The concept of psychopathy was first introduced by Cleckley in 1941, emphasising features such as reduced guilt and impaired empathy (see Blair & Viding ^{Reference Blair, Viding, Rutter, Bishop, Pine, Scott, Stevenson and Taylor1} ). Research into psychopathy waited some four decades until Hare developed a questionnaire measure to identify it. Some dozen years later the psychopathy construct was extended to childhood by Frick and colleagues. ^{Reference Frick, O'Brien, Wootton and McBurnett2} Systematic research into psychopathy in childhood during the past decade or so has identified several important distinctive features that have separated it from antisocial behaviour as a whole. Thus, twin studies have shown a high heritability that exceeds that for antisocial behaviour unaccompanied by psychopathy; ^{Reference Viding, Jones, Pual, Moffitt and Plomin3} a randomised controlled trial has shown that there is a worse response to treatment when antisocial behaviour is associated with psychopathy; ^{Reference Hawes and Dadds4} there is a modest degree of continuity between psychopathy in early adolescence and adult psychopathy as assessed some years later using a difference measure; ^{Reference Lynam, Caspi, Moffitt, Loeber and Stouthamer–Loeber5} and brain imaging studies in adults and adolescents who exhibited both psychopathy and antisocial behaviour have produced striking findings with respect to amygdala dysfunction and the orbitofrontal cortex. ^{Reference Blair, Viding, Rutter, Bishop, Pine, Scott, Stevenson and Taylor1} However, it is crucial to recognise that the findings so far lack diagnostic specificity.

The paper by Dadds et al in this issue ^{Reference Dadds, Allen, Oliver, Faulkner, Legge and Moul6} is innovative in several different respects. First, by focusing on a sample of 4- to 8-year-olds it shows that callous–unemotional traits (thought to index psychopathy) can be manifest quite early in childhood. The concept of psychopathy has usually involved an assumption that its origins lie early in life but it is a new finding that it can be measured in early childhood. Second, a novel ‘love task’ was used as a social ‘press’ for interactive eye gaze. This fits in with Dadds’ suggestion from earlier research that lack of attention to the eyes may underlie the fear recognition deficits in child psychopathy. Third, findings show that the lack of attention to eyes is not associated with any difference in the mother's interactions with the children, implying that the deficit lies in the child rather than in a parenting feature.

Putting all these findings together, it is clear that a good case exists for recognising the existence, and importance, of psychopathy in childhood. It is surprising, therefore, that neither ICD-10 nor DSM-IV include a diagnosis of psychopathy for any age period. It sort of sneaks in through the side door by the inclusion of callous–unemotional traits in the criteria for antisocial personality disorder, but that is most unsatisfactory in that it leaves no room for diagnosing antisocial behaviour in adult life when it does not include these features. So what are the problems that have held back a more general acceptance of psychopathy? Four stand out. First, there are several different questionnaire measures of psychopathy with only moderate agreement among the different scales (see Blair & Viding ^{Reference Blair, Viding, Rutter, Bishop, Pine, Scott, Stevenson and Taylor1} ). There is also uncertainty on the extent to which there should be reliance on self-report ratings or other informant-report ratings. ^{Reference Kumsta, Sonuga–Barke and Rutter7} Second, it remains to be shown that the relevant features can be reliably and validly assessed clinically – the evidence at the moment relying almost entirely on questionnaire scores. Third, many, but not all, of the questionnaire measures assume that psychopathy is intrinsically related to antisocial behaviour. However, both the empirical findings in childhood ^{Reference Kumsta, Sonuga–Barke and Rutter7,Reference Rowe, Maughan, Moran, Ford, Briskman and Goodman8} as well as broader considerations ^{Reference Skeem and Cooke9} cast doubt on this assumption. Many individuals with psychopathy have not shown either oppositional/defiant disorder or a conduct disorder. Fourth, it cannot be assumed that all forms of apparent lack of concern for others reflect psychopathy. ^{Reference Jones, Happé, Gilbert, Burnett and Viding10}

To begin with, Jones et al ^{Reference Jones, Happé, Gilbert, Burnett and Viding10} noted that individuals with autism have difficulty understanding the perspective of others and consequently may react in a seemingly cold and uncaring manner in real-life situations. Their findings nevertheless suggest that this differs from callous–unemotional traits in that psychopathic features are associated with difficulties in resonating with other people's distress, whereas autism is characterised by difficulties in knowing what other people think. Second, profound institutional deprivation in early life has been associated with callous–unemotional features in adolescence, years after the children were adopted into well-functioning homes. The strong implication is that the institutional deprivation led to the callous–unemotional traits. Whether or not these are truly psychopathic features remains an open question and the mediating causal mechanisms are still poorly understood. However, the findings hark back to Bowlby's notion of ‘affectionless psychopathy’ associated with a very disrupted family upbringing (see Blair & Viding ^{Reference Blair, Viding, Rutter, Bishop, Pine, Scott, Stevenson and Taylor1} ).

There is now ample evidence to justify the inclusion of psychopathy (both in childhood and in adult life) in psychiatric classifications but, in view of the uncertainties noted above, it would probably be best to be in a grouping of disorders specified as requiring further testing.