the sufficient conditions of deterrence. Necessary conditions of deterrence logic are quite different from sufficient conditions of deterrence efficacy. Many offenders are not deterred by their basically accurate understanding of law and punishment, and even by the experience of punishment. And they are not likely to be deterred in the future (perhaps ending criminal behavior when they get older and less inclined). Someone lacking the ability to respond to deterrence incentives is very different from someone who simply ignores them. Any dangerous offender is someone not likely to be deterred (someone not apt to respond to deterrence-related incentives). Recidivism in itself suggests little responsiveness to deterrence incentives. If all that matters for commitment eligibility is whether or not someone is likely to be deterred, then dangerousness alone should be the basis of eligibility and all dangerous and/or recidivating offenders should be committed (contra Foucha).3 [Mental disorder diagnoses often amount to little more than alternative labels for dangerousness. It is difficult to imagine the dangerous offender who would not receive some DSM diagnosis, especially with “Not Otherwise Specified” categories.]
This is why “serious difficulty controlling behavior” is, and must be the pivotal concept in SVP logic, and why “mental disorder” and “likely sexual violence” are derivative notions. Anyone attracted to harmful acts and seriously limited in capacity to resist these impulses even if he wanted to, is a person disordered by any reasonable standard of typical human abilities, even for criminals. The label we use for it is immaterial, DSM or not. By any reasonable expectation, someone suffering from such a condition is extremely dangerous, whether he is near the beginning of his formal criminal career (placing him in an actuarial category from which only a small percentage of sampled offenders have been observed to recidivate) or well into his career (such that a substantial percentage of his sampled peers have been observed to recidivate). His risk is not logically contingent on group characteristics (e.g., a percentage reconvicted), much less conceptually dependent on samples of offenders whose mental conditions were not assessed in detail (and who were released many years ago in an era of shorter sentences and periods of probation, less public awareness, and little treatment according to current standards).
In this writer’s experience at least, few if any SVP evaluations conducted by mental health experts – “competent professional” in statutory language – specifically assess for the possible absence of, or serious deficiencies in, those psychological abilities or capacities considered necessary for deterrence logic. They either ignore this fundamental issue or toss off shallow assumptions based on the mere fact of recidivism, which addresses only someone’s likelihood of being deterred (sufficient conditions). For example, repeat sex offenders are casually said to “lack control” given that they repeatedly offended despite being punished. Unfortunately, this phrase carries no useful clinical or legally significant evaluative meaning because it does not discriminate between “lacking in ability to control,” which is relevant to civil commitment logic, and “lacks will to control,” which is not.
Courts do not need expensive evaluators telling them the obvious fact that someone who committed crimes did not control himself enough to avoid committing crimes. Nor do they really need experts to say that a substantial percentage of offenders with multiple past offenses are apt to re-offend. Actuarial research or not, most sensible observers would consider it risky to leave offenders like this completely free in the community after they finish their sentences. What fact finders need in an SVP trial is a mental health expert who can help them determine whether an obviously risky offender is “beyond control,” in a manner that makes him stands out from other offenders who have sexually recidivated. Quantifying the obvious is fine; it is useful from a criminological perspective and it is useful for tailoring treatment/supervision strategies to different offender groups. If SVP commitment was based on a prediction of future crimes, use of reconviction rates is superior to clinical judgment alone.4 But SVP status is not merely suitability for a given level of treatment or monitoring. It is a clinical/normative assessment whose outcome (except in Arizona) is years of lost freedom with no end in sight.5 SVP status does not depend on evaluators coming up with definite predictions that someone will reoffend. It is contingent on proving in a specific individual’s case the presence of so severe an impairment to sexual self-regulatory functions that any reasonable person understanding the nature of this condition would have grave concerns about further unregulated and harmful sexual behavior given this person’s attractions. These are tough issues that do not lend themselves to direct measurement or mechanical procedures, which is precisely why mental health experts are needed rather than criminologists or actuaries.

Principle of Parsimony

Philosophers and scientists are familiar with the Principle of Parsimony, or “Ockham’s Razor,” which holds that the best or most economical theory or explanation is the one that covers all relevant facts with the fewest assumptions and constructs. The principle of parsimony can be applied to explaining behavior, crime in particular. The vast majority of criminal acts, including many shocking and repugnant, seem well explained (psychologically) by the same concepts we use to explain most human behavior: motive, intention, attitude, habit, character, and circumstance. Predatory and harmful behaviors of wide variety are not uncommon worldwide and therefore would seem to be within the wide “normal” limits for human beings as a group. Unless criminality in general – or particularly heinous crime, or chronic criminality – is considered clinically disordered just as such, there is no compelling reason to bring in concepts from psychiatry or psychopathology to account for why an individual engages in criminal activity. For this reason, any account of a person’s crimes that depends on ideas from abnormal psychology is not the simplest of explanations. It might be the best explanation, however, provided that important features of the person’s offending behavior or state of mind are not explained by other concepts, features that are abnormal even for criminals.
Consider the crime of rape. One simple explanation attributes this behavior to motives, attitudes and habits not uncommon among criminals. The rapist wants sex, encounters resistance, feels entitled, and decides to take what he wants anyway. This is often secondary to his general approach to life, which is to live in the moment, selfishly taking whatever is wanted, whenever it is wanted. Whether the crime of rape happens once in the offender’s lifetime, or more than once, this is a plausible account. A paraphilia diagnosis, unless used merely as an alternative label for this sort of crime, would just introduce an additional construct without there being additional facts not explained by the simpler account.
On the other hand, there are rape patterns not well explained by this account. These rapists have not only repetitive rapes and the motivation to continue raping (rape-related fantasies and urges) but also unusually strong motivating feelings and little experience coping with feelings in pro-social manner. These individuals have conditioned and habituated themselves to the point of feeling “driven,” even compulsive. They may or may not want their urges or an end to their behavior, and their attitudes might be similar to those who feel more in control of their criminal impulses. The point is that regardless of the interdependence of offense-supportive attitudes and deviant urges, the attitudes of these rapists no longer play a primary role in the reason they act. They rape because the mental condition that generates their urges and the pressure these urges place on (or against) deliberation, has taken on a life of its own, capable of “overwhelming” or “overpowering” the person’s ability to weigh options according to perceived values. Such a rapist would likely be someone with multiple assaults closely proximate in time, dozens of sexual assaults over the years, a need for rape fantasies in order to climax, and an abnormal frequency of masturbation (e.g., more than once per day, and always to violent fantasies). He feels uncomfortably pressured by his urges to rape, anxious and distracted when not acting out. This man seeks out non-consenting victims to coerce according to a repetitive sexual script that preoccupies him in fantasy. Understanding this kind of pattern requires more than reference to sexual desire, attraction, thoughts, and urges, coupled with the willingness to pursue gratification regardless of law. A sexually abnormal or atypical condition is suggested. This is a condition not common to criminals, even rapists.

Principle of Narrow Class

Supreme Court opinions in Hendricks and Crane provide what is essentially a legal principle of parsimony governing sexual dangerousness and SVP status. Essentially, this principle requires the state to regard criminally disposed persons as acting voluntarily when they commit crimes (and therefore responsible for their actions), unless proven otherwise. The Court presumes the “typical” offender is not significantly impaired in his ability to obey the law, even if he is chronically unwilling to exercise that ability. The burden of proof does not fall on the side arguing that someone is volitional, but on the side arguing that he is impaired. This way, offenders are treated like any other citizen in being moral agents and choosers (factors within one’s control), whatever their different impulses or attractions, external circumstances, or backgrounds (factors often beyond one’s control). Contrary to some interpretations of the Court’s argument (Faigman, 2004) it is too strong to say that the Supreme Court is presuming an abstract doctrine of “free will” in its concept of an average or typical offender. The Court is merely assuming normal human agency unless proven otherwise. This is both logical and practical in any society that values both community safety and personal autonomy.
Distinguishing a “narrow” class of dangerously impaired sex offenders from a larger “typical” class of volitionally dangerous offenders, forces the state to argue parsimoniously when it presents an SVP case. The notion of a “typical recidivist” sets a logical and legal default setting for offering the court an explanation of someone’s past crimes and current dangerousness. According to the Court (Crane), “typical recidivists,” unlike SVP recidivists, are “dangerous but typical (persons)…who are perhaps more properly dealt with exclusively through criminal proceedings.” Distinguishing SVP offenders from this larger group of dangerous criminals is necessary “lest civil commitment become a mechanism for retribution or general deterrence.” The Court is directing the state to operation from the presumption that any sexual recidivist is like most criminal recidivists in meeting the necessary conditions of deterrence logic (and fairness), unless the state can prove otherwise. Civil commitment requires successful proof of SVP status (with a burden as high as “beyond reasonable doubt” in some states, e.g., Arizona).

Typical Recidivists

Unfortunately for those conducting actual SVP cases, the Supreme Court in Crane is vague with its terms. Who are “typical recidivists”? Are typical recidivists (TR offenders) repetitive criminals but not sex offenders? Are they general criminal recidivists, but recidivists with one sex offense? Or are they typical sexual recidivists (TSR offenders)? Some authors (e.g., Vognsen & Phenix, 2004, interpreting California’s SVP statute) construe “typical recidivist” as a chronic criminal who might have one opportunistic sex crime but generally offends for criminological reasons as opposed to special attraction to deviant sex.6 Other authors (e.g., Colb, 2002) allow that typical recidivists might have multiple sex offenses and even deviant sexual interests.
The most logical interpretation would seem to be that of the Arizona Supreme Court in Leon G. (389 Ariz. Adv. Rep. 6, 59 P.3d 779, 2002) which follows Crane and construes “typical recidivists” as “dangerous but typical (persons who) choose to commit acts of sexual violence (emphasis added).” In other words, typical recidivists are sex offenders who differ from SVP sex offenders by their lack of volition-related impairment. No mention is made of typical recidivists being only one-time sex offenders or sex offenders with fewer sexual offenses than SVP offenders. So it would appear there are “typical” sexual recidivists. Qua voluntary, TSR offenders satisfy the necessary conditions of deterrence because they are capable of obeying the law, as much as any criminal is capable, however unwilling they may be to do so.7

Sexually Violent Persons

In contrast to typical recidivists, SVP offenders (according to Crane) are individuals “unable to control their dangerousness” who possess the “critical distinguishing feature of…serious disorder (consisting of) a special and serious lack of ability to control behavior.” Unlike the higher threshold of “irresistible impulse” (the now discarded second prong of legal insanity), this lack of control need not be a “total or complete lack of control (italics in original text).” According to the Court, “the line between an irresistible impulse and an impulse not resisted is probably no sharper than that between twilight and dusk…most severely ill people, even those commonly termed psychopaths, retain some ability to control their behavior.” [The issue of whether psychopaths have volitional or other impairment is discussed later.] To commit someone as an SVP, the state must give “proof of serious difficulty controlling behavior.”
The Court acknowledges a lack of specific criteria for “serious difficulty” when they say, “(I)nability to control behavior will not be demonstrable with mathematical precision.” Most offenders with “serious difficulty” will possess “volitional impairment,” i.e., inordinate difficulty restraining criminal impulses through voluntary choice. However, the Court does not draw clear lines here, given that volitional impairment was the relevant to the Hendricks case and the Court usually narrows its attention to issues raised by the specific case prompting their review. Thus, the issue of what kinds of impairment or incapacitation should be included under “serious difficulty” has yet to be addressed. As such, the Court allows for the possibility of a wide category, without making this definite just yet: “(We) do not draw a clear distinction between the purely emotionally sexually related mental abnormality and the volitional…Nor, when considering civil commitment, have we ordinarily distinguished for constitutional purposes among volitional, emotional, and cognitive impairments. Here, as in other areas of psychiatry, there may be considerable overlap between a defective understanding or appreciation and [an] ability to control behavior.” This ambiguity allows individual states to determine the nature of impairment relevant to SVP status. Following the Arizona Court of Appeals in its interpretation of Crane, the Arizona Supreme Court (Leon G.) regards SVP mental disorder as “making” or causing serious control difficulty by virtue of “impair[ing] or tend[ing] to overpower the person’s ability to control his or her behavior.” This interpretation seems sensible and supports a mostly volition-based interpretation of “serious difficulty.”

TSR and SVP Prototypes

For any SVP evaluator taking the law seriously, the Court’s reasoning demands a parsimonious chain of clinical reasoning. The evaluator must first rule out the possibility that someone is dangerous yet typical, before reaching an opinion that he is dangerous because of atypical impairment. So before considering the nature of an SVP pattern of sexual offending (and current sexual mental status), the evaluator should consider the characteristics of a TR/TSR sexual pattern. Although this is no simple matter, it seems possible to formulate a plausible example of a TSR offender:
Imagine a man who has molested several children over the years. Suppose he is attracted to children but also to adults, so he does not “need” children for sexual satisfaction. His urges toward children are not isolated or rare (they are “recurrent”) and they can be powerfully stimulating (“intense”), at least as much as non-deviant sexual fantasies can be stimulating to the average person (or typical criminal). However, these feelings are not preoccupying, obsessive, or severely distressing, and the man does not feel driven, compelled, unpredictably impulsive, or pressured when he is aroused or in close proximity with an object of attraction. He does not masturbate any more frequently than the average man his age and he can climax to non-deviant fantasies. This man can feel attracted and aroused by a young child, something the average man (and typical criminal) does not, but he can also remain calm enough to think about whether pursuing gratification is “worth it.” He can be sexually aroused and interested without being trapped in sexual “tunnel vision.” This man exploits some opportunities for deviant sex but not others. Sometimes, even often, he is impulsive and rash in his offending, like other criminals (especially when intoxicated). At other times, he is careful and deliberate. Either way, this man can at least think beforehand whether the circumstances are right. For example, he is capable of exercising enough patience to manipulate a child into a good opportunity for sex rather than acting immediately on impulse. This offender might have a variety of non-sexual offenses in addition to his sexual offenses, and the decision-making processes behind these offenses are not unlike those behind his sex crimes. In this respect our TSR offender is similar to many “career” criminals.
Our typical sexual recidivist would satisfy DSM criteria for Pedophilia, and probably Antisocial Personality Disorder. He is highly dangerous sexually by any reasonable consideration based on his attitudes, attractions, and past behavior.8 Nonetheless, it cannot be argued that this man’s DSM mental disorders are the primary causal factors behind his dangerousness, rather than choice. Paraphilia and personality disorder, as these disorders present in this individual, do not rise to the threshold of SVP “mental disorder” because they do not cause this man “serious difficulty” controlling sexual behavior. This man’s paraphilia can be said to “make” him attracted to children, but it cannot be said to make him molest, or be likely to molest, because his decision-making abilities are still intact. We can say that his paraphilia “disposes” or “predisposes” him to sexually offend, but in the weak sense of causing ongoing incentive to offend, not in the strong sense of overpowering volitional capacity to not offend. If anything can be said to make this man dangerous, it is his attitude-determined, and voluntary, willingness to do as he pleases without respect for law or other people. Regardless of whether clinicians call this a mental disorder, it consists of purposive actions and decisions not qualitatively different from those of many dangerous criminals. This individual is dangerous, but willfully dangerous, not dangerous “beyond control”.
Now consider the atypical sexual recidivist (a potential SVP). An example would be a man who is sexually fixated on children. He has offended against children many times. He does not have attractions to adults. This man masturbates multiple times per day and spends substantial time cruising to see (or obtain) children. He feels anxious or restless when not engaged in sexual pursuits and he is seldom without deviant thoughts, even when he would like to be. This pattern might wax and wane depending on events in his life, but the pattern is never dormant for long. This man feels preoccupied by his sexual thoughts, distracted, sometimes tormented, and masturbation does not seem to satisfy or allow him to be comfortable with non-sexual pursuits. Such an individual might be both an opportunist at times and a deliberate predator at other times. He exploits opportunities when they arise and also seeks out opportunities. This man is not a true obsessive-compulsive; his urges to molest are not as compelling as the urges of a compulsive “checker” or hand-washer, or someone who is bulimic. He does have some degree of control, for example, he can groom a child and bide his time. But he cannot be without some kind of sexual pursuit or activity for long. This man is more easily aroused than the average man when confronted by an object of sexual interest and his physiological reactions might be unusually strong given his circumstances, e.g., the mere sight of a child in close proximity might cause his heart to pound. Such an offender might well have non-sexual crimes in his background and he might have much the same antisocial attitudes as our TSR offender. Or he might not. But regardless of attitude, this man has more difficulty thinking straight in sexual situations than he has in non-sexual situations, even given his own baseline of meager thoughtfulness and personal responsibility.
Our atypical, potentially SVP recidivist certainly satisfies DSM clinical criteria for Pedophilia, and perhaps also some personality disorder (APD or not). However, his paraphilia is much more than a recurring pattern of powerful fantasies and urges (and related behavior). There is a crucially important additional aspect of the disorder in this case, i.e., its hypersexual severity, which makes this man’s psychological condition a SVP-relevant “mental disorder.” Using an analogy from electrodynamics, we might say that this man’s deviant interest is the current behind his sexual offending. The hypersexuality provides the voltage. It is the voltage, not the current, “overpowering” this man’s ability to use the switch and restrain sexual behavior. Our TSR offender has the same current but not the same voltage; he can use his switch if he chooses. Roughly speaking, it is this man’s hypersexuality that drives his paraphilia to the point where he is “made” dangerous beyond typical criminal volition, even for sex offenders.9

Ramifications of SVP Parsimony for Evaluations

If this line of argument is taken seriously, then most SVP evaluations fail to support the claim that someone meets criteria for Sexually Violent Person. At risk of only some exaggeration, the average SVP evaluator probably assigns a paraphilia diagnosis to any sexual recidivist, provided that two or more of the individual’s sex crimes are similar in kind and at least six months apart (if that). A diagnosis of personality disorder – Antisocial Personality Disorder in particular – will be attached to most any pattern of diverse recidivism, perhaps with additional relationship difficulties, unstable employment, and substance abuse. [Personality Disorder NOS, Antisocial Traits, is commonly seen when there is little information about juvenile criminality or other behaviors.] Thus, for many evaluators, their assignment of a paraphilia-related “mental disorder” will not distinguish SVP sexual recidivists from TSR sexual recidivists. Their diagnosing a personality disorder will not distinguish SVP from TR (general) recidivists.
Even if SVP recidivists were the only recidivists assigned diagnostic labels, diagnostic decisions based solely on some threshold number of overt behaviors will imply nothing about the functional capacity of someone’s control-related mental processes. So they will not support the opinion that the person was “overwhelmed” by urges. In most cases, external criminal behaviors demonstrate no more about mental states or processes than the mere presence of urges or attractions. In general, they indicate nothing about intensity of urges and attractions, nor do they imply anything about the level of psychological resources available to the person to avoid acting on sexual impulses, intense or not.
In most evaluations, then, the only basis for discriminating between SVP and non-SVP offenders is risk assessment, which in recent years is construed in mostly actuarial terms. Of course, risk assessment can help exclude from SVP categorization many (not all) of those offenders who pose less than high risk.10 But the Hendricks and Crane opinions are concerned with discriminations within the high risk group, not with distinguishing high from low risk. In Leon G., the Arizona Supreme Court addresses this issue of danger threshold, but only to clarify an ambiguous term (“likely”). The Arizona Court describes SVP offenders as “extremely” dangerous and “highly probable” to do acts of sexual violence. However, typical recidivists are still distinguished from SVP recidivists by means of “serious difficulty,” not “likely”.11

SVP Logic

The Court’s principle of parsimony is essentially two principles: (1) SVP only if not TR or TSR. (2) SVP must be a narrow class. Based on this reasoning, sexual recidivism itself cannot be considered prima facie reason to think the recidivist has serious trouble controlling his behavior. All else being equal, it should be considered (prima facie) reason to think the recidivist is capable but unwilling to control his behavior. This is the simplest explanation of a person’s crimes and related propensities and the only explanation that does not require proof once guilt has been established. What requires proof is any explanation involving inability (or unusually limited ability) to obey the law. The state must prove that the specific features of an offender’s sexual pattern cannot be explained by appeal to ordinary criminal volition coupled with sexual motivation (deviant or not); that the only acceptable explanation attributes the person’s past behavior and current proclivities to a kind of serious disturbance affecting volition-related processes. The same reasoning should be used by the expert whose evaluation is the basis of an SVP case.
Post-Crane, SVP evaluators usually do add statutory language specifically addressing “serious difficulty.” But rarely do their reports contain argumentation for this opinion beyond the evidence already given for (DSM defined clinical) mental disorder and mostly actuarial-based risk classification. As noted by Mercado, Schopp, & Bornstein (2005), when issues of control capacity are discussed at all by clinicians, there is much overlap, if not outright conflation with typical qualities of criminal activity, such as impulsivity, risk-taking, short-sightedness, and low self-monitoring – in other words, all the qualities we associate with the “typical” recidivist. Some evaluators argue for “serious difficulty” by coupling high actuarial percentages with the additional fact (if it is a fact) that the particular offender sexually recidivated within a few weeks or months of most recent release. This kind of reasoning is still superficial. The mere fact of recidivism within weeks/months of release does nothing to distinguish the volitionally (or otherwise) impaired from the merely reckless. Evaluators who emphasize time-to-recidivism (unless, perhaps, the offending occurs within hours or days of release) would seem to be assuming that impaired offenders are somehow capable of resisting “overpowering” urges for a rather long time (weeks or even months). Or perhaps offenders prone to overwhelming urges manage to never actually have them until shortly before the crime.12 In any event, these are assumptions not supported by any body of research of which this writer is aware.13

More Logic

Philosophers distinguish “order of knowing” from “order of being.” The first is usually the reverse of the second. The order of being or existence is the direction of causal or material relationship among entities or phenomena in the world. Simplistically put, cause comes first, then effect. In an SVP context, mental disorder comes first and causes the general effect of impairment to volition-related mental processes related to conforming one’s actions to the law. This in turn results in a more specific effect, namely, an extremely dangerous tendency. Order of knowing is the direction of our reasoning in coming to understand causal relationships in the world, and also the necessary direction of demonstrating or proving to someone (e.g., to a fact finder) that the world should be understood in this way. For knowledge, and proof, we reason backward from effects to causes.
Logically, then, the state, and its evaluator, must argue from effect back to cause. The fact finder must first be presented with strong evidence of the effect (impairment) in order to have good reason to believe there is an underlying cause (incapacitating mental disorder). So the state must show convincing evidence of “serious difficulty” before it can prove “mental disorder.” Not the other way around. And it cannot presume that proof of a DSM defined mental disorder coupled with proof of high likelihood of recidivism proves that the former “makes” the latter true (via impairment). The state must prove that someone’s past sexual history and current state of mind (the effect) shows the existence of serious impairment (cause), which can only be a mental disorder (given that the average person, and typical offender, is not impaired), and that this impairment, coupled with freedom of movement around potential victims, is clearly dangerous. Only in this way does the state really show that the offender is “made” or caused to be dangerous by his impairment, i.e., mental disorder, as opposed to his risk coming only from an enduring liking for sexual offending and attitudes dismissive of law and the rights of other people. Such a person is truly someone the state can describe as dangerous “beyond control” rather than simply dangerous.
Unfortunately, typical SVP reasoning confuses order of knowing with order of being and tries to argue from cause to effect. Arguing from an alleged cause to an alleged effect fails to prove the cause being alleged, and thus the existence of the effect. It is not uncommon for evaluators to simply look at the criminal sexual record and diagnose a DSM-defined paraphilia. Doing this drains the diagnosis of any explanatory function because it has become an alternative label for that record, or a label for more than one deviant urge (given the fact of more than one deviant act). The evaluator uses the same evidence – the same record – coupled with a bit of demographic information (age; enduring live-in relationships or not) to classify the offender into a group from which previously studied samples (or just one sample) showed a substantial sexual reconviction or rearrest rate when tracked over a long period following release. Then the two – DSM paraphilia and statistical classification – are said to be causally “linked,” as if the one being true and the other being true establishes that the first caused the second. All this comes from the criminal record, and it proves nothing. In fact, all that has been shown is: (1) someone with more than one similar sex offense likes a deviant kind of sex, (2) a substantial percentage (maybe over half) of persons liking deviant sex and willing to indulge in it despite threat (or even experience) of prison, can be expected to have more deviant sex. These claims cannot establish impairment (cause) or “out of control” dangerousness (effect) because they can fit both the SVP and the TSR offender.14 So the evaluator has told the court nothing beyond the obvious, and the state is left with no case.

The Argument for SVP Status

Few if any SVP evaluations and SVP court cases reason in the correct logical direction, if the point is to show that someone is highly dangerous because of impairment to his ability to control sexual behavior, secondary to mental disorder. Most evaluators and county attorneys presenting SVP cases do little more than the following by way of argument: (1) Offender X has a DSM-construed paraphilia (based on at least two similar sex offenses). (2) Offender X is “likely” to continue sexual violence (which is equated with the claim that X falls into a high-risk category according to one or more actuarial instruments). Therefore, (3) Offender X is likely to do sexual violence because of mental disorder. Therefore, (4) Offender X is likely to do sexual violence because he has “serious difficulty” controlling behavior.
The Crane decision allows this direction of argument, but only because it makes the assumption that some clinical mental disorders (paraphilias) include “serious difficulty” as an intrinsic feature. According to the U. S. Supreme Court (and the Arizona Supreme Court in Leon G.) the state need only “link” (connect causally) likely sexual violence to mental disorder and a jury will necessarily “infer” the presence of “serious difficulty” and attribute the person’s dangerousness to volitional/emotional impairment. Although the state is not required to make a distinct argument for “serious difficulty”, the judge or jury must make a specific determination as to its presence and connection with risk. But as discussed, without first showing the presence of impairment, there is no way to prove someone’s dangerousness is “caused” by a mental disorder rather than typical volitional processes for dangerous criminals. Arguing in the direction outlined above would seem to violate parsimony rather brazenly. Is the Court not requiring parsimony after all?
Parsimonious reasoning is still required. The Court requires SVP offenders to be distinguished from typical recidivists and it requires that SVP status create a narrow class. The Court preserves its parsimony principle by introducing an assumption about those clinical disorders most relevant to SVP “mental disorder.” The effect of this assumption is to obliterate the distinction between atypical and typical dangerous offenders, thus removing the only element that keeps SVP “narrow.” At bottom, the Court allows the state to assume the key fact to be found, i.e., the presence of “serious difficulty”, whenever a relevant DSM clinical diagnosis has been established. In other words, SVP-relevant mental disorders are presumed to be inherently incapacitating at an SVP-relevant level of legal significance. The Court’s assumption about paraphilias appears in its description of Hendricks: “an individual suffering from pedophilia, a mental disorder that critically involves what a lay person might describe as a lack of control.” The behavior of sex offenders is then described as “compulsive, repetitive, driven (which) appears to fit the criteria of an emotional or psychiatric illness…(i)t is often appropriate to say of such individuals, in ordinary English, that they are unable to control their dangerousness.” Support for this claim is one article from a now dated sex offender treatment handbook.15 This would seem to be remarkably shallow reasoning were it not for the likely fact the Court does not consider its assumption to be controversial. In one sense they are probably right. The belief that paraphilias are intrinsically compulsive, driven, or “out of control” forms of sexuality is a stereotype dating back to the 19th Century (Jenkins, 1998). Laypersons might indeed regard paraphilias as involving little ability to control sexual impulses. Then again, the public is also likely to assume that virtually all sex offenders are highly dangerous and that anyone who has engaged in a sex offense has a disorder of sexual deviance. If popular opinion is our standard, mental health expertise is not needed.
Given this assumption, the Court in effect allows the state (and the SVP evaluator) is permitted to proceed like this:
(1)X has two or more similar sex offenses. Therefore:
(2)X is enduringly attracted to (has urges toward, is sexually aroused by) a deviant form of sexual activity or object, and has acted accordingly with a non-consenting person. Call this Paraphilia1.
(3)If Paraphilia1, then Paraphilia2.
Call Paraphilia2 an enduring and victimizing pattern of sexual preoccupation with a deviant object/activity, and/or overwhelming (“overpowering”), driven, compulsive, or grossly impulsive sexual urges or behaviors toward a deviant sexual object/activity. Therefore:
(4)X has a mental disorder compromising voluntary control over sexual behavior
(5)X is involuntarily dangerous (or less than typically voluntary for recidivists) to whatever extent he is dangerous (which is argued separately and usually involves actuarial classification).
The soundness of this argument depends on premise (3). This is the only thing allowing the state to reason from presence of (DSM defined) mental disorder to “serious difficulty,” or from the twin facts of mental disorder and dangerousness to the causal connection of (or “link” between) mental disorder to dangerousness. If (3) is false, the “link” is absent and the argument fails.
Without assuming (3), the state must argue in reverse order if it is reason parsimoniously. It must actually show “serious difficulty” in order to show dangerousness “beyond control,” and also to show the presence of SVP-relevant (volitionally impairing) mental disorder, which allows the state to show non-voluntary dangerousness because of mental disorder (or dangerousness because of impairing mental condition rather than choice). The argument would proceed like this:
(1)X has a sexual history (including but not restricted to his documented sexual offenses) that is unusual for (even) sexual recidivists.
(2)X’s sexual behavior and/or current mental condition show signs of unusual driven-ness, obsessive/compulsiveness, or gross impulsivity when compared to other sex offender patterns.
(1) and (2) are interrelated and provide the basis for:
(3)X’s sexual behavior and/or current sexual mental state is “beyond control” (i.e., “out of control,” less-than-typically voluntary or volitional, even when compared to other sex offenders).
Then the issue of mental disorder:
(4)Given the type of victim in X’s offenses and/or type of sexual activity, X has a paraphilia according to DSM clinical criteria.
(3) and (4) provide the basis for:
(5)Not only does X have a paraphilia, but he has a paraphilic condition so severe that it impairs his ability to control sexual behavior rather than simply motivating him to engage in sexual crime. Therefore:
(6)X has an SVP “mental disorder.”
Now the state must “link” SVP-mental disorder to extreme dangerousness. First, establishing dangerousness:
(7)If released unconditionally X will have access to potential victims. [Barring special circumstances, e.g., poor health.] Therefore:
(8)Situations of high imminent risk are expected to occur for X (sooner or later).
(5) and (8) are true; therefore:
(9)Any rational person would have the confident expectation that X would engage in more sexual offending if released unconditionally. [Whether or not this will actually happen is unknowable. But using a qualitative understanding of “probable”, i.e., reasonable to expect; not reasonable to expect otherwise, it is probable that X will commit acts of sexual violence if released.]
Now the “link”: (5) and (9); therefore:
(10)X has a mental disorder that both motivates sexual offending and impairs mental processes related to voluntary self-control, to the point where additional sexual offenses are only to be expected from X, eventually if not imminently (even if he wanted to be lawful). Therefore:
(11)X has a mental disorder that “makes” or causes him to be extremely dangerous, as opposed to his dangerousness coming from the simple interaction of sexual desire and criminal volition.


Implicit in the Court’s assumption about paraphilias is the notion that deviant attractions are not only abnormal in their object of sexual interest but also in their strength of interest or drive toward that object. Certainly popular stereotypes of sexual “deviants” portray such persons as having both illicit sexual tastes and unusually insatiable sexual drive, as if abnormality of sexual interest somehow generates abnormality in sexual drive. [Or some might suggest it is the other way around: If someone’s sexual interest or drive becomes too strong it will expand into deviant directions; see Carnes, 1983.] Either way, the assumption seems to be that people who act on deviant urges more than once somehow possess a craving for deviant sex that has gotten to the point where normal decision-making and self-regulatory processes are seriously undermined and unreliable.
By way of contrast, it is easy to imagine a non-deviant man who gets involved in affairs and is likely to continue doing so, much to the detriment of his life and the lives of other people. Presumably, family courts and custody evaluators would not automatically assume that this man suffers from (almost) uncontrollable urges or an unusual lack of ability to engage in practical reasoning, such that he would have extraordinary trouble avoiding affairs if he wanted and tried to do so. He would not be diagnosed with Infidelity Disorder, or civilly committed to protect marriages. Of course, affairs are not illegal. But the forensic significance of a sexual pattern is not relevant to its empirical dynamics and it is a baseless assumption about these dynamics that is driving the forensic (civil) significance of deviant sexual patterns. In some cases, a person having multiple affairs might actually be impaired volitionally (he or she might be hypersexual; an example is found in DSM-IV under Sexual Disorder NOS). However, few clinicians or courts would simply assume incapacitation in the serial adulterer, just because that person is a serial adulterer.
However popular it might be, the assumption that paraphilias are intrinsically impairing to sexual self-control lacks scientific basis. I am unaware of any significant body of research supporting any of the following claims: (1) All or most sex offenders are, compared to other criminals or men in general, abnormally driven, compelled, dependent (“addicted”) or grossly impulsive in their sexual behavior; (2) All or most sex offenders who are highly dangerous are abnormally driven, compelled, dependent, or grossly impulsive in their sexual behavior; (3) Paraphilic persons are all or mostly sexually driven, compulsive, dependent, or grossly impulsive to an abnormal and clinically severe degree. This writer’s experience in sex offender civil commitment centers offers examples of both patterns discussed in an earlier section of this paper (SVP vs. TSR prototypes). Even among the highest risk offenders, there are numerous men with paraphilic interests and arousal, but only normal range or even below normal range frequency and felt-urgency of sexual thoughts and impulses. There are also men with true paraphilic obsessions and “overwhelming” if not literally uncontrollable deviant urges.
Rightly, the authors of DSM warn explicitly against drawing legal/normative inferences from any diagnosis. The authors of DSM-IV (p. xxiii) and DSM-IV-TR (p. xxxiii) are clear in stating that diagnostic criteria imply nothing intrinsically about whether or not someone’s impairments are sufficiently severe to meet any particular set of legal criteria related to capacity to regulate behavior. Even for those disorders in which some kind of impaired self-control is typically present, too much variation can be found among clinical presentations of a given disorder to presume anything about an individual’s ability to control behavior of legal significance. Whatever the failings of SVP evaluations done for the defense (a subject worthy of its own paper), these evaluators correctly emphasize the caution recommended in DSM. On the state’s side, DSM warnings fall on deaf ears.
The issue here is more than scant data. Empirical investigations presuppose prior definitions. Definitions presuppose foundational questions: Should the term ‘paraphilia’ be restricted to only those patterns with impulses toward deviant objects fairly described as “overwhelming” or “out of control,” or should any sexual pattern carry the label provided it has a deviant object and the person has acted accordingly with a non-consenting person? Some experts (e.g., Kafka, 1991) view paraphilia as akin to non-paraphilic hypersexuality. It is not clear that all or most experts agree.
If experts in sexual disorders widely accepted the Court’s stereotype about paraphilias, we would expect DSM diagnostic criteria for paraphilias to resemble those for Obsessive Compulsive Disorder or Substance Dependence, or at least one of the impulse control disorders, e.g., pathological gambling. Criteria for all these latter disorders specify multiple specific symptoms strongly suggestive of severely limited ability to control relevant thoughts and behaviors (albeit to no particular degree of severity), such as unwanted preoccupations and driven behaviors, escalating involvement in a behavior in order to achieve a pleasurable effect, and repeated unsuccessful efforts to stop a behavior.
Clinical DSM criteria for paraphilias are nothing like this. The phrase, “recurrent, intense fantasies or urges,” is mildly suggestive of abnormal amount and strength of sexual ideation and drive. But nothing specific follows about one’s ability to resist sexual urges or capacity for reasoned decision-making with regard to sex. A fantasy is “recurrent” if it happens more than once. Non-deviant fantasies are recurrent as well. How many sexual patterns of any kind in males who are physically healthy and not depressed, involve only isolated or rare thoughts/impulses? The nature of “intense” fantasies or urges is anyone’s guess. Non-deviant fantasies and urges are often vivid and powerfully arousing, as well as frequent (in young males at least). Even if “recurrent, intense” is meant to convey abnormality in degree, and not just kind, of sexual impulse, a criminal record with a few (or even several) documented sex crimes cannot justify any specific opinion about the degree of psychological pressure there was behind these behaviors or how much it interfered with the person’s contemplation of options and potential consequences. If the offender does not report recurrent intense fantasies/urges then the record alone seldom supports the inference that they exist, or ever existed.

Personality Disorder

Personality disorder is typically included on the list of relevant mental disorders in SVP statutes. The presumption seems to be that if paraphilia is not enough to make someone highly dangerous beyond control, then paraphilia mixed with personality disorder does. In states that specify personality disorder as an SVP disorder, an offender can possibly satisfy criteria for civil commitment without paraphilia. The most frequently cited personality disorder is Antisocial Personality Disorder (APD), psychopathy in particular. Other personality disorders are also seen.
In fact, the inclusion of personality disorder only weakens the state’s case. An SVP case resting on personality disorder flies in the face of Foucha and borders on sophistry. In a nutshell, telling a jury that a sex offender has a personality disorder is little more than announcing that the person has plenty of reasons to continue deliberately and voluntarily making criminal decisions, whatever his motivations. Several considerations are important to keep in mind when thinking about personality disorders and SVP statutes.
By the logic of Crane, the state must establish that an offender is not a typical recidivist. SVP-disorders, defined as disposing someone to high sexual risk by virtue of impairment, must delineate a narrow category of criminal recidivist. If personality disorder has any logical or legal significance in determining SVP status, then it must help discriminate the dangerously impaired (atypical recidivist) from the volitionally dangerous (typical recidivist). Are we to assume that TR and TSR offenders lack personality disorder? The Supreme Court acknowledges that a high percentage of inmates meet criteria for Antisocial Personality Disorder (APD), so APD alone cannot distinguish a small subset of recidivist. Coupling APD with sex offenses (or high actuarial scores) does nothing, as multiple sex offenses will earn a paraphilia diagnosis from most if not all evaluators, with or without APD. Even if personality disorder were not allowed to stand alone as an SVP disorder, it is far from clear how the coupling of a widely shared characteristic among offenders, such as APD, with one less widely shared, such as paraphilia, is capable of delineating a narrow class. Why would the less common characteristic not be enough by itself (provided there is the relevant impairment)? Moreover, how is it that an SVP offender with supposedly little control over criminal sexual impulses can be said to have severe personality disturbance, but not the TR or TSR offender who persists in choosing sexual crime when he could make decisions that are both legally safer and less severely harmful to others? One would think it is the second offender who has the more inflexible and maladaptive set of attitudes and habits overall. At least in theory, the SVP offender might not want to act as he does. Qua impaired, he “cannot help” himself. In contrast, the voluntary offender does as he pleases. One would think that “serious difficulty” has something to do with an unusually severe “weakness of will,” not intractable willfulness. The latter would seem to be as common to personality disorders (Cluster B at least) as the former, if not more so.
It is important to keep in mind that DSM-defined personality disorder (PD) diagnoses are simply labels for seriously maladaptive patterns of thought, feeling, and behaviors. A pattern of voluntary deliberate behavior can be just as inflexible and maladaptive as non-voluntary reactive pattern. Nothing about the concept of a PD as such implies abnormally limited capacity for exercising volition. Quite the contrary. Unlike Axis I disorders, PDs involve ego-syntonic patterns, giving the individual ample reason to choose the behavior he does. That said, it is true that personality disordered individuals are notorious for their seeming inability to learn from the recurring bad consequences of their choices (see Doren, 2002). Is this not a genuine impairment?
Yes, in a clinical context, but not in any clear legal/normative context. The issue in SVP cases is whether an impairment of this kind – if indeed this is the actual nature of personality-related impairment – should carry special legal significance in isolating a narrow class of dangerous offender. There is little reason to think so. Individuals with PDs do not, just as such, suffer from the learning impairments of retarded or psychotic persons, or even those of individuals with specific learning disorders. In many areas these individuals learn as well as anyone else with similar IQ. Not surprisingly those areas where these individuals seem not to learn – acting as if they cannot learn – are areas closely connected to self-concept, perceptions of others, and attachment. Rather than some experiential learning deficit, the real problem would seem to be schema-based attitudes and severe emotional immaturity. Clinically, these are genuine problems and very difficult for the person to change. However, normatively and legally, they amount to no more than flawed character. Character defect does not lessen moral agency in any other area of law (responsibility), nor does it prevent someone from understanding the rudiments of law and society’s response to law-breaking, and incorporating this knowledge into decision-making (necessary conditions of deterrence). Difficulty changing one’s character – something anyone would have – does not equate with unusually great difficulty obeying the law when compared to the typical chronic criminal.
Finally, personality organization would have to be quite severely disturbed to render someone criminally dangerous “beyond control,” and then only at certain times and in certain circumstances. One can imagine someone with severe Borderline Personality Disorder (BPD) in the throes of rage or despair, or perhaps a schizotypal patient on the verge of psychosis. But such cases are not at all analogous to the narcissistic molester believing he is qualified to “teach love” to a child, or the antisocial rapist who thinks women “ask for it.” However reckless and impulsive, criminal behavior is rarely like the self-cutting or suicidal behavior often seen in inpatient BPD. Criminals frequently experience excitement and pleasure in their crimes, as well as ego-enhancement. And their behavior is often deliberate and pre-planned. These are motives predisposing criminal volition and thus factors that the state must explain away in arguing for the presence of a true volition-related impairment.
A much more convincing case for “serious difficulty” controlling sexual behavior would be an offender the state can show to be lacking in personality-related disposition to feel above rules and entitled to have anything he finds attractive. The sexual behavior of such an offender would then appear anomalous compared to his normal behavior, suggesting a source other than the person’s normal decision-making processes that genuinely interferes with even serious attempts to weigh the pros and cons of a desired behavior before acting.
None of this implies that individuals with a PD never have important impairments in sexual self-regulation capacity and decision-making ability. The point is only that one cannot argue that such impairment is the reason for the individual’s sexual dangerousness until it is first shown that such propensities are not better accounted for by the simpler explanation, namely, the offender is prone to sexual crime because of his ego-syntonic patterns of voluntarily choosing behavior, which tend to be inflexible and maladaptive.


The argument might be made that only genuine psychopathy should be considered to meet SVP standards of PD. APD includes (most) psychopaths in forensic settings but is not specific to them. When people think of sexually violent “predators,” they frequently have in mind someone cruel (sadistic) and remorseless (psychopathic). Research does suggest some degree of physiological difference between the brains of psychopathic criminals and those of non-psychopathic criminals, and non-psychopaths generally. However, perhaps unfortunately for county attorneys bringing SVP cases (and their evaluators), these findings do little to resolve the issue of how much and exactly what kind of impairment is found in psychopathy, much less whether it has implications for such legal and normative notions of incapacitation as “serious difficulty.” This is not the place for an exhaustive review of specific studies. An interested reader might start with R. D. Hare’s 2003 manual for the PCL-R, which nicely summarizes this research. A general overview will suffice:
One body of laboratory research supports the “response modulation hypothesis”, which holds that psychopaths fail to accommodate important environmental cues while engaging in goal-directed behavior (Gorenstein & Newman, 1980; Kosson & Harpur, 1997; Newman, 1997, 1998; Newman et al., 2003). According to this hypothesis, the psychopath’s perseverative and dishibitory behavior comes from his failure to attend to new information in a changing environment. Other researchers suggest that psychopathic behavior comes from an overly strong behavioral activation system, a hypothesis apparently better supported than the idea that psychopaths have a weak behavior inhibition system. In this view, psychopathic behavior is more the result of hypersensitivity to reward than hyposensitivity to punishment (Arnett et al., 1997). Some authors (Newman et al., 2002) argue that psychopath is not so much related to general deficits in cognitive or emotional processing but to selective attention to the primary demands of a situation, with limited processing of incidental cues that would provide other people with perspective on behavior. [Interestingly, confirming data come only from “low anxiety” psychopaths. Apparently, “high anxiety” psychopaths do not fit the model.]
From this research we might say that many (not all) psychopaths behave as if they are heavy trucks barreling down the highway. Given its momentum, a truck is harder to stop than other vehicles (hypersensitivity to reward). Mirrors and brakes are functional (inhibition systems are not impaired as such) but the driver is too focused on what is directly ahead to see any need to use them (tunnel vision).
But consider logical parsimony in actual civil (or criminal) cases. The above-mentioned research raises too many questions to offer clear support for the idea of volitional or “emotional” impairment in psychopaths, much less to a threshold level of normative or legal significance. First of all, is the behavior of psychopaths the structural product of physiological mechanisms directing attention and goal-pursuit, or more a matter of character traits? Personality and character are obviously affected by brain physiology, but parsimonious explanations of goal-directed human behaviors (such as crime) do not typically require resort to physiology. Psychopaths might indeed have unusual brain structure. But for purposes of court proceedings, what features of psychopathic criminal behavior are not well accounted for by a personality or character based account, such that we must introduce descriptions of neurophysiology? After all, arrogant and egocentric drivers who feel they own the road are apt to drive dangerously, and for all the same kinds of reasons as those noted above.
Moreover, while the physiological models above might explain a date rape by a psychopath – failure to adjust one’s goal (sex) to changing circumstances (the victim says no) – they have questionable relevance to spree or serial rapes. The latter are less opportunistic and impulsive and more premeditated and predatory. For the paraphilic rapist, there is no stumbling into violent sex after failing to process signals from a soon-to-be victim. Forced sex is his goal. Perhaps psychopaths have above-average expectations (for criminals) about getting what they want, and they might pursue their ends more vigorously than others. But this is not the same thing as being compulsively driven or “overwhelmed” with urges.
Cutting the opposite direction, a second body of research, focused on criminal behavior, suggests that psychopaths might actually possess greater control than other criminals. According to Woodworth and Porter (2002), psychopaths engage in more instrumental and “cold-blooded” murder than other killers, whose violence tends to be reactive and connected to high emotional dsyregulation (“crimes of passion”). They argue that the impulsivity of psychopaths possibly owes less to limited capacity for control and more to conscious decision-making after rapid consideration of consequences. This seems consistent with research suggesting that the violence of psychopaths often occurs as a “macho” display of power or vengeance. According to Walsh (1999), psychopaths experience increased self-esteem after violence, as evidenced by their reported feelings of excitement, power, and justification. In contrast, non-psychopaths report anxiety, guilt, and fear; emotions we would expect to follow non-deliberate violence by someone who “lost control”. Although ‘poor behavioral controls’ is an item on the PCL-R, the scale’s developer, R. D. Hare, does not believe that psychopaths are acting uncontrollably. According to Hare (1993) psychopaths know what they are doing and do not lose control just anywhere or at any time. Given the arrogance and supreme self-confidence often displayed by psychopaths, it would seem very difficult to argue that their dangerousness comes from incapacity rather than the natural choices of a vastly inflated ego unencumbered by conscience. Lack of conscience/empathy would not constitute SVP-relevant “emotional” impairment unless there was a basis for thinking that typical recidivists lack this impairment. No such basis is apparent. Even if we assumed that only SVPs are psychopathic, non-psychopaths might also feel little if any remorse or concern for others.
Finally, even if research on (small) samples of psychopaths revealed genuine physiological impairments in those individuals, an evaluator would be making a large leap in logic if he or she assumed that anyone scoring high enough on the PCL-R will be likewise physiologically unusual or organically impaired. A defense attorney will surely ask the obvious question, “Well, doctor, did you specifically assess for these alleged brain abnormalities in this case?” Of course, the answer will be no. At most, the expert can say only that if the psychopath is volitionally or emotionally impaired in a legally relevant manner, the nature of this impairment is apt to be of the sort researchers are now investigating. This is akin to saying that if paraphilias constitute a volition-related impairment, then the impairment would be one of “overwhelming” urges. Of course, the issue before a court in an SVP case is not the possible truth of a hypothetical statement about sex offenders, but the actual nature of an individual’s condition.

Evaluators and Volitional Impairment

After a detailed review of the legal and empirical literature related to “volitional impairment,” Mercado, Schopp, and Bornstein (2005) conclude that there is too little scientific evidence available to arrive at general opinions about volition and volitional incapacity based on anything more than theory. [In Hendricks the Court simply took the word of Hendricks that he could not control his deviant urges.] The authors recommend that clinicians provide the court with “descriptive explanations” of symptom patterns and offender characteristics, but not information or opinions meant to be directly relevant to volitional capacity. Otherwise, they overstep their science and professional ethics. It is up to the court to draw meaning from this testimony for an inherently ambiguous legal construct. Miller et al. (2005) and Zander (2005) would seem to agree.
This point is well taken. Unfortunately, it is not realistic. SVP courts (at least in Arizona) want more than just descriptive explanations of symptoms, behaviors, and clinical disorders. Evaluators are often, if not always, called upon to give explicit opinions about all three SVP constructs – “mental disorder,” “likely,” “serious difficulty” – as these notions relate to individual cases. In other words, the expert is called upon to form clinical opinions about mental status, impairment, and risk, and then measure the degree or severity of these characteristics against legal standards. The final result is an opinion about SVP status, which is extra-clinical and legal/normative in nature.
Another possible approach is for SVP evaluators to offer an interpretation of “serious difficulty” (as well as “mental disorder” and “likely”) in light of the spirit of their state statutes and Supreme Court opinions, and then marshal evidence as to the relevance of this interpretation (or lack thereof) to a particular offender’s case. The expert’s report should spell out his or her working interpretation of legal concepts (which need not be “bright line” criteria), as well as explaining how this interpretation relates to the empirical and clinical information and opinions presented in the psychological evaluation itself. Ideally, the expert’s report will contain a psychosexual evaluation in which clinical evidence is explained and interpreted to support clinical opinions, followed by a section on legal interpretation of relevant SVP concepts applied to these opinions. In this section, clinical assessments are explicitly measured against an articulated interpretation of legal standards to produce a legal/normative opinion about SVP status.
Of course, the evaluator should provide all the appropriate caveats. The most important caveat is that neither the legislature nor the courts have provided specific criteria for determining how SVP concepts apply to individual cases. A second important caveat is the lack of specifically volition-related psychosexual research and related measures of volition-related sexual impairment. A third caveat is that “beyond a reasonable doubt” (the legal burden of proof in some SVP statutes, e.g., Arizona) is a legal standard applicable to the state’s burden in making an SVP case and the fact finder’s own assessment of that case as a whole. It is not a clinical or professional standard of “proof” and would preclude almost any conclusion being drawn if used by a clinician to interpret each and every piece of clinical evidence in the formation of a clinical opinion.
Ethically, evaluators would also seem obligated to inform the court whether their basic assumptions are contrary to those grounding the constitutionality of SVP statutes according to Crane. This is more specific than the expert’s opinion about the wisdom of these kinds of laws. The issue is the expert’s assumptions about sex offenders and sexual disorders and whether these beliefs confound the Court’s principle of parsimony. If clinicians believe that all sexual recidivists, or all paraphilic sex offenders, or all highly dangerous sex offenders are “volitionally” impaired just as such, or that such impairment (to the degree required by SVP statute) is in some way inherent to paraphilia or personality disorder (consistent with SVP concepts in some states but contrary to DSM warnings) then they should make this clear up front. Clinicians should be clear when they do not acknowledge any meaningful empirical or clinical differences between SVP sexual recidivists and “typical” sexual recidivists, or if they only distinguish SVP from non-sexual criminal recidivists without sex offenses or recidivists only one sex offense (i.e., if they believe there are “typical recidivists,” but no “typical sexual recidivists,” in contrast to SVP recidivists). It is then up to the court to determine whether the evaluator’s findings are more probative than prejudicial. In this writer’s opinion, if right at the outset evaluators cannot make the key distinction upon which these laws turn, it is difficult to see how their work will be anything but prejudicial.

Serious Difficulty Controlling Behavior

While Crane does not define SVP impairment or specify threshold criteria for “serious difficulty” or “volitional impairment,” mental health professionals are not doomed to utter agnosticism about these ideas. According to the Oxford American Dictionary (2002), ‘volition’ is defined as the exercise of will, or power of willing. Someone acts volitionally if they act voluntarily. Someone’s behavior is “voluntary” if that person acts intentionally, or according to his or her “own free will.” It would seem to follow that a person is impaired volitionally if they possessed deficient “willpower” capacity to exert one’s will to resist acting on impulses. This is not deficient exercise of will, as in a lack of effort and persistence. Many criminals fail to restrain criminal urges because they have never really tried. Deficient willpower is deficient ability to exert one’s will and persist to the point of effectiveness, even if a person wants and tries to do so. The person who is impaired in this way, to the point of legal significance, is not someone who merely failed to make lawful decisions. But neither is it true that he is someone who necessarily wanted to do right. Some volitionally impaired persons try and fail to avoid bad behavior; others like bad behavior and will never try a different path. The issue is whether an individual would be as capable as other criminals of avoiding crime if he or she tried to be lawful. If yes, then this individual is criminally volitional. He may or may not always remain so. If no, then the person is volitionally impaired, even if he resembles other criminal recidivists in his attitudes and values. These offenders will likely never try to stop their behavior. They are still impaired, however, because were they to try, they would likely fail.
Of course, when there is no evidence to suggest that an offender has ever tried to control his urges, it is difficult to know how to assess what difficulty he could or would have if he ever tried self-restraint. Whatever the evidence, parsimony requires that we regard any given offender as a typical recidivist unless compelling reason exists to think otherwise. In other words, without strong evidence to the contrary, we should think about any given dangerous recidivist that he is an individual who, if he were to try to avoid crime, would have no greater difficulty than any other dangerous offender.
So what kind of evidence might be to the contrary? How can the expert ever know what someone would have trouble doing (or not doing) were he to try? One clue is the existence of a sexual history so extensive that it indicates preoccupation and unusual frequency and urgency of sexual urges. Even among sexual recidivists, and those in civil commitment centers, there is wide variability in extent of sexual background. Whether or not the behavior is ego-syntonic, sexual ideation and behavior in the “driven” sex offender takes up so much time and energy in his life that it is not reasonable to think that change of attitude and attempts at control would stop the behavior. So even if the offender will not talk about his inner states to the evaluator, or is not reliable in his reports, the amount and frequency of sexual activity in his background can, depending on their degree, be strong evidence of impaired ability rather than low willingness to control oneself. This is true even when there is little evidence the offender ever tried to control sexual behavior. Some offenders are impaired and unwilling, and will never realize their limitations until they try control. [Of course, as noted, in many cases there is simply too little reliable information about someone’s full sexual history, or not enough sexual behaviors in that history, to clearly indicate an impaired pattern without the offender self-reporting relevant sexual mental states, past and present. If the offender refuses, or is unreliable, then SVP status cannot be determined to any reasonable degree of certainty.]
An analogy can be drawn with smoking. It seems relatively straightforward to assess someone’s ability to control smoking behavior when that person is believable in her expressed disgust for her behavior and credible in reporting many serious but futile attempts to stop. Much more challenging (and more representative of sex offenders) is the ego-syntonic smoker who has never tried to stop. Is there any way to tell how much trouble this person would have if she ever tried to stop? For some of these smokers, there is simply no way to tell if they would have any greater difficulty than the typical smoker. They do not smoke any more than the average smoker and many smokers can stop if they are serious about doing so. Then there are smokers who smoke a great deal even for smokers. We have no evidence of attempts or even a desire to stop, but we do have evidence about the lengths these individuals will go to smoke, the time they spend smoking, the frustration or discomfort they feel when they cannot obtain tobacco (or cannot leave a non-smoking area), and perhaps the fact they no longer feel pleasure in smoking. All this suggests much more than a bad habit. Any reasonable observer would expect that these individuals would fail in any attempt to stop (even if the future happens to be otherwise) and that these smokers would need treatment to have any realistic chance of success. These individuals are actually incapacitated, not just “weak.” [Weakness of will is not impaired willpower ability.]
Evaluators should have some sense of the ballpark here. The Court has at least fenced in the field of “serious difficulty”. At one extreme are persons who would probably qualify for the now discarded “irresistible impulse” prong of legal insanity, e.g., individuals with dementia or frontal lobe damage, mental retardation (more than mild), psychosis, mania, or especially severe obsessive compulsive disorder or extreme physiological dependence on substances. These persons have no meaningful control whatsoever over at least some of their behavior, in the sense of being able to avoid behaviors they know to be illegal. This highest level of impairment would certainly qualify as “serious difficulty,” if anything would. But most persons with “serious difficulty” will not be this extremely impaired. At the other extreme are the “typical” serial criminals, often with have long histories of impulsivity, recklessness, lack of future orientation, and little concern for law, other people, or consequences. However difficult it is for these persons to change their behavior, or their character, they are capable of responding to deterrence-related incentives and do not warrant special legal standing. Most persons with “serious difficulty” controlling their behavior fall between the extremes of irresistible sexual impulse and reckless sexual behavior.
Mental health professionals have empirical/clinical models, or analogous clinical syndromes, for a kind of impaired self-control falling in between complete helplessness in the face of impulse and typical criminal impulsivity and disregard for consequences. These models should be utilized in thinking about sexual patterns. One possibility is chemical dependency (of a psychological if not physiological kind). Another is binge eating disorder. Another would be pathological gambling (noted in DSM). Full-blown obsessive-compulsive disorder (e.g., found in inpatient settings) might be too severe for this middle category, but the less severe forms of OC behaviors might well be examples. Finally, there is the phenomenon of “sex addiction” or “compulsive sexual behavior,” which sexual disorders experts have conceptualized as analogous to (if not an actual species of) these other disorders. All of these involve, intrinsically, some degree of genuine disability in behavioral control, not just impulsivity or refusal to exercise control. This is not to say that if an offender has a sexual behavioral pattern akin to these disorders that he automatically satisfies the “serious difficulty” standard of SVP law (recall DSM’s admonition). But he will have a pattern that renders this issue a serious empirical question, as well as a legal/normative issue, for the ultimate finder of fact.


Given a plausible working, if not precise, understanding of “serious difficulty,” it is hard to imagine how a sex offender who lacks retardation or severe mental illness (the vast majority) can have genuine emotional/volitional impairments meeting SVP standards without having some form of hypersexuality. Mercado, et al. (2005) correctly note that there is little psychological knowledge, consensus, or assessment methodology exists from which an evaluator can evaluate volitional impairment per se. But the same is not necessarily true, or not as true, about hypersexuality, which, unlike paraphilia, is a clinical construct involving impaired control capacity as an inherent feature, albeit to no particular legal standard of severity. Little research has been conducted on hypersexuality and there are to this writer’s knowledge no validated measures to date. However, there is a substantial body of conceptual work as well as detailed case studies, enough to ground some systematic assessment procedures. At the very least, these procedures will far exceed the current standard of assessing this condition found in most SVP evaluations (often nothing at all).
Not yet specifically recognized in DSM, hypersexuality will probably get its own diagnostic label at some point. [At the present time, when no paraphilia diagnosis assigned, sexual disorders experts use Sexual Disorder NOS for coding.] Whether we should define paraphilias as deviant forms of hypersexuality – not done in DSM-IV or DSM-IV-TR – or more broadly such that there can be hypersexual and non-hypersexual variants of a paraphilia, an evaluator will most often have to find explicit evidence of hypersexuality in order to make the case that an offender is actually impaired in his sexual control, paraphilia or not. As noted, exceptions might exist in the presence of other debilitating conditions, such as when there is gross sexual impulsivity but normal range frequency of sexual ideation or behavior. Some retarded persons, or individuals with bipolar or psychotic illness, might fall into this category. But many SVP statutes specify a very narrow list of mental disorders that can count as SVP-relevant. Usually these are narrowed to paraphilia and personality disorder. Someone sexually dangerous largely because of bipolar illness or mental retardation might qualify for a civil commitment of some sort, but not (technically) under an SVP law, at least not according to the logic of Crane. The same is true for chemical dependency disorders.
Several models have been proposed to describe hypersexuality: sex addiction (Carnes, 1983; Goodman, 1992, 1998); compulsive sexual behavior (Quadland, 1985; Coleman, 1990, 1991, 1995), and sex/mood dysregulation (e.g., Kafka, 1991). These can be referred to as “Axis I” (descriptive, not etiological) models in that they seek to understand hypersexuality as either a subtype of, or a disorder analogous to, an already recognized Axis I disorder, e.g., chemical dependency, obsessive-compulsive disorder, bulimic or binge-eating disorders, or some kind of impulsive control disorder (e.g., pathological gambling; see Goodman). While most of these papers are conceptual or focused on case studies, they do provide detailed guidance in assessing hypersexuality, particularly the works of Eli Coleman and Aviel Goodman.
This writer has proposed a general definition of hypersexuality that is (largely) neutral to descriptive theoretical models (Montaldi, 2002). It is important to bear in mind that a definition is not a set of inter-rater reliable or validated diagnostic criteria. Definitions do not operationalize constructs to the degree necessary for reliable identification of the phenomenon being defined. [This is why there must be operationalization of constructs in developing an empirical study.] As noted, however, there are numerous published papers that address the specific issues of diagnostic assessment. Hypersexuality is defined as follows:

  1. Excess of sexual behavior to the point of severe distress and/or impairment (the latter defined to include harm to others), with:
  2. Either of the following conditions:
    (a) Impaired micro-control over sexual behavior: The pattern of sexual behavior is unmanageable in that the person is frequently unable to stop or alter any given instance of a sexual behavior without abnormally severe distress and/or impairment of functioning (even if only temporarily); or
    (b) Impaired macro-control: The person is capable of behaving differently on any given occasion without undue distress or impairment, and the behavior may be desired, but he or she persists in the behavior despite the predictable negative consequences of sexual/romantic choices. As a result, they do not (or act as if they cannot) learn from sexual experience. The result is a pattern of behavior that is both maladaptive and inflexible.

Condition (b) pertains to what this writer calls an “Axis II” variant of hypersexuality in which someone’s sexual or romantic pattern is more similar to a personality disorder than an Axis I disorder (whether or not the person meets full criteria for a personality disorder diagnosis).
In general, Axis II impairments (macro-control problems) will not satisfy SVP standards. However, impaired micro-control would seem to be (or cause) a condition approaching “serious difficulty” controlling behavior. Someone with little micro-control would be expected to present with a sexual pattern similar to the non-sexual symptoms listed in DSM criteria for pathological gambling, chemical dependency, or some aspects of Obsessive Compulsive Disorder. These individuals would not lack all control, but what control they have is abnormally tenuous and unreliable.
In sum, a person who is likely to commit sexual violence because of impairment rather than (or more than) criminal volition, is most often someone who suffers from impaired micro-control over sexual behavior. He will likely suffer from some form of hypersexuality, whether paraphilic or not. Evaluators who do not find enough evidence to support a reasonably certain professional opinion that a particular offender is hypersexual (and not just paraphilic or personality disordered) cannot logically conclude that the offender is importantly different from the “typical recidivist,” whatever that offender’s actuarial score, list of risk factors, or DSM diagnoses.


This paper addressed some crucial issues in the logic of SVP law and related psychosexual assessment. The arguments offered here are made in full awareness of the fact that SVP centers would house many fewer residents if SVP evaluations and legal proceedings followed the logic of Crane but without its assumption about paraphilias. Indeed, many of the offenders fitting the public’s image of “sexual predator” would probably not be committed. Finally, this writer is well aware of the difficulty if not impossibility of assessing hypersexuality when offenders are unwilling or unreliable in reporting mental states and sexual history beyond their documented criminal acts.
Nevertheless, the constitutional requirements for legitimate civil commitment are what they are. Certain assumptions about sexual disorders are baseless, regardless of their popular acceptance. If information is lacking, then it is lacking. Clinical judgments should not be made from inadequate evidence or unsupported presumption, and no clinician should offer opinions based on legal concepts he or she feels unwilling or unable to interpret and investigate with the assistance of psychosexual information. It is the responsibility of clinicians to indicate to the court what can be said about an offender’s capacity for sexual self-restraint, and what cannot be known with reasonable professional certainty. In many but not all cases, the issue of whether an offender has abnormally limited ability to control sexual behavior simply cannot be determined. In these cases, the most that can be said is that the offender is motivated to continue offending, has supportive attitudes and habits, and is thus highly disposed to continue doing sex crimes. The offender might also fall into a category found to have the highest sexual reconviction rate according to actuarial instruments, and possess one or more additional dynamic or (non-actuarial) static risk factors. But if logic and constitutional law are taken seriously, this will not be enough to establish SVP status. Parsimony demands compelling evidence of SVP-impairment; otherwise, an offender should be regarded as provably different from the typical dangerous offender. If this does not permit the commitment of as many dangerous persons as we would like, then perhaps civil management through mental health law is not the best way to accomplish community protection from sexual offending.16


  1. American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders – text revision. Washington, DC: Author.
  2. American Psychiatric Association. (1994). Diagnostic and statistic manual of mental disorders (4th ed.). Washington, DC: Author.
  3. Arnett, P. A., Smith, S. S., and Newman, J. P. (1997). Approach and avoidance motivation in psychopathic criminal offenders during passive avoidance. Journal of Personality and Social Psychology, 72, 1413-1428.
  4. Colb, S. F. (2002). The Supreme Court’s decision in Kansas v. Crane: Which “dangerous” people can be incarcerated? Retrieved January 30, 2002, from FindLaw’s Legal Commentary. externer Link in neuem Fenster folgt
  5. Coleman, E. (1990). The obsessive-compulsive model for describing compulsive sexual behavior. American Journal of Preventive Psychiatry and Neurology, 2, 9-14.
  6. Coleman, E. (1991). Compulsive sexual behavior: New concepts and treatments. Journal of Psychology and Human Sexuality, 4, 37-52.
  7. Coleman, E. (1992). Is your patient suffering from compulsive sexual behavior? Psychiatric Annals, 22, 320-325.
  8. Coleman, E. (1995). Treatment of compulsive sexual behavior. In R. C. Rosen & S. R. Leiblum (Eds.), Case studies in sex therapy. (pp. 333-349). New York, NY: Guilford.
  9. Doren, D. (2002). Evaluating sex offenders: A manual for civil commitments and beyond. Thousand Oaks, CA: Sage Publications.
  10. Doren, D. (2006). When statistical analyses just will not do. Sex Offender Law Report, 7, 49-64.
  11. Faigman, David L. (2004) Laboratory of justice: The Supreme Court’s 200-year struggle to integrate science and the law. Times Books. New York: Henry Holt and Company.
  12. Foucha v. Louisiana, 504 U.S. 71 (1992).
  13. Gillies, D. (2000). Philosophical theories of probability. New York, NY: Routledge.
  14. Goodman, A. (1992). Sexual addiction: Designation and treatment. Journal of Sex & Marital Therapy, 18, 303-314.
  15. Goodman, A. (1998). Sexual addiction: An integrated approach. Madison, CT: International Universities Press.
  16. Gorenstein, E. E., and Newman, J. P. (1980). Disinhibitory Psychopathology: A new perspective and a model for research. Psychological Review, 87, 301-315.
  17. Hare, R. D. (2003). Hare Psychopathy Checklist – Revised (PCL-R): 2nd edition, technical manual. Multi-Health Systems. Toronto, Ontario.
  18. Hare, R. D. (1993). Without conscience: The disturbing world of the psychopaths among us. New York: The Guilford Press.
    In re the Matter of Leon G., 59 P.3d 708 AZ S.Ct. (2002).
  19. Janus, E. S. (2006). Failure to protect: America’s Sexual Predator Laws and the rise of the preventive state. Cornell University Press: Ithaca, NY.
  20. Jaynes, E. T. (2003). Probability theory: The logic of science. New York, NY: Cambridge University Press.
  21. Jenkins, P. (1998). Moral panic: Changing concepts of the child molester in modern America. New Haven, CT: Yale University Press.
  22. Kafka, M. P. (1991). Successful antidepressant treatment of nonparaphilic sexual addictions and paraphilias in men. Journal of Clinical Psychiatry, 52, 60-65.
  23. Kansas v. Crane, 534 U.S. 407 (2002).
  24. Kansas v. Hendricks, 521 U.S. 346 (1997).
  25. Kosson, D. S., and Harpar, T. J. (1997). Attentional functioning of psychopathic individuals: Current evidence and developmental implications. In J. A. Burack and J. T. Enns (Eds.). Attention, development, and psychopathology. (pp. 379-402). New York: Guilford Press.
  26. Martin v. Reinstein, 987 P.2d 779 AZ App. (1999).
  27. Mercado, C. C., Schopp, R. F., & Bornstein, B. H. (2005). Evaluating sex offenders under sexually violent predator laws: How might mental health professionals conceptualize the notion of volitional impairment? Aggression and Violent Behavior, 10, 289-309.
  28. Miller, H.A., Amentia, A. E., & Conroy, M.A. (2005). Sexual predator evaluations: Empirical evidence, strategies for professionals, and research directions. Law and Human Behavior, 29, 29-54.
  29. Montaldi, D. F. (2002). Understanding hypersexuality with an Axis II model. Journal of Psychology & Human Sexuality 14, 1-23.
  30. Newman, J. P. (1997). Conceptual models of the nervous system: Implications for antisocial behavior. In D. M. Stoff, J. Breiling, and J. D. Maser (Eds.). Handbook of Antisocial Behavior (pp. 324-335). New York: John Wiley and Sons, Inc.
  31. Newman, J. P. (1998). Psychopathic behavior: An information processing perspective. In D. J. Cooke, A. E. Forth, and R. D. Hare (Eds.). Psychopathy: Theory, research, and implications for society (pp. 81-104). Dordrecht, The Netherlands: Kluwer Academic Publishing.
  32. Newman, J. P., and Lorenz, A. Z. (2002). Response modulation and emotional processing: Implications for psychopathy and other dysregulatory psychopathology. In R. J. Davidson, K. Scherer, and H. H. Goldsmith (Eds.). Handbook of affective science (pp. 1043-1067). Oxford University Press.
  33. Newman, J. P., Brinkley, C. A., Lorenz, A. R., Hiatt, K. D., and MacCoon, D. G. (2003). Psychopathy and psychopathology: Hare’s essential contributions. In H. Herve and J. Yuille (Eds.). Psychopathy in the third millennium: Theory and research.
  34. People v. Burris, 102 CA App. 4 (2002).
  35. Prentky, R. A., Janus, E., Barbaree, H., Schwartz, B. K., & Kafka, M. (2006). Sexually Violent Predators in the courtroom: Science on trial. William Mitchell College of Law: Legal Studies Research Paper Series, 50, 1-37.
  36. Quadland, M. C. (1985). Compulsive sexual behavior: Definition of a problem and an approach to treatment. Journal of Sex & Marital Therapy, 11, 121-132.
  37. Schopp, Robert F. (2001) Competence, condemnation, and commitment. American Psychological Association. Washington, D.C.
  38. Sreenivasan, S., Weinberger, L. E., & Garrick, T. (2003). Expert testimony in Sexually Violent Predator commitments: Conceptualizing legal standards of “mental disorder” and “likely to reoffend.” The Journal of the American Academy of Psychiatry and the Law, 31, 471-485.
  39. Vognsen, J., & Phenix, A. (2004). Antisocial Personality Disorder is not enough: A reply to Sreenivasan, Weinberger, & Garrick. The Journal of the American Academy of Psychiatry and the Law, 32, 440-442.
  40. Walsh, T. C. (1999). Psychopathic and nonpsychopathic violence among alcoholic offenders. International Journal of Offender Therapy and Comparative Criminality, 43, 34-48.
  41. Woodworth, M., and Porter, S. (2002). In cold blood: Characteristics of criminal homicides as a function of psychopathy. Journal of Abnormal Psychology, 111, 436-445.
  42. Zander, T. K. (2005). Civil commitment without psychosis: The law’s reliance on the weakest links in psychodiagnosis. Journal of Sexual Offender Civil Commitment: Science and the Law, 1, 17-82.

Author Note

Daniel F. Montaldi, Ph.D., is the evaluator for the Arizona Community Protection and Treatment Center. The ACPTC treats Sexually Violent Persons for the state of Arizona.
Dr. Montaldi completed a postdoctoral fellowship under Eli Coleman, Ph.D., and Michael Miner, Ph.D., at the Program in Human Sexuality, University of Minnesota Medical School.
Correspondence concerning this article should be addressed to the author at E-Mail Link


1 See externer Link in neuem Fenster folgt for complete texts of these and other court opinions cited in this paper.

2 The possibility of “emotional” impairment as one kind of “serious difficulty” might suggest otherwise, but only superficially. Supreme Court opinions on SVP laws leave open the possibility that “serious difficulty” includes more than “volitional impairment,” but only because the cases in question did not call for specific attention to the issue. [Hendricks claimed to have no control over sexual urges, suggesting a volition-based impairment.] “Emotional impairment” might seem like a way to treat psychopaths as impaired when they otherwise seem fully aware and (coldly) controlled. However, most criminal courts would consider lack of concern to be an aggravating, not a mitigating factor. It does not render someone dangerous “beyond control.” Deterrence logic does not presuppose empathy, conscience, or respect. It presumes only self-interest, volition, and awareness of society’s rules and institutions. Psychopaths have all of these. The issue of psychopathy and “serious difficulty” is discussed in a later section.

3 This point was lost on California Court of Appeals in People v. Burris, 102 CA App. 4 (2002). Ironically, the Court acknowledges that Hendricks “focused on lack of control because this factor serves to distinguish those recidivist violent sexual offenders who should be dealt with civilly from those who should be dealt with criminally”. The Court does not explain why “recidivist violent sexual offenders” properly deal with civilly are persons unlikely to be deterred, whereas recidivist violent sexual offenders properly dealt with criminally are not. The latter would also seem to “lack control,” at least in the sense of “failure to exercise control.”

4 Clinical judgers apparently err against mechanical or actuarial judgers not by missing many reoffenders (false negatives) but by predicting too many reoffenders (false positives). See Doren, 2002.

5 Doren (2006) draws a distinction between prediction and probability estimates. He is right that giving a numerical probability to a potential event (e.g., reconviction for a sexual offense) is not the same thing as predicting the event, given that the estimate might be correct even if the event does not occur. So even if there is no rain today, an estimate of 60% chance of rain might be an accurate reflection of rain “risk,” given data concerning past frequencies of rainy days relative to frequencies of days like today. But it is also true that people make predictions based on probability estimates. No climatologist keeps a job for long by just coming up with probability estimates. He or she is expected (when working for a news agency) to be a weather forecaster. Forecasters make cautious predictions (e.g., of storms, or where a storm will hit) based on a set of complex probabilistic estimations. So an evaluator giving a probability estimate about an offender’s chances of sexual recidivism is not avoiding prediction. If the estimate is high, then the evaluator is giving the basis for a cautious but reasonable (not certain or absolute) prediction that the person will reoffend at some unspecified point in the future if he is released (and not watched or restricted). SVP dangerousness is supposed to be high probability of future sexual violence. It is difficult to see how it is coherent to be, on the one hand, reasonably certain that someone is highly dangerous, yet on the other hand to be neutral or agnostic about whether this person would do harmful acts if left to his own devices in a world full of vulnerable persons.
Indeed, even the assumption that “likely” (or its equivalent in various state SVP laws) should be interpreted as a numerical or statistical probability threshold is questionable. Courts have noted that statutory concepts inevitably include “some vagueness” yet are intended to be understood by “the ordinary person exercising ordinary common sense” (Broadrick, 413 U.S. at 608, 93 S.Ct. 2908, cited in Martin; see below). Such language is often used qualitatively rather than quantitatively in its application. The qualitative nature of “likely” is illustrated in the Arizona Court of Appeals review of SVP terminology in Martin v. Reinstein, 987 P.2d 779 AZ App. (1999), which was later endorsed by the Arizona Supreme Court in In re the Matter of Leon G., 59 P.3d 708 AZ S.Ct. (2002). The court describes the term “likely” as “reasonably understood” and “effectively used often in our laws,” giving as examples the clause “character likely to deceive or defraud the public,” said to be an adequate standard for limiting the term “unprofessional conduct” (from a case involving the Arizona Board of Chiropractic Examiners; Lathrop, 182 Ariz. 172, 178, 894 P.2d 715, 721, App. 1995), and the clause “under circumstances likely to produce death or serious physical injury,” said to provide a clear standard by which to limit “endanger” for purposes of a child abuse statute (Poehnelt, 150 Ariz. 136, 144, 722 P.2d 304, 312, App. 1985). Neither of these uses of “likely” is intrinsically quantitative or computational, much less actuarial, and ordinary persons would not seem prone to thinking of the phenomena at issue in this manner. [This is not to say that experts conducting empirical studies of what percentages of subjects in a sample are misled by chiropractic claims or children injured in certain home environments would have nothing relevant to contribute. The point is only that the law intends jurors to evaluate phenomena potentially affecting the public good by carefully considering the individual features and circumstances of the specific case before them, and what any reasonable person would expect as a natural result of these facts. Jurors need not wait for information about how many people are actually harmed by roughly similar cases studied in aggregate.]
Clearly, legislatures intended SVP laws to confine those sex offenders any sensible observer would expect to recidivate at some point in the future if released, given the evidence available about a particular offender’s current mental condition (evidence that includes both past behavior and more recent mental states). It is not at all obvious that courts want from risk assessment only group-wise probability estimates drawn from sexual recidivism rates, provided by experts neutral as to what this means for the future of a particular offender if he is released into the community. Expectations are not definitive predictions of when a future event will occur. It is always possible that expected outcomes might not happen. We accept this as an inherent feature of expectations. But just like the non-occurrence of an event does not falsify a statement about numerical odds of occurrence, non-occurrence of an expected event does not indicate that the evidence available at the time did not support that expectation.

6 For these authors, apparently, virtually anyone with more than one sex offense is paraphilic. This position might well be consistent with the elasticity of DSM criteria, but it lacks any clear empirical basis, a point discussed later in this paper.

7 Like Crane, Leon G. is also vague, so it is possible that typical recidivists are TR offenders with only one sex offense. However, logically this would create the problem of pushing all sexual recidivists into SVP. This would make it much larger than a “narrow” category, or a third category would be created, namely, sexual recidivists who are not typical but who fail to meet full SVP criteria. Such a group is not even mentioned by Crane or Leon G., This suggests against the interpretation of “typical recidivist” as a category including only non-sexual recidivists and recidivists with only one sex offense.

8 Actuarial rankings or percentages are irrelevant to this opinion. Think about this individual as an individual: He likes sex with children. He acts out with ease. He has sexually offended before, more than once. He feels entitled to do as he pleases. He will take his chances about getting caught. Prison is not enjoyable to this man, but it is not experienced or anticipated as if it is the end of the world. Overall, this individual has gotten away with other sexual offenses. Whether or not this man goes on to reoffend, it is at least reasonable to expect more sexual crime from this offender if he is left to his own devices. In fact, it would seem more than reasonable, i.e., an expectation we would hold with confidence, though not absolute certainty. In contrast, suppose such a man falls into a lower risk category according to an actuarial instrument because (for example) he has only gotten caught for one of many offenses. Given his particular characteristics, not just those of his cohort, it is still reasonable to think that this individual is more like the minority in observed samples drawn from his actuarial category, those reconvicted for a new sex offense, rather than the majority who were not.

9 Like our typical recidivist, there is ample reason to regard our representative SVP offender as “likely” to re-offend sexually. Once again, actuarial findings are irrelevant. Think about this individual’s (not his cohort’s) characteristics: He is strongly drawn to sexual offending and has little control over deviant urges. He has acted on these urges many times in the past. Little control means he would have serious trouble stopping himself even if he really wanted to stop and struggled to do so. Given the facts of this man’s mental disorder itself, there would seem to be ample basis for expecting more offending from him if he is left to his own devices in a world full of children. If his actuarial rating is high, this only confirms what is already (more than) reasonable to expect. If his actuarial rating is low, based on cohort characteristics, his individual characteristics still easily support a reasonably confident expectation (which is absolute knowledge) that he will go on to accumulate more risk-related characteristics. All high scorers on actuarial instruments were once low scorers (i.e., earlier in their lives they had criminal histories that would have earned a lower score). Either way, actuarial information adds nothing important beyond information about the volitional nature of a particular individual’s mental disorder as it presents in his specific case.

10 Most, but not all. Someone who is genuinely “out of control” sexually but not yet falling into a high actuarial group will be missed by many risk assessments as they are currently conducted. This writer knows of one case which did not go to trial because the offender’s actuarial score was considered too low to meet the “likely” threshold. This assumes jointly that “likely” should be interpreted as a numerical likelihood and that a cohort sample’s reconviction rate is its estimated value (both questionable). In this particular case, the man had made multiple attempts to mutilate his genitals in order to escape his urges. Such cases show that “serious difficulty” controlling sexual behavior does not depend on high actuarial scores, and that the individual characteristics of a person’s mental state can be sound grounds for considering him dangerous, with or without cohort comparisons.

11 The Court construes “likely” as a concept tightly connected to “mental disorder” through “serious difficulty.” As discussed in an earlier note, the Court’s discussion of these concepts in Leon G. strongly suggests a qualitative interpretation of “likely” in SVP law. This is further suggested in the Court’s comparison of “likely” in Arizona’s SVP statute to interpretative language used by other courts, such as “reasonably certain to accrue in the future”, “probable or reasonably to be expected”. Rather than connoting a computable quantity, “probable” is used to mean that future sexual violence in SVP is more than a “mere possibility”. The Court interprets the Arizona State Legislature as intending “likely” to be a standard “somewhat higher” than “probable”, or “highly probable”, in order to restrict SVP status to a “small but extremely dangerous group of sexually violent predators”. Similarly, the Court sees legislative intent in using the term “likely” as establishing a standard “more stringent” than “reasonably expected” in non-SVP civil commitment law. This implies that these standards are comparable in kind yet different in degree. Were the SVP standard quantitative, e.g., an actuarial statistic, the Court would be comparing apples to oranges. How would “50%+ likelihood of reconviction” be more stringent than “reasonable to expect”?

12 If evaluators are also assuming that rapid recidivists are at higher risk than their actuarial cohort (and not just that risk comes from control deficits), then they are assuming that offenders composing the samples used in the development of actuarial instruments did not sexually re-offend within a few months of release. Otherwise, it is not clear how rapid recidivism constitutes an additional risk factor. The same point applies to additional victims and offenses beyond those resulting in arrest or conviction (or other characteristics not yet included on actuarial instruments). The issue is not that rapid recidivism and extra victims provides no additional information relevant to risk, just that there is no clear basis for ascribing to the offender an additional risk factor beyond actuarial findings, as in “the more risk factors, the more risk.” Instead of suggesting that such information calls for “adjustment” of actuarial results, as if to say that someone is at higher risk than others in his actuarial cohort, it is better to say that more risk-related information suggests (but does not establish) that the offender is more similar to individuals in his actuarial cohort who were reconvicted, whatever percentage they comprised, as opposed to those offenders who were not reconvicted. This claim is not quantifiable, and proves nothing by itself, but it is more sensible that saying the offender belongs in a different and more (quantitatively) risky cohort unspecified by actuarial instruments. Of course, all this would change if new data were to show that a group of offenders with a given score on an actuarial instrument, also having the feature of rapid recidivism and/or additional victims, had a higher reconviction rate over 5, 10, or 15 years post-release than a group with the same score but without these characteristics. In that event, there would be grounds for distinguishing a new cohort, one with a higher reconviction rate than previously recognized. The instrument itself should then be revised to include more items, rather than its findings “adjusted.”

13 To add further complication, criminals who quickly re-offend often do so in part because they quickly resume substance abuse. So the real issue in many cases is control over chemical use. But impaired control over sexual urges largely secondary to a substance-related disorder is not the same thing as impaired control coming from sexual (or personality) disorder, even if a sexual disorder determines the object or direction of the urge. Substance abuse disorders are not typically included on lists of SVP mental disorders.

14 It can be meaningful to say, as SVP statutes commonly state, that a particular individual’s mental disorder “makes” or “causes” him to be dangerous. Even outside of a forensic context, clinicians can make sense of such statements. It is not meaningful to say that a mental disorder, as it presents in a specific case, makes or causes the individual to have an actuarially derived risk level. The disorder can (at most) be said to cause the individual to behave in a manner that accumulates a criminal history, and that history may place the person into a category defined by history. The category itself is not what generates a risk estimate with actuarial instruments, but rather a characteristic attributed to that category – a probability of reconviction – based on inference from the observed reconviction rates in researched samples drawn from that category. [Such an inference presupposes a “frequentist” model of probability, which is considered inaccurate by contemporary probability theorists; see Gillies, 2000, and Jaynes, 2003. This is a separate issue.] It is certainly possible that there is some overarching causal factor or set of factors, perhaps sociological and psychological, responsible for a group effect, such that this factor or factors “cause” persons in these samples to exhibit a given rate (or range of rates) of reconviction for sexual crimes. But such a factor or factors cannot consist of any one individual’s mental disorder, which causes only his potential for harmful acts. Groups categorized by criminal sexual history can be composed of persons displaying different mental disorders (e.g., rapists and child molesters) or perhaps no disorder in some cases. Finally, reconviction rates characterize groups of offenders, not any one offender. Reconviction (or not) characterizes the offender himself. Probability of reconviction can therefore be said to characterize an individual offender (in a subjective or logical model of probability, not a frequentist model) but not to characterize a group or category of offender. About a category of offenders, it is possible to talk about the probability that randomly drawn samples will display a given rate of reconviction, but a probability or likelihood of reconviction (itself) is not an attribute that can be “associated with” the category (contra Doren, 2002, p. 139). Statistical risk based on rates (frequencies, percentages, proportions) is thus not attributable in any causal sense to the characteristics of an individual, including his mental condition.

15 Abel, G. G., & Rouleau, J. L., (1990). Male sex offenders. In M. Thase, B. Edelstein, & M. Hersen (Eds.), Handbook of outpatient treatment of adults: Nonpsychotic mental disorders. Specific page reference is not known.

16 Or, as a reviewer of this paper correctly noted, perhaps civil management through full confinement is not the best way to protect the public. It is possible to construct a civil commitment statute with a broad range of “less restrictive alternatives” to confinement, such as civil managed probation and outpatient treatment, half-way houses, group homes (for lower functioning offenders), and so forth. Sex offender civil commitment in Texas is outpatient by statute. In Arizona, SVP civil commitment has evolved from mostly full confinement to a community reintegration program.

Author address

Daniel F. Montaldi, Ph.D.
E-Mail Link