The concept of psychopathic personality has had a tortuous history and has varied in the extent to which it denotes a disorder of personality or socially repugnant behavior (Millon & Davis, 1996). In Britain, the nineteenth century notion of "moral insanity" inferred a disorder of the person from social deviance alone. This influenced mental health legislation and led to the characterization of psychopaths as individuals who exhibit antisocial conduct from an early age.
German psychiatrists, however, saw psychopathic personalities as a generic class of disorders of personality observable among psychiatric patients and offenders, but also detectable in law-abiding citizens. Schneider (1923/1950), for example, explicitly excluded antisocial behavior from the definition. Psychopathic personalities were defined, first, by their abnormal (i.e., statistically unusual) personality characteristics and, second, their tendency to cause suffering to themselves or others. Schneider's concept has not influenced recent concepts of psychopathy, but its legacy is the broad class of personality disorders in current psychiatric classifications. It remains relevant to psychopathy because of its influence on conceptions of the nature of personality disorder more generally (Livesley, 2001).
Influenced by psychodynamic and sociological interest in crime, North American concepts of psychopathic, sociopathic, or antisocial personality focused on a narrower and specific kind of socially deviant person while emphasizing a constellation of deviant personality attributes as the defining characteristic. Millon and Davis (1996) noted the variety of concepts proposed in the clinical literature, such as instinct-ridden characters, manipulative personalities, or malignant narcissism, but the most influential concept has been that of Cleckley (1941/1982). Cleckley rejected detailed classifications of personality disorders, including Schneider's, seeing most categories as neurotic or psychotic disorders, but he proposed a "distinct clinical entity," which he considered to be a disorder of personality.
He described psychopaths as individuals incapable of leading normal lives who cause great distress in their community, despite an outer mask of robust mental health. By analogy with semantic aphasia, he hypothesized that their attributes resulted from a "deep" disorder in semantic processing, which impaired emotional awareness of what the most important experiences of life mean to others. He proposed 16 criteria, such as absence of psychosis or nervousness, superficial charm, unreliability, insincerity, lack of remorse, egocentricity, and interpersonal unresponsiveness, although some criteria seem peripheral to the construct (e.g., suicide rarely carried out, fantastic and uninviting behavior with or without drink).
Cleckley's concept has stimulated research for the past four decades, but it is instructive to compare it with two other near-contemporary conceptions. Drawing on the concept of the social self, Gough (1948) also specified a core theoretical construct that aimed to account for the behavioral manifestations. He distilled a pattern of traits from the literature that closely coincides with Cleckley's criteria, such as impulsive, emotionally shallow, and inability to form lasting interpersonal attachments, and proposed that a pathological deficiency in role-taking ability, the capacity to identify with another's point of view, synthesized the observed characteristics. Inability to experience the social emotions of contrition and group loyalty and lack of self-control were a consequence of this deficit. McCord and McCord (1964) provided an overlapping description, identifying the primary manifestations of a "psychopathic character structure" as persistent antisocial-ity; a craving for primitive satisfaction or excitement; an inability to control aggressive impulsiveness; a lack of guilt, conscience, or sense of remorse; and a defective capacity for loving others. Persistent antisocial behavior was a necessary, but not sufficient, criterion and a function of a craving for excitement. The essential criteria were guiltlessness and lovelessness.
Although apparently similar, these concepts differ in significant ways. First, each identifies the core of the construct as a deficit, but where for Gough this is a cognitive dysfunction, for Cleckley and the McCords, it is affective. Second, they employ different embedding theoretical networks. Gough drew on the symbolic interactionism of social psychology, Cleckley on a disease model of traditional psychiatry, and the McCords on a psychodynamic developmental perspective. Concomitantly, Gough specifies the etiology of the deficit in social development, the McCords in family rejection, while Cleckley's quasi-neurological concept of semantic dementia implies a constitutional defect in the nervous system.
Third, the behavioral manifestations vary. Although Cleckley observed that psychopaths may become bored easily and seek out novelty, this does not appear among his criteria, whereas both Gough and the McCords include impulsiveness. Only the McCords specify aggressiveness as a manifestation. Moreover, only the McCords include criminal behavior as a core manifestation. They are clearly describing a type of criminal, although they do not clarify why guiltlessness and lovelessness are limited to persistent lawbreakers. Neither Cleckley nor Gough, on the other hand, see psychopathic traits as the prerogative of adjudicated offenders. Cleckley distinguished psychopathy from criminality, delinquency, consistent sexual deviation, hedonism, and clinical alcoholism, and noted that the typical psychopath ". . . is not likely to commit major crimes that result in long prison terms" (1982, p. 14). Similarly, Gough suggested that psychopaths would contribute more than their fair share to criminal behavior, but stressed that psychopathy was not explicitly dependent on illegal or asocial behavior.
Somewhat inconsistently, Cleckley regarded the DSM-II category of antisocial personality disorder as a fair summary of psychopathy. This category included the attributes of "chronically antisocial" and "hedonistic" excluded from his own concept. However, these three concepts may emphasize different, though related, personality constructs. Several of Cleckley's criteria, for example, appear in current descriptions of histrionic personality disorder (superficial charm, egocentric, emotionally shallow, suicide gestures, impersonal sex life). On the other hand, the McCords' focus on persistently antisocial behavior resulting from a craving for excitement reflects a hypothesis common in criminology that criminality results from impulsivity or lack of self-control (Blackburn, 1995). Gottfredson and Hirschi (1990), for example, contend that criminality is a disposition to seek short-term, immediate pleasure without regard for the consequences.
Although the explanatory power of Gough's construct is psychologically greater than that of Cleckley's, the latter is more reductionist and had greater appeal for hard-nosed empiricists. From the late 1950s onward, Cleckley's criteria provided operational criteria for research by psychologists, notably Lykken (1995) and Hare (1970, 1991). His specific hypothesis of semantic dementia has not itself been subjected to much empirical testing, but his proposal that an emotional deficit is the core of a psychopathic disposition has been central in research.
The PCL-R is now the most widely used measure of psychopathy. It was derived through factor analysis of Cleckley's criteria and other items and consists of 20 items each scored 0 to 2 from a semistructured interview and file information. A total score of 30 or more identifies psychopaths, but this has not been consistent in research, and a cutoff of 25 is commonly recommended in Europe. Early research focused on the PCL-R total score, but a 12-item screening version (PCL-SV) has been developed as a short form to facilitate assessment in civil populations. The PCL-R has good psychometric properties that generalize cross-culturally, and the psychopathy construct has been shown to apply to preadolescent children and adolescent boys, though less clearly to females (see Harris et al., 2001, for a review).
Hare (1991) argues that the "Cleckley psychopath" is the clinical basis for the PCL-R. However, there are several reasons for questioning the coordination of the measure with this construct. First, although the PCL-R contains several of Cleckley's defining traits, it goes well beyond them and includes several items not derived from Cleckley's construct, such as grandiosity, impulsivity, and juvenile delinquency. Factor analysis revealed two correlated factors (Harpur, Hare, & Hakstian, 1989), the first reflecting a selfish, callous, and remorseless use of others (e.g., superficial charm, egocentricity, lack of empathy); the second, social deviance or a chronically unstable and antisocial lifestyle (e.g., proneness to boredom, parasitic lifestyle, impulsivity, juvenile delinquency). For brevity, these factors are referred to as F1 and F2, respectively, in the following discussion. The former is considered to represent the core of psychopathy as described by Cleckley.
Harpur et al. (1989) found several differential correlates of the two factors. F2, but not F1, related inversely to intelligence and social class, suggesting its association with working-class criminality. F2 also related more strongly than F1 to DSM-III antisocial personality disorder, whereas F2, but not F1, correlated with narcissistic disorder. Self-report personality measures generally correlate only weakly with the PCL-R, but scales of narcissism, Machiavellianism, and lack of empathy correlate significantly with F1, while impulsivity and sensation seeking correlate with F2 (Hare, 1991).
Hare (1991) saw the total PCL-R score combining the two factors as a "unidimensional measure," and reported high correlations between this score and ratings of Cleckley's criteria. However, as Lilienfeld (1994) noted, the contribution of F2 implies that antisocial behavior is necessary to the concept of psychopathy, contrary to Cleckley's view. In fact, the description of F2 as "social deviance" is misleading, because it mixes traits (impulsivity, irresponsibility) with antisocial items (juvenile delinquency, revocation of conditional release). This factor can perhaps be construed as the disposition of self-control identified by Gottfredson and Hirschi (1990) to underlie criminality. Cleckley's construct may, then, be measured with some fidelity by F1, but the PCL-R total represents a drift away from this concept. The PCL-R may well be overall a comprehensive representation of psychopathy, but it is closer to the concept of the McCords than that of Cleckley.
A second example of conceptual drift is the current focus of research on the predictive utility of the PCL-R in assessing risk for violent recidivism. The PCL-R consistently identifies the more persistent offenders and is one of the best single predictors of both violent and nonviolent recidivism in released prisoners and mentally disordered offenders, as well as a predictor of institutional misconduct (Hare, Clark, Grann, & Thornton, 2000; Harris et al., 2001; Hemphill, Hare, & Wong, 1998; Salekin, Rogers, & Sewell, 1996). This is paradoxical given that neither aggression nor impulsivity appears as explicit traits among Cleckley's criteria. It is assumed that Cleckley's construct of an affective deficit accounts for an increased propensity for violence, but the validity of this requires first that the relationship is mediated primarily by F1 because it seems more obviously predictable from F2. The evidence is equivocal. In a meta-analysis, Hemphill et al. (1998) found that F2 correlated more highly with general recidivism than did F1, but both factors contributed equally to the prediction of violence. However, this was based on only a small number of studies, and Salekin et al. (1996) questioned the strength of the F1 relationship. Skeem and Mulvey (2001) found that although the PCL-SV total score was a relatively strong predictor of violence in a civil psychiatric sample, its predictive power came from the equivalent of F2 rather than. As Hart (1998) notes, the proximal cause of a violent crime is not psychopathy, but rather a decision to act violently. A psychopathic disposition may mediate such decision making through several possible mechanisms, such as perceptions of hostile intent, a generalized affective deficit, or lack of forethought. In the absence of research permitting a choice among these, the primacy of an affective deficit remains a tenuous assumption. For example, Woodworth and Porter (2002) found that among homicides, the offenses of psychopaths were primarily "instrumental." This was mediated mainly by F1 and was held to reflect murder "in cold blood." However, the violence of psychopaths was commonly motivated by vengeance and retribution, which are "hot" emotions more characteristic of "angry" than "instrumental" aggression, and their victims were more likely to be female. A more plausible interpretation is that their crimes reflected the narcissistic grandiosity of F1 and were an excessive reaction to threatened egotism (Baumeister, Smart, & Boden, 1996).
A further source of conceptual drift is the absence from the PCL-R of several items considered critical by Cleckley, notably absence of psychosis and absence of nervousness. The omission of these exclusionary criteria questions whether the measurement model matches that required by the construct (Messick, 1981). Cleckley's construct of psychopathy as "a distinct clinical entity" is clearly categorical. Exclusionary criteria are appropriate to categorical diagnosis, which does not require a homogeneous scale. The latter is more appropriate to a continuous variable. The measurement model of the PCL-R is ambiguous and reflects ambivalence about whether psychopathy should be construed as a discrete entity or a continuous dimension, an issue on which Hare (1970, 1998) declares himself agnostic. The total score reflects a homogeneous dimension (Hare, 1991), but it is often interpreted as the extent to which an individual meets the prototype of a psychopath and is used to create a dichotomy. This is not unreasonable for much research, but cutoff scores distinguishing psychopaths from nonpsychopaths can create the illusion that a discrete category has been identified. In fact, the common cutoff for dichotomizing the PCL-R is relatively arbitrary (Salekin et al., 1996).
The PCL-R measurement model more clearly meets the requirement of a quantitative dimension specified earlier by Hare (1970, p. 12) than that of a discrete category. In this dimensional concept, a single score on the PCL-R represents a person's degree of psychopathy, and psychopaths "exist" only as a convenient fiction representing an extreme degree. Implicitly, a dimensional conception acknowledges the absurdity of assuming that traits such as egocentricity, unreliability, or deceitfulness are limited to only a small section of humanity. However, Harris, Rice, and Quinsey (1994) argue that psychopathy is a discrete entity and not merely the end of a continuum. Applying taxometric methods to measures of psychopathy and childhood variables reflecting antisocial conduct, they found evidence for a taxon. However, chronic antisocial behavior beginning in childhood was the central feature rather than the affective and personality characteristics associated with F1. This was replicated by Skilling, Harris, Rice, and Quinsey (2002), who also included APD items. PCL-R and APD items were found to be highly correlated, and the level of prediction of violence was similar for both sets of items. Again, however, PCL-R F1 items were peripheral to the taxon. It, therefore, appears that a natural class may underlie the persistent rule-breaking assessed by APD and F2. This rediscovery of "the born criminal" should give us pause for thought.
The two-factor model has been questioned by Cooke and Michie (2001). Using confirmatory factor analysis, they developed a model in which three correlated factors are necessary to specify a superordinate construct of psychopathy: deceitful interpersonal style, deficient affective experience, and impulsive and irresponsible behavioral style. They contend that all factors contribute to a similar extent. In this model, F1 divides into the interpersonal and affective components and seven PCL-R items are deleted, most of them implicitly or explicitly entailing criminal behavior. Criminal behavior then becomes a correlate or a consequence of psychopathy rather than a core feature, consistent with Cleck-ley's views (Cooke, Michie, Hart, & Clark, in press). However, the three-factor model does not clarify whether the emotional deficit hypothesized to be the core of the disposition is reflected in the superordinate dimension or only one of the component factors. The model calls for a radical revision of the theoretical construct to accommodate this structure.
Cleckley's concept is often described as the primary or classical concept of psychopathy, as distinct from secondary psychopathy. The distinction originated in psychodynamic conceptions, which differentiated socially deviant behavior resulting from deficient conscience from similar behavior arising from neurotic conflict or other psychopathology (e.g., McCord & McCord, 1964). Influenced by the learning theories of the time, early research linked the hypothesized emotional deficit of primary psychopathy to low anxiety and secondary psychopathy to high anxiety. This was the basis for theories that nonconformity in psychopaths resulted from problems created by low anxiety or fearlessness for socializing children through child-rearing methods relying on punishment. This theory is maintained by Lykken (1995), who now replaces the notion of secondary psychopaths with that of "sociopaths" who are criminal but not characterized by high fear. Although studies using personality questionnaires support the idea of secondary or anxious psychopaths (Blackburn, 1996), this questionnaire typology relates only weakly to the PCL-R (Blackburn, 1998), consistent with findings that self-reports do not readily tap the traits of F1. There is, nevertheless, a growing literature that suggests that psychopaths are not a homogeneous group (Skeem, Poythress, Edens, Lilienfeld, & Cale, 2003).
The issue of secondary psychopathy is theoretically critical to Cleckley's construct. Although it is assumed that the PCL-R measures primary psychopathy and is, hence, inversely related to anxiety, the omission of Cleckley's exclusionary criteria questions this. If absence of nervousness is equivalent to low trait anxiety, Cleckley's use of this exclusionary item means that his psychopaths were nonanxious by definition. Newman (1998) found that predictions from his hypothesis of a response modulation deficit in psychopaths are supported only when PCL-R psychopaths with high anxiety are excluded, and Schmitt and Newman (1999) found that PCL-R and F1 and F2 scores were all independent of trait anxiety or fearlessness. They conclude that the PCL-R does not measure primary psychopathy.
Consistent with this conclusion, we also found an absence of any inverse relationship between the PCL-R and Axis I anxiety or mood disorders (Blackburn, Logan, Donnelly, & Renwick, 2003). In another study, cluster analysis of the DSM-III personality disorders among violent offenders indicated three groups who scored equally highly on the PCL-R and its two subfactors (Blackburn & Coid, 1999). One group was primarily narcissistic and reported the lowest levels of Axis I anxiety and mood disorders. Another displayed borderline, avoidant, and other personality disorders associated with trait anxiety and reported the highest levels of anxiety and mood disorders. These two groups justify the distinction between primary and secondary psychopaths. These findings seem to reflect a failure of the PCL-R to provide an unbiased measure of Cleckley's construct. It could be argued, however, that the empirically derived PCL-R is a better measure of psychopathy and that Cleckley's inference of an emotional deficit in psychopaths was an artifact of excluding nonanxious people from his psychopathic category.
Our distinction suggests that primary and secondary psychopaths are phenotypically distinct groups having a similar extreme position on the psychopathy dimension, but distinguished by opposite extremes on an orthogonal dimension of anxiety. Some writers, however, make a genotypic distinction (see Skeem et al., 2003). Mealey (1995), for example, proposed two types falling at extremes of a continuum of "sociopathy," a product of evolutionary pressures manifested in predatory social interactions associated with emotional unresponsiveness. Primary sociopaths are a small but stable number of "cheaters" selected for in every culture through frequency dependent selection. Secondary sociopathy reflects a less extreme position on the genetic continuum and is the outcome of environmental conditions in which criminal behavior as a cheating strategy is more likely among individuals at a competitive disadvantage. In an addendum, Mealey suggested that primary sociopathy is reflected in F1 of the PCL-R, and secondary sociopathy in F2. However, as these are correlated variables, it is difficult to reconcile this suggestion with her proposal of two discrete types on a single continuum.
Porter (1996) proposed a similar etiological distinction between fundamental psychopathy, an inability to form interpersonal bonds together with a lack of empathy and conscience resulting from genetic predisposition, and secondary psychopathy, where the same outcome results from traumatic interpersonal experiences producing dissociation of affect. However, as Porter noted, etiological theories specifying different pathways to the same phenotypic outcome are untestable except by longitudinal research. Moreover, while such theories have implications for primary intervention, phenotypic discrimination of psychopaths based on variation along theoretically relevant personality dimensions has more immediate implications for construct validity, prediction, and secondary intervention.
DSM-II described antisocial personalities as chronically antisocial, lacking in personal or group loyalties, callous, hedonistic, emotionally immature, and lacking responsibility or judgment. The DSM-III criteria, however, marked a return to the moral insanity tradition. They were a list of socially undesirable activities virtually devoid of reference to personality, prompting Millon (1981) to describe the category as "an accusatory judgment rather than a dispassionate clinical formulation" (p. 181).
These criteria were influenced by Robins (1978), who defined psychopathy as failure to comply with social norms, regardless of the psychological substrate, a definition that would equate all forms of social dissent with psychopathy (Levenson, 1992). More than other personality disorder categories, APD illustrates the folly of operational definitions (Leahey, 1980). Operational definitions are neither "operational" as originally conceived, nor are they "definitions" in a meaningful sense because the definition comes from the concept that binds the criteria. Although Robins (1978) appeared to accept Cleckley's concept, she defended her definition of psychopathy by arguing that "we rely on behavior to infer the psychological substrate" (p. 256). However, our prior assumptions about the substrate determine the behaviors we attend to in the first place. The substrate determining Robin's criteria does not clearly reflect Cleckley's construct. The related notion of antisociality similarly lacks any clear theoretical construct from which the manifestations are derived.
The DSM-IV criteria moved slightly in the direction of Cleckley, or more precisely, the PCL-R. Antisocial personality disorder is considered synonymous with psychopathy and is "a pattern of disregard for, and violation of the rights of others" (American Psychiatric Association, 1994, p. 629). The operational criteria defining APD are a mixture of socially deviant behavior (illegal behaviors, lying, physical fights, inconsistent work behavior, early conduct disorder) and concrete indicators of the traits of impulsivity, deceitfulness, recklessness, and lack of remorse. DSM-IV notes that individuals with APD frequently show traditional psychopathic criteria of callousness, arrogance, lack of empathy, or glib, superficial charm. However, because only three of the formal criteria are required for diagnosis, many individuals who meet the criteria for APD frequently do not in practice show these characteristics. Antisocial personality disorder is clearly a measure that is poorly coordinated with Cleckley's underlying construct and an illustration of conceptual drift.
The continued emphasis in APD on antisocial acts as primary manifestations of psychopathy fails to capture the traditional defining personality traits for two reasons. First, it confounds the two conceptual domains of personality and social deviance or antisociality. These are not competing conceptualizations of the same phenomenon as some maintain (e.g., Skilling et al., 2002) because they reflect different frames of reference and different areas of inquiry. The domain of personality is stylistic variations between people, while social deviance refers to departures from cultural and moral standards (Blackburn, 1988). It is of more than passing interest to inquire how personality contributes to persistent social rule-breaking. However, to define the former in terms of the latter confounds the dependent with the independent variable and precludes any understanding of the relationship.
Second, specific acts have multiple determinants and are poor criteria of a stable underlying disposition (Blackburn, 1995; Block, 1989; Gottfredson & Hirschi, 1990). Criminality or antisociality may be construed as a disposition of the person, but antisocial acts are only one of several possible manifestations of this disposition. Gottfredson and Hirschi (1990) conceptualized this disposition as a tendency to seek short-term, immediate pleasure, a disposition manifest in stable traits such as impulsivity or recklessness, which APD specifies. This same tendency is held to underlie using drugs, being in accidents, and smoking, but these are events arising from opportunities and external conditions as much as the attributes of the person. If antisocial behavior is simply a possible but not inevitable outcome of psychopathy, then social deviance is neither necessary nor sufficient to the definition.
It is not surprising that there is a close correspondence between APD and F2 of the PCL-R. However, the PCL-R criteria overall are more clearly weighted with Cleckley's traits, and, at best, APD provides no more than a crude approximation to Cleckley's concept. In terms of psychometric utility, the balance of evidence also suggests that the PCL-R is more valid for practical purposes than APD (Cunningham & Reidy, 1998). Nevertheless, both APD and the PCL-R are biased representations of the construct they purport to represent.
Was this article helpful?