Handbook on Sexual Violence (41 page)

Kelly’s prevalence continuum of sexual violence, which combines all types of violence, can then be thought of within this formulation.

Behaviour patterns associated with

different classes of sexual violence

Table 7.3
Reconceptualising Kelly’s prevalence continuum

Core behaviours of sexual violence

The emboldened behaviours having more than 70 per cent frequency are undifferentiating of a particular manifestation of sexual violence whereas those with lower frequencies begin to look like more specific forms such as obscene phone calling or sexual abuse. There are no behaviours in Kelly’s formulation that map onto the Canter MO and signature levels, but these are theoretically possible if looked for as demonstrated in Table 7.2.

To recapitulate the argument thus far, Canter’s (2000) theorising proposed a model to distinguish between types of crime (say burglary, robbery, and violence). Youngs (2004) was able to demonstrate this showing clear distinctions between property-based crime and those crimes committed against the person. Thereafter, within a particular crime such as rape, behaviours can be differentiated by general to increasingly particular actions grouped around a number of overarching themes (Canter and Heritage 1990). The present chapter takes these ideas to present a model conceptualising the different subclasses within sexual violence. Thus it should be possible to focus on the broader area of sexual violence and show the within variations of its different manifestations. This is represented in Figure 7.2.

Domestic violence

Sexual harassment

Signature

Modus operandi

Behavioural pattern Typicality

Stalking

Sexual murder

Rape

Figure 7.2 A model differentiating subclasses of sexual violence (adapted from Canter 2000)

Such a model hypothesises that each subclass of sexual violence would have behaviours that are typical of all classes of sexual violence, represented by the dense shading at the centre, as indeed is shown by Kelly. Thereafter behaviours reduce in frequency and begin to separate such that there are distinguishable behavioural patterns associated with different types of sexual violence, as implied by Kelly. Then behaviours become less frequent and begin to be associated with different styles of crime typifying offenders. It is at this level that types of offender can be identified and differentiated such as the erotomanic or predator stalker (Mullen
et al
. 2000) or the exploiter, violent and controlling sexual murderer (Salfati and Taylor 2006). Finally, behaviours that are low frequency would identify specific offenders by virtue of singular, signature behaviours, not indicated by Kelly’s prevalence continuum. This is exemplified by the particular method of killing, a ligature, used by railway murderer John Duffy (
The Independent
, 3 February 2001).

The position of the subclasses of sexual violence around the model is also

significant. Adjacent classes would expect to have more behaviour in common and that there is a more direct connection in terms of escalation across boundary behaviours than with a type of sexual violence further round the circle. As shown previously in Table
7.2, there are overlapping behaviours that feature in more than one subclass of sexual violence. Moreover, there also may be some crossover by a perpetrator, for example the intimidating aspects of sexual harassment at work parallel the intimidation of domestic violence at home. Domestic violence may lead to murder as can rape. The potential escalation and crossover between types of sexual violence can be represented diagrammatically as shown in Figure 7.3.

Sexual harassment

stalking

Domestic violence

Murder

Rape

Figure 7.3
Potential crossover between subclasses of sexual violence

Thus sexual harassment or domestic violence may migrate to stalking, stalking in turn can result in murder, as indeed may rape or domestic violence. Moreover, some behaviour may occur in more than one subclass, for example it is likely that obscene phone calling may feature in stalking, sexual harassment or domestic violence. Stealing some item from the victim may occur in rape, murder or stalking. Demeaning the person by putting them down in front of others may occur in sexual harassment and domestic violence. Adjacent subclasses would be expected to share more behaviours

than those further away around the circle. This then adds a qualitative distinctiveness to the Kelly prevalence continuum. There is a relationship between behaviours in their likelihood to occur together. Moreover this suggests different pathways, for example an offender may start by sexually harassing, and then engage in stalking, leading to rape and eventually murder. Alternatively, there may be a measure of domestic violence in a relationship which then escalates to rape and ultimately murder (as appeared in the case of John Duffy). It is also possible that an individual may only engage in one specific form of sexual violence.

Finally, it should be possible to identify themes within each subclass of sexual violence. Canter (2000) employs the term theme to overcome the problem with types or taxonomies in that often there is not a ‘pure’ type, rather it is more likely there will be some degree of overlap between behaviours. As has been demonstrated above, themes have been used to describe criminality in general and the crime of rape in particular. It is likely then that themes can be found that are associated with other manifestations of sexual violence such as expressive or instrumental themes identifying homicide behaviours (Salfati 2008).

Empirical support for qualitative differentiation

Salfati and Taylor’s (2006) idea of a continuum proposes that crime scene behaviours are associated, to a greater or lesser extent, with the type of crime, and as a function of the association can be placed on the continuum. This in effect suggests there is a stronger relationship of behaviours that are adjacent to each other and a weaker one between those behaviours that are further away on the continuum. They sought to differentiate between the behaviours of rapists and sexual murderers. By using a multivariate statistical analysis they were able to map rapist and murderer behaviour along a continuum as follows.

Crime scene behaviours Blindfold

Binding

Ripping clothing Bringing weapon Stealing

Vaginal penetration

Removing all victim’s clothing Stealing valuable item

Forensic awareness Single wound Sexual activity Multiple wounds

Weapon found at scene Non-controlled violence Anal penetration

Object inserted

}
rapist

}
murderer

Figure 7.4
Continuum of rapists’ and murderers’ behaviours (after Salfati and Taylor 2006)

The behaviour at the top is typical of a rape and less likely found to typify murderers’ behaviour whereas anal penetration and inserting an object were more likely to be associated with sexual murder than rape. The emboldened behaviours were found in both offences. The typical rape behaviours appear to be more instrumental in controlling the victim whereas in the murder part of the continuum, behaviours appear more expressive and hostile acts of violence. The sexual behaviours at the end of the continuum are especially violent and degrading. Salfati and Taylor conclude that this does demonstrate a progression and represents a movement away from planned controlling behaviours to more extreme hostile and expressive acts perpetrated against the victim.