I would like to offer three comments on Kendell's useful conceptual exploration of personality disorder (Reference KendellKendell, 2002). First, reduced life expectancy, which Kendell passes on to us, sceptically from Scadding, as a core, defining feature of disorder is implausible. As this criterion refers to aggregate data about a social group, not a claimed causal link about a particular individual, it prompts an odd conclusion. For example, both male gender and poverty predict (reduced) longevity. Does this mean that being male or poor are medical disorders? Such a medicalisation of material or existential disadvantage would surely stretch a metaphor very thinly.

Second, a categorical diagnostic approach (disordered/non-disordered) makes us a hostage to fortune when researching interventions. If we are obliged to ask the categorical question ‘is personality disorder treatable?’, it will produce a predictably ambiguous answer (Reference Dolan and CoidDolan & Coid, 1993). From this flows an understandable ambivalence about the willingness to ‘treat’ among general psychiatrists (Reference Cawthra and GibbCawthra & Gibb, 1998) and even among some forensic psychiatrists (Reference CopeCope, 1993). If we asked a different sort of question, such as, ‘can we reduce the re-offending rates of sex offenders using this specific intervention’, we might get a useful probabilistic answer about trying to change some people who habitually offend our moral order in a particular way. For example, it is cost-effective to offer psychological interventions (note: not ‘treatment’) to detained sex offenders as a group, even though risk prediction at the individual level remains problematic on release.

Third, the ambiguities Kendell correctly exposes about the relationship between personality disorder and mental illness also apply to the permeable boundary with normality. Common aspects of parliamentary life (e.g. sexual and financial ‘sleaze’ and the routine impression-management of politicians), some sport (e.g. boxing and hunting) and some private sexual activity (e.g. consensual sadomasochism) overlap strongly with DSM criteria for variants of ‘dramatic’ personality disorder. In my view, this points to the logical superiority of a dimensional over a categorical approach (Reference PilgrimPilgrim, 2001).

Readers may correctly spot that this dimensional preference is predictable from a psychologist, which highlights that the ‘nature’ of ‘personality disorder’ is bound up with the constructs favoured by particular professional groups. However, Kendell, a psychiatrist, also argues that a dimensional view makes more sense (he calls them ‘graded traits’) — suggesting that a categorical approach has now failed us all, both scientifically and pragmatically. The category of personality disorder is not inherent to those who gain the label, but is a by-product of our professional discourse. A further indication of this point is that whether a detected child molester becomes a prisoner or a patient is a function of multi-party professional judgements. Thus, ‘personality disorder’ is socially negotiated — it does not exist ‘out there’ waiting to be discovered. If we go looking, we find ‘it’, in vast amounts, via circular psychiatric epidemiology (Reference KullerKuller, 1999), particularly in prison populations. In my view we should abandon the concept of personality disorder altogether and appraise whether and how society (not just mental health professionals) can respond correctively to the wide range of role/rule violations it subsumes.