Mixed Emotions: An Incentive Motivational Model of Sexual Deviance

Existing Models

Over the past decades, numerous theories and models on the onset of sexual offending have been developed (see Ward, Polaschek, & Beech, 2006, for a comprehensive overview of theories/models and their respective empirical support). Although the broad consensus is that sexual offending behavior is a complex and multifaceted phenomenon, certain basic commonalities can be drawn from the existing models. First, from Finkelhor’s (1984) precondition model to Seto’s (2017a) motivation-facilitation model, all existing theories recognize an important motivating role for sexual deviance, mostly in the shape of illegal sexual interests^a and illegal sexual preferences^a as a driving force behind sexual offending. (“Illegal” refers to sexual interests that are illegal if acted upon [e.g., children]; the interest itself is of course not illegal.) Second, all existing sex offender models suggest that disinhibition significantly increases the likelihood that someone will commit a sexual offense. Disinhibition can take the shape of a temporary state, for example, intoxication (e.g., Seto, 2017a) or a more stable trait (e.g., Hall & Hirschman, 1992; Ward & Siegert, 2002), rendering the offender incapable of controlling the driving force of his illegal sexual interest, despite the mores of society. There is also a more deliberate form of disinhibition in the shape of antisociality, for example, lack of empathy (Marshall & Barbaree, 1990) or detachment (Ward, Hudson, & Marshall, 1996), which renders the offender, regardless of capability, unwilling to inhibit acting on his illegal desires.

A third commonality among most current sex offender models with the exception of evolutionary approaches (more on this in “Sexual Offending Behavior is Sexual Behavior”) lies in their predominant description of sexual offending behavior as a variant of offending behavior, drawing on research comparing sexual offenders to nonsexual offenders or nonoffenders. The factor disinhibition, both in the shape of impulsivity and antisociality, is indeed equally common in models related to nonsexual offending behavior (e.g., Andrews & Bonta, 2010) and there appears to be considerable overlap between sexual offending behavior and other offending behaviors. This is reflected in the often broad criminal careers of sex offenders (e.g., Lussier, LeBlanc, & Proulx, 2005), the possibility to predict future sex offending from general offending in juveniles (Beaudry-Cyr, Jennings, Zgoba, & Tewksbury, 2017; Zimring, Jennings, Piquero, & Hays, 2009), and the higher nonsexual than sexual recidivism rates for sex offenders (e.g., Hanson & Bussière, 1998; Smallbone & Wortley, 2000). However, sexual offending at the same time is regarded as a very typical form of offending behavior in need of its own models and theories (Ward et al., 2006). Overall, the disinhibition factor is regarded as closely related to offending behavior in general while the sexual deviance factor is seen as typical for sexual offending. The available recidivism research seems consistent with this assumption: Disinhibition-related factors predict all types of reoffenses while sexual deviance–related factors predict sexual reoffenses specifically and make no significant contribution to the prediction of nonsexual offenses (e.g., Brouillette-Alarie, Babchishin, Hanson, & Helmus, 2016). Sexual deviance appears to be the key factor that makes the offender a sex offender. Consequently, one of the most interesting questions in sex offender theory becomes, “What causes sexual deviance?”

Some sex offender models include precursors to the development of sexual deviance in the form of early negative experiences, such as childhood maltreatment and more specifically childhood sexual abuse (e.g., Malamuth, 1996; Finkelhor, 1984; Marshall & Barbaree, 1990; Ward & Beech, 2006). Childhood sexual abuse is theorized to either directly influence the development of pedophilic interest^a or indirectly through disturbed attachment styles (Ward et al., 1996), lack of self-esteem, intimacy deficits (Finkelhor, 1984; Marshall & Barbaree, 1990), or poor social skills (Ward & Siegert, 2002). These traits are thought to create a social focus on children who are regarded as safe and accepting, while these traits at the same time may limit sexual access to adults (Finkelhor, 1984). It remains unclear how the social focus on children becomes a sexual interest, although Marshall and Barbaree (1990) suggest this may happen through a process of aggression displacement, reconfirmed by conditioning effects. None of these suggested causal routes have been empirically validated, although an association between childhood sexual abuse and abuse perpetration is often found (Jespersen, Lalumière, & Seto, 2009).

Ward and Beech’s (2006) integrated theory of sexual offending (ITSO) takes a step further back from early childhood experiences to innate vulnerabilities as precursors of risk factors for sexual offending behavior. However, examples provided by the authors refer to innate vulnerabilities for general criminal behavior, not for sexual deviance. Similarly, Kafka’s (2003) neurobiological theory implies an innate biological precursor of sexual offending behavior by suggesting that pedophilia is associated with dysregulation of serotonergic receptor sensitivity. However, Kafka assumes that this dysregulation results in a propensity toward behavioral disinhibition (Soubrie, 1986) and specifically toward violence and impulsive aggression (Kavoussi, Armstead, & Coccaro, 1997), thus again linking the biological precursor to disinhibition rather than to sexual deviance. It is worth noting that this is not in accordance with the clinical observation that pedophilic offenders often do not seem to act impulsively and violently but tend to groom their victims. The presence of deviant sexual interests and the acting out of those interests appear to be two different things. Whereas the acting out is likely influenced by disinhibition, the existence is probably not. The existence of nonoffending pedophiles attests to this, and recent research suggests that they distinguish themselves from offending pedophiles by better inhibition (Kärgel et al., 2017).

As Ward et al. (2006) conclude in their overview of current sex offender theory, the available models mostly provide a good overview of what we know about the (correlational) relationship between risk factors and sexual (re)offending but have not yet yielded many suggestions or testable hypotheses regarding the causal pathways between these risk factors and sexual offending behavior. The same applies to our knowledge of sexual deviance. Possible risk factors for the development of sexual deviance have been suggested, such as childhood sexual abuse; however, we are not sure through which mechanisms childhood abuse may lead to pedophilic interest.

More in general, most current models have not yet offered convincing explanations as to why so many, and so many relatively normal men commit sexual offenses (Smid, van Beek, & Troelstra, 2011). Equally interesting and unanswered is the question why most perpetrators of sexual offenses do not seem to persist, as it is becoming increasingly clear that even untreated (Schmucker & Lösel, 2015) and possibly unpunished (Bailey, Bernhard, & Hsu, 2016) sex offenders in majority do not persist in their sexual offending behavior. And although some of the current models do provide a feedback loop where sexual satisfaction resulting from offending behavior further entrenches deviance and increases the tendency to sexually offend, they do not provide a “way out.”

An Incentive Model of Sexual Motivation

To compare sexual offending behavior to normative sexual behavior, we will use an incentive model of sexual motivation (for an extensive description, see Toates, 2014). Throughout the previous century, a drive and regulation view of sexual desire largely dominated the literature (e.g., Freud, 1955; Kaplan, 1995; Watson, 1930). In this view, sexual desire is seen as an intrinsic drive triggered by a bodily deficit or imbalance, comparable to hunger or thirst. Not having sex is assumed to increase tension and the desire for sex, while having sex is assumed to relieve tension and lower sexual desire. The organism is seen as regulating to keep itself in a sexual balance: homeostasis. Having too much or too little sex is perceived as a cause of physical and psychological problems.

However, little empirical support has been produced for this drive model. While thwarted sexual desire may give rise to feelings of frustration, not having sex is not detrimental to physiological functioning (e.g., Ågmo, 2008; Beach, 1956) or the cause of disease (e.g., Leitzmann, Platz, Stampfer, Willett, & Giovannucci, 2004) nor does it appear to lead to “sublimation” in the shape of increased creativity (Kinsey, Pomeroy, & Martin, 1948). Moreover, a lack of partner sex does not increase masturbation frequency (Laumann, 1994) or nocturnal orgasms (Kinsey, Pomeroy, Martin, & Gebhard, 1953). There is just as little indication for the possibility of “getting sex out of the system,” that is, reducing sexual desire by having sex. Neither in the short term, because having a lot of sex does not decrease sexual desire (beyond the limited physical refraction period), nor in the long term, for example, there is no evidence for less adultery in societies permissive of premarital sex (Suggs & Marshall, 1971). Experimental research shows that although sexually aroused subjects subsequently rate attractive individuals as still more attractive, the individuals they initially found unattractive are rated as even less attractive when sexually aroused (Istvan, Griffitt, & Weidner, 1983). This polarization effect contradicts the existence of an intrinsic sexual drive, searching any adequate outlet.

An incentive motivational approach steps away from the intrinsic drive and regulation approach and instead emphasizes an important initiating role for the stimulus:

An alternative to the view that we are pushed by something arising within the tissues of the body and then happen upon a suitable means of discharge is termed the incentive view of motivation [. . .]. That is to say, people are pulled by external stimuli, thoughts about them and their associations. [. . .] In the view advanced here, rather than internal factors, such as hormone levels, driving the animal, human or otherwise, they modulate the power of the incentive to exert a pull. [. . .] Such a view has now largely replaced drive interpretations inside scientific sexology (Both, Everaerd, & Laan, 2007). (Toates, 2014, pp. 103, 107)

In general, there is no motivation without a stimulus that predicts reward or danger (Schultz, 1998) and this also applies to sexual motivation^b (Key terms regarding sexual motivation are marked with ^b and defined in Table 2). People, much like animals, become initially sexually aroused by neurochemical and hormonal changes following the perception of a competent stimulus^b (Herbert, 2001; Pfaus, 2009). In a process referred to as liking^b (Berridge, Robinson, & Aldridge, 2009; Robinson & Berridge, 1993), parts of the brain, that is, insula and the orbitofrontal cortex (Craig, 2002; Morris, 2002; Sumich, Kumari, & Sharma, 2003) interpret the neurobiological changes caused by the competent stimulus as initial sexual arousal^b, a process involving the endogenous opioid system and other reward systems (Pfaus, 2009).

OPEN IN VIEWER

Table 2.

Terms and Definitions Concerning the Presented Incentive Motivational Model.

Sexual motivation	Sexual motivation is a construct, used to explain the generation of sexual behavior, involving the processes of liking, wanting, and (dis)inhibition.
Competent stimulus or stimulus with incentive salience	Stimulus (in vivo or fantasy) signaling possible sexual reward.Partly evolutionary determined and partly the result of experience and learning processes
Initial arousal	Initial arousal in brain and body triggered by the competent stimulus, not yet entering awareness, including (low-level) genital arousal.
Liking	Appreciation of the stimulus, consisting of feedback of the initial arousal to relevant brain parts that interpreted it as sexual arousal and trigger/facilitate the wanting process. Unconscious in the early stages and conscious in the later stages
Wanting	Dopamine-driven approach tendency toward the competent stimulus. Unconscious in the early stages and conscious in the later stages.
Sexual desire	The combination of liking and wanting, forming the (subjective) experience of a magnetic pull toward the sexually attractive stimulus, in search of closer acquaintance/stronger sexual stimuli
Inhibition	Automatic or deliberate suppression of sexual desire or an “escape” to alternative places or activities
Sexual arousal	Sexual arousal explicitly experienced by the individual as such.
Genital arousal	Arousal measurable as increase in erection or vaginal blood flow, low level and unconscious in the early stages, more prominent and conscious in the later stages.
Sexual gratification	Usually orgasm/ejaculation, sometimes another type of peak moment that involves consummatory pleasure and induces a period of recovery.

Simultaneously, the processing of the competent stimulus results in physiological changes preparing the body for action/appetitive behavior, a process referred to as wanting^b (Berridge et al., 2009; Robinson & Berridge, 1993) involving the limbic system and mesolimbic dopamine pathway (Pfaus, 2009). Dopamine is the central neurotransmitter in this motivational process, involved in both reward-signaling and the initiation of motor responses (Phillips, Stuber, Heien, Wightman, & Carelli, 2003). Testosterone is known to sensitize these neurological pathways, making it more likely that a given stimulus will trigger a wanting reaction (Hermans et al., 2010). Dopamine does not seem to be involved in assessing the valence of a stimulus (liking), but is involved in the tendency to approach a stimulus (wanting; Berridge, 1996). Various studies have shown facilitating effects of dopamine on sexual motivation in animals (Melis & Argiolas, 1995) and in humans (Both, Everaerd, Laan, & Gooren, 2005; Meston & Frohlich, 2000).

In other words, the processing of a competent stimulus automatically prepares an organism for sexual activity, by means of two processes, liking and wanting, underlined by two distinguishable neurobiological processes. Liking refers to the positive affective response to a certain stimulus, and wanting refers to the tendency to approach such a stimulus. Together they form what we will call sexual desire^b, a magnetic pull toward the competent stimulus. In general, a stronger stimulus (e.g., a more explicitly sexual or a more preferred stimulus) will elicit a stronger reaction. Competent stimuli may be perceived in the environment or may be imagined in case of sexual fantasy. The evaluation of stimuli, which is largely subconscious and automatic, is influenced to a great extent by an individual’s personal history of sexual reward and sexual gratification^b (Pfaus, Kippin, & Centeno, 2001). This means that personal experience plays an important part in the evolvement of stimuli that elicit the strongest reaction: the most competent stimuli (Toates, 2009).

The presence of a competent stimulus (actual or imagined) will trigger liking and wanting, but whether or not the individual proceeds to take action depends on the individual’s aptitude for (dis)inhibition^b. (Dis)inhibition involves quick subconscious and automatic elements, increasingly automated through the process of socialization and learning as well as a slower more conscious and deliberate appraisal of opportunities and risks and possible rewards and detriments, open to cognitive processes, for example, weighing the consequences (Toates, 2009). Serotonin is reported as the neurotransmitter most closely related to these inhibition processes (Dalley & Roiser, 2012; Pfaus, 2009).

Note that the model implies that sexual desire is caused by initial sexual arousal rather than sexual desire causing sexual arousal, although acting on this desire will in turn reinforce arousal. Instead of assuming that people feel (intrinsic) sexual desire and therefore set out to find stimuli that provide sexual arousal, the model assumes that competent stimuli arouse people and are consequently interpreted as sexually desirable. However, it is important to note that people are cognitively evolved; they do not live in the moment, but remember yesterday and look forward to tomorrow. Instead of merely feeling or not feeling sexual arousal, people may want to become sexually aroused and actively seek out stimuli to arouse them; they may “want to want” (Toates, 2014, p. 139).

Besides sexual arousal, an incentive motivational approach is applicable to a wide range of emotions (Craig, 2002; Morris, 2002; Sumich et al., 2003) and appetitive behaviors (Georgiadis & Kringelbach, 2012). Functional imaging studies show that subjective experiences of different emotions like anger, disgust, and anxiety are associated with activation of the same brain structures, indicating a generalized route for emotional states and not to a specific feedback route for sexual arousal (Both et al., 2007). Experimental research has shown that both sexually arousing and anxiety-inducing stimuli evoked different subjective appreciation, but comparable general somatic motor system responses, which prepare the organism for action (Both, Everaerd, & Laan, 2003).

Reviews report ample evidence for the involvement of distinguishable brain structures, activation, and neurotransmitters in sexual arousal (e.g., Georgiadis, Kringelbach, & Pfaus, 2012) and other emotional processes (e.g., Georgiadis & Kringelbach, 2012) that are in line with an IMM. Experimental research has shown that stronger sexual stimuli evoke proportionally larger wanting reactions (Both, Boxtel, Stekelenburg, Everaerd, & Laan, 2005), which are in turn associated with the increased likelihood of engagement in actual sexual behavior (Both, Spiering, Everaerd, & Laan, 2004). Although it is not the only possible perspective, we believe an incentive model of sexual motivation provides a useful framework for the description of sexual behavior including sexual offending behavior, with the ability to produce new and testable hypotheses.

The Sexual Motivation Cycle

It is important to recognize that sexual arousal^b is (mostly) not a goal in and of itself; the goal of sexual arousal is usually sexual gratification. However, sexual arousal is not an on/off button that guarantees gratification once arousal is present. The road from initial sexual arousal to orgasm rather has been described as a complex and subtle interplay of inhibition and disinhibition (Pfaus, 2009). We propose a model in which, from the perception of a competent stimulus to the actual engagement in sexual interaction and the achievement of gratification, the motivational cycle of liking, wanting, and disinhibition is repeated multiple times (see Figure 1).

OPEN IN VIEWER

Figure 1.

A schematic representation of the sexual motivation cycle.

The individual’s brain is evolutionary tuned to detect relevant stimuli that signal the possibility of sexual reward, assumed they are not currently engaged in other important activities that vie for their attention (Toates, 2014). Such a competent sexual stimulus triggers initial arousal in brain and body that is (unconsciously) interpreted as sexual (liking) and incites an approach tendency toward the attractive stimulus (wanting), which may be acted upon if not inhibited. The individual approaches the desired stimulus in an effort to make a connection, for instance by flirting. If contact is achieved, this provides a reward in the shape of a stronger stimulus which in turn invokes a stronger reaction, leading to closer contact. Initial arousal in the early stages to a still relatively mild stimulus or arousal that is sufficiently (automatically) inhibited can be latently present but may remain unconscious to the individual and may not be recognized as sexual arousal. Arising behavior can nonetheless be placed under the denominator of sexual behavior.

Subsequently, the motivation cycle is repeated to achieve stronger, more sexually explicit stimuli (kissing, touching, etc.) and to gradually increase sexual arousal while decreasing inhibition. Gradually, the sexual arousal will become strong enough and the inhibition low enough to enable fully aware genital arousal^b. Continuation of the sexual interaction will further increase arousal and lower inhibition, preferably neither too slow nor too fast, until gratification is achieved. Theoretically, the individual can step out of the sexual motivation cycle at any time if sufficient inhibition is incited (automatic or deliberate); however, this is likely to get increasingly more difficult when the individual is closer to orgasm.

Consistent with this assumption of a cyclic and gradual process of sexual motivation, studies in rats have found that levels of dopamine gradually increase throughout the process of approach, anticipation, and copulation (Mas, Fumero, Fernandez-Vera, & Gonzalez-Mora, 1995), while at the time of ejaculation serotonin is released and dopamine release is inhibited (Hull et al., 1999). To further test this assumption, it is recommended that future research aim to cover all possible stages of sexual motivation, from initial liking and approach tendency to orgasm, in addition to the current predominant focus on genital arousal triggered by the use of explicit sexual stimuli. As an example, viewing time measures, used for the assessment of sexual interest, appear to rely on the earlier stages of sexual motivation. These instruments use nonsexual stimuli depicting potentially attractive individuals, assuming that the attention given to these stimuli reflects someone’s sexual interest in them (e.g., Schmidt, Babchishin, & Lehmann, 2017). This type of assessment can be expanded to the more direct assessment of approach tendencies toward attractive stimuli by means of an approach avoidance task (AAT; Weidacker et al., 2017).

Sexual Offending Behavior

The proposed IMM can be applied to the description of sexual offending behavior and its relationship to normal sexual behavior (Smid & Wever, 2016). First, the motivational model assigns an important role to (dis)inhibition and purports that (dis)inhibition is not just a significant factor related to offending behavior but is also a crucial factor in sexual behavior. Second, the sexual deviance factor is prominent in the model and can be described in terms of stimulus preference and liking. Sexually deviant interests can be understood as a liking reaction to a deviant stimulus. Taking an intermittent position between evolutionary/biological and learning models, the IMM assumes that competent stimuli are an evolutionary given to a certain extent but that there is room for adaption and refining by means of learning and conditioning. From an evolutionary perspective, people will be more attracted to characteristics signaling greater evolutionary fitness or “mate value” (Symons, 1995), such as greater physical health (e.g., smooth skin) and reproductive potential (e.g., younger women). However, some determinants of stimulus attractiveness are culturally determined, such as body size (Buss, 2003), and others are idiosyncratic, reflected in the finding that mere exposure also increases stimulus attraction (Meston & Buss, 2009). Furthermore, prior research has shown that even straightforward classical conditioning can, to some extent, attach nonsexual stimuli to sexual arousal. This has been shown both in animals (e.g., Pfaus, Erickson, & Talianakis, 2013) as well as in humans (Brom, 2016; Brom, Both, Laan, Everaerd, & Spinhoven, 2014). Another important feature of the IMM is that it describes sexual arousal itself as an emotional reaction to a stimulus, comparable to other emotions. This creates a new perspective regarding mechanisms of sexual self-regulation as well as general emotion regulation in relationship to sex. The mechanisms of sexual self-regulation will be further elaborated on in “Sexual Deviance and Deficiency” and “Sexual Coping: A Two-Way Street”. But first we will use the model to further discuss the prevalence of sexual offending behavior and sex offender recidivism.

As mentioned, an IMM implies an important role for (dis)inhibition in normal sexual behavior. A desired deviant sexual stimulus will not be approached if adequate inhibition occurs. And a normative sexual stimulus will not be (further) approached if the liking and wanting is inadequately overinhibited. Disinhibition is, however, not the entry point of the cycle. For someone to unsuccessfully inhibit sexual action toward an illegal stimulus (and eventually commit an offense), he must first be at least somewhat sexually attracted to that illegal stimulus. In other words, an offender needs to first appreciate the stimulus to reach the point of (dis)inhibition. Illegal sexual interest is therefore crucial to sexual offending, like in other sex offender models.

Research shows that most prevalent sexual offending behavior is directed against girls/women above 10 and below 25 years old, clearly and consistently peaking around age 14 to 15 years, regardless of the age of the perpetrator (Felson & Cundiff, 2014; Finkelhor, Ormrod, & Chaffin, 2009; Snyder, 2000). Felson and Cundiff (2014) conclude that “the risk of sexual victimization increased dramatically with sexual maturation and declined dramatically after the teenage years” (pp. 282). Not coincidentally, teenage and young adult women are also the most common competent stimuli for men, and the authors go as far as to suggest that the pattern they found in the ages of rape/abuse victims “reflects male sexuality” (pp. 273). Research shows that (some) illegal sexual interest can be quite common. And although there is ample discussion on whether or not sexual interest in pubescent children is adaptive from an evolutionary perspective (e.g., Seto, 2017b vs. Rind, 2017), sexual arousal to stimuli involving pubescent girls from age 13 years and up is commonly found in phallometric research as is sexual arousal to rape stimuli (Lalumière, Harris, Quinsey, & Rice, 2005). Although the levels of these illegal interests may vary in strength, some level of interest is so common that it appears to be (even) more prevalent than the associated offending behaviors. This means that disinhibition, in all its modalities, is the key factor for the occurrence or nonoccurrence of the associated sexual offending behaviors.

As discussed in the “Existing Models” section, there is ample support for the importance of disinhibition in sexual offending in general. For instance, intoxication, causing a state of disinhibition, is present in a substantial proportion of sexual assaults. Around one third of all sex offenses (Kraanen & Emmelkamp, 2011) and in 40% to 70% of college sexual assaults (Abbey, Wegner, Woerner, Pegram, & Pierce, 2014) are committed under the influence of alcohol or drugs. Relatively normal men can be capable of uncharacteristic behavior when under the influence of alcohol or drugs. This behavior can be limited to (one-time) sexual offending behavior without being associated with other antisocial behavior, because (dis)inhibition is a core feature of sexual motivation. Based on an IMM, we would expect to see this type of offense more in situations where the sexual motivation cycle is already well underway, for example, parties with many attractive individuals and possibilities to approach. And counterintuitive as it may seem, we would also expect to see this behavior more often involving a victim that the perpetrator actually really likes, for example, dating situations or (casual) relationships. These expectations match the literature on this topic (e.g., Gidycz, Warkentin, & Orchowski, 2007). As far as we know, an IMM is the only model that provides a clear offense path for this group of offenders.

Most sexual offenses are directed against victims that are generally considered attractive; however, a smaller but still substantial part of sexual offending behavior is directed against girls and boys under the age of 12 years. Although there is a positive relationship between children’s age and prevalence, younger victims are more rare (Felson & Cundiff, 2014). And a small part of sex offenses consist of extreme forms of rape, like sadistic rape or lust murder. Very young children and extreme violence or cruelty are not general or common competent stimuli for men. Research indicates that some level of sexual interest in prepubertal children can be found in approximately 20% of men (Green, 2002; Seto, 2008) and 4% to 10% have been reported to actively fantasize about prepubescent children (Ahlers et al., 2011; Dombert et al., 2016). Ahlers et al. (2011) also report sadistic fantasies in 22% of their subjects. It is unclear how many of those are interested in extreme violence, torture, or murder, but that is likely to be less. The IMM, consistent with other models, accounts for these types of offenses, by suggesting a more important role for sexual deviance: Most men will not have these sexual interests and find no need to inhibit sexual behavior when faced with a possible stimulus of this kind. Those who do have these interests still need to disinhibit to actually commit an offense. This combination of factors renders these offenses far less likely to occur.

Sexual Deviance and Deficiency

Sexual deviance is defined in many different ways in the literature. Sometimes the term is used to refer to any amount of sexual interest in any illegal or unusual sexual stimulus^a (e.g., Stable-2007; Hanson, Harris, Scott, & Helmus, 2007), other times it is used to refer to an exclusive preference^a for a specific illegal stimulus, but often it is described as something in between. However, the assessment of sexual deviance, especially the implicit assessment, nearly always involves the assessment of individual preference for illegal stimuli (phallometric assessment, Implicit Association Tests, Viewing Time measures; Smid et al., 2011). This is what we refer to in this article as illegal sexual preference^a: an assessed preference by means of stronger erectile responses, longer viewing times, quicker reaction times to illegal stimuli than to normative stimuli.

But can an illegal sexual preference really be best and only understood as an exceptionally high level of interest in an illegal stimulus? If we take a closer look at the sexual preference studies that report raw data, these studies nearly always find both higher than average levels of interests in illegal stimuli and lower than average levels of interest in normative stimuli in deviant groups. Often, but certainly not always, the low level of normative interest even emerges as the proportionally larger effect (e.g., Becker, Kaplan, & Tenke, 1993; Blanchard et al., 2002; Howard, Longmore, Mason, & Martin, 1994; Nunes, Firestone, & Baldwin, 2007; Poeppl et al., 2011; Renaud et al., 2014; Rice, Harris, Lang, & Chaplin, 2012; Seto, Cantor, & Blanchard, 2006; van Leeuwen et al., 2013). It seems that deviant groups are at least as much characterized by remarkably low interest in normative stimuli as by a remarkably large interest in deviant stimuli. Even if we account for the probability that lower-than-expected responses to illegal stimuli may be partly due to deviant offenders “faking good,” it is still noticeable that they seem to be unable to fake good on the normative stimuli. Based on existing research, it is difficult to definitively assign primacy to either deviant overarousal or normative underarousal with regard to deviant preferences. Results vary (e.g., Blanchard et al., 2009; Chaplin, Rice, & Harris, 1995; Fromberger et al., 2013) and included subjects as well as used measures and measurements units (e.g., z-scores) may not be optimal for this purpose. New studies are required, preferably including “raw” measurements among relatively young subjects, who are not yet fully settled and conditioned in their preferences.

Nevertheless, the hypothesis that illegal sexual preference is associated with a lack of normative sexual interest is as valid as the assumption that illegal sexual preference is driven by an exceptionally high level of illegal interest. This provides us with an alternative perspective of illegal sexual preference as a deficiency. This applies to both pedophiles and paraphilic coercive rapists, with pedophiles often showing low interest in adults and rapists often showing low interest in consent. With respect to rapists it has been noted before that their “indifference to consent” is central to their sexual proclivities (Harris, Lalumiere, Seto, Rice, & Chaplin, 2012).

A deficiency perspective of sexual deviance is more consistent with clinical experience, where our deviant patients often do not strike us as people with extra special traits, but as men who are deficient in various ways. More importantly, this deficiency perspective provides us with a potential explanation for the common occurrence of occasional sexual offending versus the far more rare habitual sexual offending. Given the very common interest in some illegal stimuli (i.e., teenage girls and nonconsent), relatively many men will sexually offend at one point in their lives due to trait/state disinhibition. But most of these men are more likely to leave these potentially attractive stimuli alone for the sake of the consequences and instead focus on the legal stimuli, which also provide them with enough sexual motivation. However, those men who lack rewarding alternatives because they are not sufficiently triggered by normative stimuli will be far more likely to persist in their offending behavior even in the face of dire consequences. Or alternatively, they may resort to finding visual material reflecting their deviant preference online, which they may think of as less harmful or less risky. Finkelhor (1984) includes a lack of access to normative stimuli as a key factor in his model. However, access to normative stimuli will not help if there is a low sexual interest in these stimuli. Note that this assumption means that habitual deviant sex offenders could be very different from “common” sex offenders, to the point that they may even be opposite in some ways. It has been noted before that existing data contradict the use of a single model of arousal for all sex offenders (Barbaree & Marshall, 1991). Therefore, we recommended that studies on sexual deviance select and describe their included samples carefully and avoid the use of heterogeneous groups of “sex offenders” or “child molesters.”

Sexual Coping: A Two-Way Street

The IMM suggests that sexual arousal is an emotion that can influence other emotions as well as be influenced by other emotions. The fact that other emotions can be influenced by sex has been widely recognized by the sex offender field for quite some time. The Relapse Prevention Model (e.g., Pithers, 1990; Pithers, Marques, Gibat, & Marlatt, 1983) assumes that high-risk situations, which form the first step in the sex offender relapse sequence, most frequently consist of negative emotional states that trigger sexual fantasies and (offending) behavior as a consoling mechanism. Proulx, McKibben, and Lusignan (1996) find that the frequency of masturbation to deviant sexual fantasy is indeed significantly associated with negative moods in sex offenders. More recently the more general idea of “sexual coping” was introduced (Cortoni & Marshall, 2001), referring to the predominant tendency to use sexual gratification (often masturbation) to relieve negative emotions of sadness, anger, frustration, and so forth. The nature of the relevant sexual fantasies or conduct in sexual coping is not necessarily deviant (see Figure 2). This sexual coping tendency is present in a substantial part of sex offenders and is regarded as an important risk factor for sexual reoffending (e.g., Stable-2007; Hanson et al., 2007). At its core, the suggested IMM poses the assumption that the connection between sexual self-regulation and emotional self-regulation is a two-way street. A person can regulate their emotional states by means of sexual activity (sexual coping) as well as regulate their sexual arousal by means of other emotions (deviant arousal).

OPEN IN VIEWER

Figure 2.

A schematic representation of self-regulation by means of sexual behavior.

Drawing further on this reciprocal connection between sexual arousal and other types of emotional arousal, it can be hypothesized that sexual coping will not solely be used to down-regulate high negative emotional arousal but may also serve to up-regulate positive emotional arousal (see Figure 2). We argue that this mechanism underlies a wide variety of fetishisms. Fetishists often use simmering, low-level sexual arousal to achieve an enduring comfortable arousal level. In this case sexual gratification is not the goal, but the sexual arousal itself is. Or, to be more accurate, the general positive arousal is the goal and is achieved by sexual arousal. The resulting emotional state may be so subtle that it is not clearly experienced or recognized as sexual arousal by the individual, who sometimes emphasizes that their fetish “is not sexual” it just makes them “feel good,” even if their fetish appears to be obviously sexual in nature (i.e., female masking: http://www.channel4.com/programs/secrets-of-the-living-dolls).

This mechanism of emotional up-regulation by means of (subtle) sexual arousal is consistent with the extensive comorbidity of paraphilias with Axis I mood disorders (e.g., Kafka & Hennen, 2002), although these disorders could also develop as a consequence of the deviant sexual desires and behaviors. The reported comorbidity between paraphilia and attention deficit hyperactivity disorder (ADHD; Kafka & Hennen, 2002) is however not as easily explained as a consequence. This correlation is consistent with the suggestion that people with ADHD suffer from low general arousal levels and tend to look for more stimuli to achieve higher and more pleasant levels of arousal (Sergeant, Oosterlaan, & van der Meere, 1999). Stimulants that increase the levels of dopamine are known to ameliorate the symptoms of ADHD (Faraone & Buitelaar, 2010). We hypothesize that some individuals achieve this increase in dopamine and heightened general arousal by means of sexual arousal.

The differentiation of motivations behind the use of sexual coping has not yet been the topic of empirical research. Future research may include assessment of parameters of sexual coping behaviors, for instance, time spent on sexual behavior versus sexual outlet. We hypothesize that among individuals who (predominantly) use sexual coping for down-regulation (SC-DR), time spent on sexual behavior will be paralleled by sexual outlet. In individuals who (predominantly) use sexual coping for up-regulation (SC-UR), time spent on sexual behavior will not be paralleled by sexual outlet.

Interconnectedness of Sexual Self-Regulation Factors

Various sexual regulation problems have been discussed separately; however, based on the model it is clear to see how they may be interrelated in various ways. Though some people may predominantly use sexual coping to down-regulate and others to up-regulate, clinical observation suggests that both types of sexual self-regulation sometimes co-occur, resulting in a comprehensive form of sexual self-regulation, infiltrating all aspects of life and taking up considerable amounts of time. We also expect an association between SC-UR and sexual deviance, because, as described in the “Sexual Coping: A Two-Way Street” section, we hypothesize that SC-UR will often involve forms of fetishism. Because SC-UR has not been researched as such, future research will have to shed light on this topic.

SC-DR becomes problematic if it is an individual’s only coping mechanism and particularly so at times when the individual experiences a lot of stress to cope with. A persistent focus on sexual behavior, as a result of sexual coping or otherwise, may lead to incentive sensitization (Berridge & Robinson, 1995; Berridge & Robinson, 2016). Mesolimbic dopamine sensitization renders “wanting” systems hyperreactive to sexual cues and contexts, leading to an increased urge to act (wanting) not matched by a corresponding degree of positive appreciation of the stimulus or even the expected outcome (liking). This mechanism of “empty wanting” posits the essence of addiction and has been described in many forms of addictive behavior (Berridge & Robinson, 2016) including compulsive sexual behavior (CSB; Voon et al., 2014). Based on an IMM, we further hypothesize that while CSB will result in increased sexual activity/outlet and want for sexual activity, at the same time, it will likely lead to decreased sexual arousal and sexual satisfaction, due to the lag in liking. The literature provides some results consistent with this hypothesis (Janssen, Hoffmann, Goodrich, & Wilson, 2016; Laier, Pawlikowski, Pekal, Schulte, & Brand, 2013; Miner et al., 2015; Winters, Christoff, & Gorzalka, 2010), but more and more specific research on this topic is needed.

CSB, either motivated by sexual coping or not, does not necessarily involve sexual deviance, because individuals may have relatively normal functioning sexual systems and may search for adequate normative stimuli to serve their purposes. However, the lag in liking associated with compulsive behavior, will eventually call for stronger, possibly deviant stimuli. Moreover, spending an extraordinary amount of time looking for sexual stimuli online or in real life increases the chance to come into contact with deviant stimuli and gaining (more) gratifying experiences with these stimuli. Severe overuse of the sexual system may also lead to the use of deviant stimuli through satiation: Orgasm renders the sexual system less sensitive for a period of time (Georgiadis & Kringelbach, 2012), among other things by means of increased serotonin levels (Hull et al., 1999). If the individual “wants to want” to return to sexual activity within that time frame, for example, because the previous outcome did not meet their expectations, then stronger stimuli are needed to counter these satiety effects. Deviant stimuli may provide this extra arousal and using those stimuli to reach gratification will stimulate the process of their conditioning. If the deviant stimuli become sufficiently conditioned, they become part of the collection of stimuli that may independently trigger a sexual reaction. Existing evidence indeed points toward substantial comorbidity between sexual deviance and CSB (e.g., Kafka & Hennen, 1999); however, this provides little evidence for the proposed mechanism. A more specific hypothesis we can derive from our IMM is that sexual deviance may develop as a secondary condition to CSB and/or sexual coping. This means that successfully countering CSB or sexual coping may remedy sexual deviance while leaving normative sexuality intact. Consistent with this hypothesis, a review of the literature suggests that Selective Serotonin Reuptake Inhibitors (SSRIs) may be successful in treating sexual deviance without interfering with normative sexuality, but this effect appears to be “more effective in paraphilias that resemble or are accompanied by OCD symptoms” (Garcia & Thibaut, 2011, p. 777).

Summarizing, considerable overlap between sexual regulation factors is expected, yet it remains important to distinguish them from each other. Future research should aim to assess not only the nature of problematic sexual behaviors but also their function and their history of origin.

Summary and Conclusions

This article applies an incentive model of sexual motivation to sexual offending behavior. Most importantly, the proposed IMM describes sexual offending behavior as a variant of sexual behavior rather than a variant of offending behavior. Furthermore, the model describes sexual arousal as an emotional reaction to a competent stimulus, comparable to other emotions. Emotions can interact with each other. The interaction between the sexual emotion and other emotions can, consciously or automatically, serve as a means of emotional or sexual self-regulation. Not only can sex be used to reduce negative emotional states (i.e., sexual coping), but sexual arousal can also be used to increase positive emotional states or to achieve a desired level of general arousal (e.g., fetishism). Furthermore, other emotions can consciously or unconsciously be used to regulate sexual arousal (i.e., sexual deviance). These added emotions can be experienced as the actual cause of the arousal and may become a necessary condition for arousal or even independently trigger sexual arousal as a consequence of conditioning.

It is our belief that this IMM of sexual offending and sexual deviance is internally coherent and provides an innovative addition to the existing literature that increases the scope as well as the depth of our current understanding of sexual offending behavior and sexual deviance. Furthermore, the proposed model brings together many aspects from existing theories, evolutionary, neurobiological, and psychological; clinical and empirical; forensic and sexological. Comparable to evolutionary psychology, the IMM describes sexual offending as sexually motivated behavior, but more than the evolutionary perspective, it incorporates individual flexibility and idiosyncrasy. The resulting model specifically provides more plausible explanations and hypotheses on the development of deviant preferences that serve no apparent reproductive purpose. Furthermore, in hypothesizing about causal intrapersonal mechanisms, the model allows for new avenues of inquiry and provides many testable hypotheses. Because the underlying mechanisms are assumed to be similar, many of these hypotheses can first be tested among subjects other than sex offenders using other outcome measures than sexual reoffending. This circumvents some of the research problems typical of our field. Research should aim to cover all stages of the sexual motivation cycle and preferably include, besides the specific subgroup of deviant sex offenders, comparison groups of normal subjects, men with unusual but legal preferences, and men with (nonorganic) sexual dysfunctions.

Although the model provides some predictive accuracy in the form of explanations for findings in existing research, there is no actual empirical evidence available yet and subsequent empirical studies will have to rigorously test whether the model has real empirical adequacy. A weak point of the model is that it remains relatively complicated. However, we do not believe this is a consequence of making many or far-reaching assumptions but is rather due to the inclusion of phenomena like circularity, reciprocity, and multicausality. This may well be indicative of the fact that the problem the model is trying to describe is known as complex and multifaceted and must not be oversimplified.