Prrw~t.
;ndik/.
/I#
Vol.
Copyright
19, No. 5. pp. 741-752. 01995
Printed in Great Britain.
0191-8869(95)00087-9
IS THE PSYCHOPATH R. J. R. Blair’.‘,i,
Elsevier
Saence
1995 Ltd
All rights reserved
0191-X869/95
$9.50 + 0.00
‘MORALLY INSANE’?*
L. Jones,’ F. Clark,“ and M. Smith”
‘MRC Cognitive Development Unit, 4 Taviton Street, London WC I H OHT, ‘Department of Psychology, University College London, 25 Bedford Way, London WC I, ‘HMP Wormwood Scrubs, DuCane Road, London WI2 OAF, ‘Broadmoor Hospital, Crowthorne, Berkshire RGI I 7EG, England (Received 2 November
1994)
Summary-Very early accounts of Psychopathic Disorder explained the disorder in terms of adeficit within some form of moral faculty or structure (Pritchard, 1837; cited by Blackbum, British Journal of Psychiatq, 1.53, 505-512, 1988). However, such accounts fail to specify the nature of the faculty/structure or its role in the control of behaviour. Blair (Cognition, in press) has suggested that a mechanism for the control of aggression is a prerequisite for the development of morality. Specifically, he has suggested that this mechanism mediates the moral/conventional distinction for transgressions. Blair has suggested that psychopaths may lack this prerequisite, predicting that these subjects should fail to make the moral/conventional distinction for transgressions. The first goal of the present study was to replicate the findings of an earlier study (Blair, Cognirion, in press) confirming this prediction. The second goal of the present study was to extend the earlier findings, to examine the influence of the proposed mechanism on the observed moral/conventional distinction for positive acts (Smetana, Bridgeman & Turiel, The Nurure of Prosocial Development: Inrerdisciplinq Theories and Strategies. New York: Academic Press, 1983). The results closely replicated those of the earlier study for the transgression moral/conventional distinction; the psychopaths, in contrast to the controls, were failing to make this distinction. However, diagnosis did not predict performance for the positive act moral/conventional distinction; the psychopaths and the controls responded to these items similarly. The results are discussed with reference to theories of the development of the psychopath and theories of the development of morality.
INTRODUCTION
The origins of Antisocial Personality Disorder (APD; APA, 1987) have been traced back to Pritchard’s (1837; cited by Blackburn, 1988) concept of moral insanity (Blackburn, 1988; Maughs, 1941). Pritchard attempted to explain socially objectionable behaviour by reference to moral “perversion”. Indeed, as Blackburn (1988) notes, much of the debate in the 19th century centred on how a diseased “moral faculty” could explain criminal behaviour. Even relatively recently Henderson (1955) stated that a “disease in a person’s moral structure may constitute a disease as truly as some form of physical involvement”. Of course, stating that APD/psychopathy is due to a diseased moral faculty or structure does not help in understanding the disorder unless the nature and the functioning of the faculty/structure is specified in detail. In particular, it would be necessary to specify the role of the faculty/structure in the control of violent behaviour. Without such specification or, for that matter, without an account of why damage to (or abnormal development of) a specific element of the cognitive system should result in a tendency to commit moral transgressions, it is perhaps reasonable to argue that the disorder should be dismissed as a myth (Blackburn, 1988; Karpman, 1948). However, a recent account has suggested that the behaviour of the psychopath is a consequence of the lack of a specific cognitive mechanism which is hypothesized as being crucial for the control of violent behaviour (Blair, in press). Blair’s suggestion was prompted by the work of the ethologists Eibl-Eibesfeldt (1970) and Lorenz (198 1) who proposed that most social animals possess mechanisms for the control of aggression. They noted that submission cues displayed to a conspecific aggressor terminate attacks; an aggressor dog will cease fighting if its opponent bares its throat. Blair (in press) proposed the existence of a functionally similar mechanism in humans (a Violence Inhibition Mechanism; VIM) and speculated that a deficit within, or a failure to develop, this mechanism might result in the psychopathic disorder; the individual without this mechanism would not inhibit his aggression when the victim displayed distress cues. In addition, Blair suggested that this mechanism was a prerequisite for the development of certain aspects of morality; notably, for the development of the moral/conventional distinction. *Portions of this article were presented at the 100th Mewing tTo whom all correspondence should be addressed.
of the Americun
741
PsychiatricAssociation,Philadelphia,
May 1994.
142
R. J. R. Blair et ul
Within the literature on the moral/conventional distinction (see, for reviews, Smetana, 1993; Turiel, Killen & Helwig, 1987) moral transgressions (e.g. hitting another, damaging another’s property) are defined by their consequences for the rights and welfare of others. Social conventional transgressions (e.g. dressing in opposite sex clothes, talking in class) are defined as violations of the behavioral uniformities that structure social interactions within social systems. Children and adults generally state that moral transgressions are more serious than conventional transgressions or rank them as more serious than conventional transgressions (e.g. Nucci, 1981; Smetana, 1985; Smetana & Braeges, 1990). In addition, and more importantly, modifying the rule conditions (by an authority figure removing the prohibition against the act, for example) only affects the permissibility of conventional transgressions. Ss generally judge moral transgressions as non-permissible even if there is no rule prohibiting the action. If there is no rule prohibiting a conventional transgression (e.g. talking in class), Ss generally judge the act as permissible. S’s seriousness judgements do not necessarily distinguish the domains; children have been found to judge some conventional transgressions as serious as some moral transgressions (Stoddart & Turiel, 1985; Turiel, 1983). However, subject’s modifiability judgements do distinguish the domains; in the Stoddart & Turiel(1985) and Turiel(1983) studies even those children who judged the conventional transgressions as serious as the moral transgressions judged the moral transgressions as less rule contingent and less under authority jurisdiction than the conventional transgressions. The moral/conventional distinction is found in the judgements of children from the age of 39 months (e.g. Smetana & Braeges, 1990) and across cultures (e.g. Nucci, Turiel & Encarnacion-Gawrych, 1983; Song, Smetana & Kim, 1987). Blair (in press) suggested that the Violence Inhibition Mechanism (VIM) mediates the moral/conventional distinction. Blair suggested that repeated pairing of representations of the transgression with the distress cues of the victim of the transgression result in representations of the transgression becoming, through classical conditioning, conditioned stimuli for the activation of VIM. Conventional transgressions will rarely come to activate VIM since, by definition, they rarely result in victims/distress cues.* Certainly, when individuals reason about conventional transgressions they rarely make reference to another’s welfare (Turiel, 1983). Blair suggested that the inhibitory consequences (effectively a withdrawal) following the activation of VIM are experienced, through meaning analysis, as aversive [following Mandler’s (1984) position on value]. He suggested that it was this sense of aversion to the moral transgression activating VIM that resulted in the act being judged as bad. Manipulations of the transgression’s context (i.e. stating that there is no rule against the transgression) would not alter the activation of VIM by the details of the transgression itself. Thus, according to this position, the transgression would still be judged bad. Conventional transgressions would not generate this sense of aversion. They are defined as transgressions only by the presence of rules. Removal of the rule, by modifying the transgression context, and the transgression should no longer be judged as bad. Most previous investigations of the moral reasoning of the psychopath have been conducted within the paradigmatic framework of Kohlberg who described a series of developmental ‘levels’ in the *There have been several suggestions that the moral/conventional distinction is flawed as certain conventional transgressions may result in victims (e.g. Light, 1993; Rest, 1983; Shweder, Mahapatra & Miller, 1987); for example, driving on the right side of the road in the U.K. may result in other victims besides the self. In addition, there are acts which are considered to result in harm in one culture which are not considered to result in harm in another. For example, eating chicken the day after the death of one’s father is not considered to have any harmful implications in Western culture. In contrast, for the Brahmans of Orissa, India, this act is considered to harm the father’s soul (Shweder et al., 1987). However, because of this difference in considering whether the act results in a victim, the Brahmans treat the act as a moral, not aconventional, transgression (Turiel et al., 1987). Similarly, Smetana (1982) has shown that whether an individual treats abortion as a moral transgression or as a conventional transgression is determined by whether he/she judges the act to involve a victim or not. Thus, it is correct to consider that the moral/conventional distinction is flawed at a philosophical level; all transgressions commonly regarded as conventional can be considered to result in victims and all transgressions commonly regarded as moral can be considered to not involve victims. However, it is wrong to consider that the moral/conventional distinction is flawed at a psychological level. As Smetana (I 985) shows, if an individual judges (learns that) a transgression causes victims he will process that transgression as moral, if he judges (learns that) a transgression does not cause victims he will process that transgression as conventional. The implication is that driving on the right-hand side of the road is considered as a conventional transgression in England because the potential victims of this act are not taken into consideration. Should these potential victims be taken into consideration the act will be treated as a moral transgression (cf. Davidson, Turiel & Black, 1983). The moral/conventional distinction could only be considered to be flawed at a psychological level if transgressions could be identified that individuals judged caused victims but which they processed as conventional. However, when Ss are asked why they had considered acts which they had processed as conventional as wrong they rarely mention the welfare of others (Smetana, 1993).
Is the psychopath
‘morally insane’?
743
child’s moral reasoning (Kohlberg, 1969; Colby & Kohlberg, 1987); in effect, a description of the changes in the child’s theories of morality with age. While it appears clear that the moral reasoning of delinquents is at a lower level than that of normal controls* (see Blasi, 1980; Trevathan & Walker. 1989) it is more debatable whether the moral reasoning of psychopaths is at a lower level than that of criminal controls. Fodor (1973) found that the moral reasoning of psychopathic youths was at a lower level than the moral reasoning of other delinquents. Campagna and Harter (1975) found the moral reasoning of sociopaths to be lower than that of non-incarcerated normals, even when controlling for mental age. Jurkovic and Prentice (1977) found that psychopaths give evidence of less mature moral reasoning than other groups of delinquent and normal youths. However, Lee and Prentice (1988) only found that delinquents responded at a lower level than non-delinquents; the psychopaths did not reason at a lower level than the other delinquent groups. In addition, while Trevathan and Walker (1989) observed a tendency for the psychopaths to reason at a lower level than non-psychopathic controls this was not significant. However, they did observe that both groups of delinquents scored at a significantly lower level than non-incarcerated controls. Thus, while it is clear that criminal groups may reason at a lower level than non-criminal controls, it is uncertain whether the moral reasoning of psychopaths is lower than that of other criminal groups. Only one previous study has examined the moral/conventional distinction in psychopaths (Blair, in press). This study confirmed two hypotheses of the VIM position: that the psychopath was less likely to make the moral/conventional distinction and that he was less likely to use other-person based reasoning in his theories about transgressions. The first goal of the present study was to attempt to replicate the previous findings as the Ss numbers in the previous study had been small (N = IO in each group). The second goal of the present study was to extend the previous findings. The moral/conventional distinction is made in children and adults for both transgressions and positive actions. For example, Smetana, Bridgeman and Turiel (1983) found that moral positive actions (e.g. comforting a young child) were ranked by Ss as more good to do than conventional positive actions (e.g. wearing the school’s uniform). The VIM position makes no direct predictions about the moral/conventional distinction in positive actions neither does it make specific claims as to why we should approve of moral positive actions. It could be argued that moral positive actions are evaluated as good if they result in the termination of a situation evaluated as bad because of the activation of VIM (i.e. a distressed other). In this case, it would follow that if VIM is a prerequisite for the development of the moral/conventional distinction in positive acts, and if the psychopath lacks VIM, then the psychopaths should fail to make the moral/conventional distinction for both transgressions and positive acts. Alternatively, it could be argued that VIM has no role in the development of approval for moral positive acts, that VIM is only involved in the development of disapproval for transgressions. In this case, it would follow that even if the psychopath lacks VIM, he will reason about moral positive acts similarly to non-psychopathic controls.
METHOD
Design The experiment involved a 2 X 2 X 2 factorial design. The independent variables were the two different subject groups (psychopaths and non-psychopath incarcerated controls), the two different domains of story (moral and conventional) and the two types of act (transgression and positive). The dependent variable was the S’s responses to the questions about the transgression situation. Subjects All the Ss were resident in either Broadmoor Special Hospital or Wormwood Scrubs prison. The files of 130 patients in Broadmoor Special Hospital and 150 men incarcerated in Wormwood Scrubs prison were read to obtain a Revised Psychopathy Checklist (PCL-R) score in accordance with the guidelines of Hare (199 1). These files were extensive and contained psychiatric, psychological and social worker reports. Wong (1988) has shown that PCL-R scores derived entirely from file data can be reliable. In addition, the PCL-R scores of twenty-two inmates of Wormwood Scrubs prison were *Though the reasons for this are not (see Blasi, 1980).
R. J. R. Blair et al.
744 Table
I.
Means
for each of the subject
GOUp
IQ:
Psychopath Non-psychopathic
PCL-R
control
score = score
on Hare’s
criteria
(standard
(WAIS-R)
deviations
in parentheses) PCL-R
Age
score
92.5
30.35
(15.871
(7.24)
CI .62)
97.90’
31.25
II.20
( 12.49)
(8.10)
(2.95)
(199 I) Revived
Psychopathy
32.00
Checklist
also calculated by a clinician resident in this institutions. The inter-rater agreement with the file ratings of the first author was 0.94. There was 100% agreement in category (psychopath or not) assignation. Ss receiving a score of 30 or higher were classified as psychopaths (range 30-34), and Ss scoring 20 or below were classified as non-psychopaths (range 4-l 9). This procedure resulted in the generation of a list of potential psychopathic and non-psychopathic Ss. 20 psychopaths and 20 non-psychopathic controls, selected from this list, participated in the present study. An additional 7 psychopaths and 4 non-psychopathic controls, selected from this list, refused to participate. The groups were 1Q and race matched (all Ss were D/E social class). All of the Ss were serving life sentences for murder/manslaughter and had been incarcerated for at least 18 months. None of the Ss were, or had been, psychotic or taking psychotropic medication. All the Ss were male. Full S characteristics are shown in Table 1. Materials The stories used to measure the moral/conventional distinction were all taken from the developmental literature. The four moral transgression stories involved a child hitting another child, a child knocking another child down in the playground, a child smashing a piano with a hammer and a child taking another child’s bag. The four conventional transgression stories involved a boy wearing a skirt, two children talking in class, a child walking out of the classroom without permission and a child who stops paying attention to the lesson and turns their back on the teacher. The four moral positive stories involved a child sharing food with another child, a child comforting another child who was sad, a child returning a skateboard to the child who had lost it and a child giving money to charity. The four conventional positive stories involved a child putting his hand up to talk in class, a child wearing the school’s uniform, a child joining the back of the queue for lunch and a child putting his bag in his own locker. The Ss responses to the questions were recorded on standard scoring sheets. Procedure Ss were tested in interview rooms attached to the S’s ward. Before the study commenced the Ss were introduced to the experimenter and informed about what they were to do. S consent forms were taken. Before any of the transgressions scenes were read to the Ss, they were informed that all of the scenes would occur within a school environment. It was decided to place the transgressions scenes within a school environment, as opposed to a ward or other adult environment, because piloting had shown that teachers were regarded by the Ss as legitimate authority figures for children. Some Ss did not regard nurses for example, as legitimate authority figures for other adults. Each of the transgression scenes was read out to the S one at a time. The order of presentation of the transgression scenes was randomized across Ss. After the transgression scene had been presented, the S was asked four questions: (1) “Was it right or wrong for X to do Y?” (examining the S’s judgement of the permissibility of the act). (2) “On a scale of one to ten, how right [or wrong depending on the answer to (l)] was it for X to do Y?” (examining the S’s judgement of the seriousness of the act). (3) “Why was it right [or wrong depending on the answer to (l)] for X to do Y?” (examining the S’s theories about the act). If the S had said that the act was wrong in answer to question
(l), the S was then told:
Is the psychopath ‘morally insane’?
745
Table 2. A description of the justification categories Justification category
Description
Other’s welfare
Any reference to the welfare of the victim (e.g. “It will hurt him”; “Prevents the other feeling hungry”) Any reference, even implicit (e.g. “It’s not acceptable to do that”; “Because it’s the honest thing to do”), to rules Any reference to the rudeness of the act (e.g. “It’s bad manners”: “It’s good manners”) Any reference to the disruption caused by the transgression (e.g. “It will distract the class”; “It will not disturb the others”) Any reference to the act showing/proving something (e.g. “It shows honesty”: “It shows kindness”) Any other response
Normative references Rudeness Disorder statements Shows/proves Other
“Now what if the teacher said before the lesson, before X did [the transgression], anybody can Y if they want to. Anybody can Y.”
that “At this school
and then asked a final question: (4) “Would it be O.K. for X to do Y if the teacher says X can ?” (examining jurisdiction). All responses
were recorded
by hand on a standard
scoring
the rule’s authori
sheet.
Scoring procedure The scoring procedure followed that commonly used in the literature (e.g. Smetana, 1981; Smetana & Braeges, 1990). The answers to all questions, except (2), were scored categorically. Yes responses were assigned a score of 0, and no (not OK) responses a score of 1. Ss could thus achieve a cumulative score of between 0 and 4 for each of the domains for each of the questions. Question (2) was scored according to the value (between 1 and 10) the S had given that transgression. The justifications of the Ss were scored according to categories similar to those used in previous research (e.g. Smetana, 1985). The justification categories are shown in Table 2. Two coders scored all justifications, and inter-rater reliability was high 94%.
RESULTS Table 3 presents the means and standard deviations of moral and conventional judgements for questions (l), (2) and (4) for both S groups for the transgression and positive act stories respectively. 2(Domain) X 2(Group) ANOVAs were performed on the S’s responses for each of the three questions for both the transgressions and the positive acts. The results of the ANOVAs for the transgressions stories closely replicated those previously obtained in the earlier study (Blair, in press). The three ANOVAs revealed main effects of domain for all three judgements: permissibility, F( 1,79) = 10.88, P = 0.005; seriousness, F( 1, 55)* = 5 1.92, P < 0.01, and: authority jurisdiction (modifiability), (F( 1,79) = 43.5, P < 0.001). Moral transgressions were judged significantly less permissible, more serious and less authoriry dependent than conventional transgressions. There were no significant group differences for any of the three judgements and only one significant domain X group interaction which was for the authority jurisdiction judgement [F( 1,79) = 11.76; P = O.OOl]; see Table 3. Moreover, and very interestingly, a two way ANOVA on the aurhoriry jurisdiction judgement for the moral transgressions revealed that the psychopaths were more likely to judge the moral transgression permissible than were the controls [F( 1,39) = 3.90; P < 0.061; see Table 3. However, as in the earlier study (Blair, in press), a simple effects analysis using two way ANOVAs to examine the moral/conventional distinction of the two groups independently revealed the real differences in the response patterns of the two groups. The non-psychopathic controls made a significant moral/conventional distinction on all three criterion judgements [permissibility P < 0.05; seriousness F( 1,27) = 13.7, P < 0.005; authoriiy jurisdiction F( 1,39) = 5.52, [F( 1.39) = 45.88, P < O.OOl]. However, the psychopaths did not make a significant moral/conven*This question
was not administered
to 7 of the psychopaths
and 7 of the non-psychopathic
controls.
R. J. R. Blair er al.
746
Table 3. The means and standard deviations of the criterion judgements for both transgressions and positive acts
Permissibility M C
Group
0.98 (0. I 1)
Psychopath Non-psychopathic
0.04 (0.09) 0.01 (0.06)
0.03 (0.08) 0.03 (0.08)
1a0
control
Psychopath Non-psychopathic
(0.00)
0.89 (0.22) 0.93 (0.14)
control
for both subject groups and
Criterion judgement Seriousness C M Tranrgresslona 37.07 (5. IO) 35.36 (5.56) Positive acts 37.50 (4.24) 34.36 (5.01)
30.00 (6.59) 25.50 (8.27)
Authority M
0.83 (0.36) 0.99 (0.06)
juriadictlon c
0.68 (0.34) 0.5 I (0.31)
36.29 (3.60) 34.09 (3.59)
M = Moral; C = Conventional
tional distinction on either the permissibility [F( 1,39) = 2.48. P = n.s.1 or, crucially, the authority jurisdiction [F( 1,39) = 0.183, P = ns.]. Though, the psychopaths did make a significant moral/conventional distinction on the seriousness judgement [F( 1,27) = 10.08; P < 0.00.5]. An analysis of the individual subject data on the u&ho&y jurisdiction question reveals the difference in the pattern of responding of the two S groups even more clearly (see Table 4). Table 4 shows how many of the Ss made the moral/conventional distinction on this item. As can be seen from this table it was not that all of the psychopaths were making less of a moral/conventional distinction than all of the controls but that many of the psychopaths (11 out of 20) were not making a moral/conventional distinction at all. In contrast 13 out of the 20 non-psychopathic controls were making a clear moral/conventional distinction on this item. A 2 (group) X 3 (distinction class) Chi Square confirmed these apparent group differences (;r’ = 8.13; d.f. = 2; P < 0.05). As regards the positive act stories, the picture was rather different (see Table 3). Against predictions, the ANOVAs for permissibility and seriousness revealed no domain effects but there was one significant group effect for the seriousness judgement [F( 155) = 7.79; P < 0.051; surprisingly, the psychopaths rated both moral and conventional positive acts as significantly better to do than the non-psychopathic controls. As can be seen in Table 3, it is apparent that neither group were making a moral/conventional distinction for these positive stories for these questions. JustiJication
categories
Table 5 shows the Ss’ proportionate use of justifications for the (combined) moral and conventional items for the transgression and positive act stories. As regards the transgression stories, normative references were the dominant mode of responding for both groups for both moral and conventional transgressions. As regards domain differences, other’s welfare reasoning was more commonly used, at least by the non-psychopathic controls, to justify moral items while rudeness was more commonly used, to justify conventional items. Indeed, a 2(group) X 2(domain) ANOVA, performed on the other’s welfare justification category, revealed a main effect for domain [F(1,79) = 15.44; P < O.OOl]. This ANOVA also revealed a main effect of group (F( 1,79) = 10.54; P = O.OOS]. As found in the earlier study (Blair, in press), psychopaths are significantly less likely to justify items by references to the other’s welfare. There was also a significant group by domain
Table
4. The
number
No distinction
Group Psychopath Non-psychopathic
of psychopaths and non-psychopaths in each moral/conventional distinction’ categories
control
II 3
of the
‘quality
Quality of distinction Mild distinction
Clear distinction
4 4
5 I3
of
No distinction = The moral transgressions were judged in the same way as the conventmnal transgressions; mild distinction = 1 more conventional than moral transgression was judged authori dependent; clear distinction = 2 or more, conventional than moral transgressions were judged authority dependent.
Is
the
psychopath
‘morally
747
insane’?
Table 5. The proportionate use by the psychopath and the non-psychopathic controls of the justification categories for both the transgreaaions and the positive acts (in percentages)
Moral
Psychopath Conventional
Non-psychopathlc control Moral Conventional
Welfare Norms Rudeness Disorder Shows/proves Other
3.15 92.5 0.00 I .25 0.00 2.50
Transgressions 0.00 62.5 21.25 8.75 1.25 6.25
21.5 66.25 0.00 2.50 0.00 3.15
2.5 52.5 21.25 3.75 7.50 12.50
Welfare Norms Rudeness Disorder Shows/proves Other
16.25 27.50 0.00 0.00 40.00 16.25
Positive acts 0.00 57.50 11.2s 26.25 3.75 1.25
25.0 3x.75 0.00 0.00 28.75 7.50
I .25 50.00 1375 I5.00 6.25 13.75
Norma = Reference to normative references.
interaction [F( 1,79) = 8.43; P = 0.011. This was a product of the fact that this difference between the groups was only present for the moral items; no conventional items were justified through references to other’s welfare. Examination of other’s welfare justification by the individual S for the transgression stories revealed that 2 of the psychopaths and 11 of the non-psychopaths used this form of justification at least once. Of the 2 psychopaths who used this justification, 1 made a clear moral/conventional distinction (see Table 4) while the other did not. 8 out of the 11 non-psychopaths who used the other’s welfare justification made a clear moral/conventional distinction. As regards the positive act stories, the picture was again different (see Table 5). Again normative references were the dominant response pattern for both groups for both types of positive act. As regards domain differences, reference of other’s welfare and what was being shown/proved (the actions were good because they “showed honesty/kindness to others”) were more commonly made by both groups to moral positive acts. Both groups of Ss justified conventional transgressions with reference to rudeness or disorder. A 2(group) X 2(domain) ANOVA, performed on the other’s welfare justification category, revealed a main effect for domain [F( 1,79) = 23.44; P < O.OOl]. Both groups of Ss were more likely to refer to the other’s welfare when reasoning about a moral positive act than a conventional positive act. However, there was neither a significant group effect [F( 1,79) = 1.62; n.s.] nor a significant group by domain interaction [F( 1,79) = 0.82; n.s.1. Indeed, a two way ANOVA comparing the use of the other’s welfare justification category by the psychopaths for the moral transgression and positive act stories showed that the psychopaths were significantly more likely to make other’s welfare justifications for the positive act stores than for the transgression stories [F( 1.39) = 4.97; P < 0.051. The lack of a group difference in the use of the other’s welfare justification is particularly noticeable when analysing the individual Ss; 9 out of the 20 psychopaths vs 10 out of the 20 control Ss made other’s welfare justifications for the positive act stories. An additional analysis was performed on the other’s werfare justifications of both groups for the positive act stories. According to the VIM model, individuals make reference to other’s welfare when justifying why moral transgressions are bad because the victim’s distress cues are the stimuli which activate the withdrawal response which results in the evaluation of badness (see Blair, in press). However, individuals may make reference to other’s welfare when justifying why moral positive acts are good in two ways: by referring to the pain of the individual who is about to be helped or referring to the pleasure of the individual once he has been helped. For example, when justifying why comforting another is good the S may say either that it is good because the person is sad or that it is good because the person will be made happier. The VIM model predicts only that those without VIM would be less likely to make reference to a person’s sadness than individuals with VIM; there is no reason why individuals without VIM should not make reference to a person’s happiness on being helped. Thus, the other’s welfare justifications of both the psychopaths and the non-psychopathic controls were re-analyzed according to the focus of the concern for the other; whether it was other’s pain or other’s
748
R. J. R. Blair et al. Table 6. Number of ues by the psychopaths and pleasure-based other’s we&are Pain-based
Group Psychopath Non-psychopathic
and the non-psychopathic controls of pain-based reasoning for the positive moral acts justification
Pleasure-based 7 4
6 16
control
Justification
pleasure based. The results of this analysis are presented in Table 6. As can be seen the psychopaths were significantly less likely than the controls to make other’s pain based justifications. However, they were as likely as the controls to make other’s pleasure based justifications. A 2 (group) X 2 (justification type) Chi Square confirmed these apparent group differences (xl = 4.062; d.f. = 1; P < 0.05). Moreover, an analysis by individual S for the psychopaths revealed that 3 of the 4 Ss who made pain-based other’s welfare justifications were Ss who had made at least a mild moral/conventional distinction (see Table 4). In contrast, 3 of the 6 psychopaths who made pleasure-based other’s welfare justifications were Ss who had not made a moral/conventional distinction at all. Individual
item PCL scores and the moral/conventional
distinction
In the earlier study (Blair, in press), the S’s scores for both groups on various items of Hare’s (1985) Psychopathy Checklist were compared with various measures of their moral/conventional distinction. This form of analysis was also conducted here. Table 7 shows those PCL-R (Hare, 1991) items which significantly correlated with quality of the moral/conventional distinction (as indexed by the authority jurisdiction question), amount of other’s werfare justifications for the transgression stories and amount of other’s welfare justifications for the positive act stories and their intercorrelations. Table 7 shows 4 significant correlations of individual PCL-R items with quality of the moral/conventional distinction and amount of other’s welfare justifications for the transgression stories. Lack of remorse/guilt correlated with both measures. Early behavioral problems correlated with amount of other’s welfare justifications for the transgression stories. Criminal versatility correlated with quality of the moral/conventional distinction. All of these correlations were also significant in the earlier study (Blair, in press). There were no significant correlations with amount of other’s welfare justifications for the positive act stories. In addition, it can be seen that total PCL-R score, quality of the moral/conventional distinction and amount of other’s we&u-e justifications for the transgression stories all correlated significantly with each other. None of these items correlated significantly with tendency to make other’s welfare justifications for the positive act stories.
DISCUSSION
The present study examined the form of the moral/conventional distinction for transgressions and positive acts made by psychopaths and non-psychopathic controls and the categories used by these Ss when they justify their judgements. The study revealed: First, and in line with the findings of the earlier study (Blair, in press), that while non-psychopaths did make the moral/conventional distinction for transgressions stories, psychopaths did not. Secondly, and in line with the results of the earlier
Table 7. Individual items on the PCL-R which Ggnificantly (P < 0.01) correlated with quality of the moral/conventional distinction, tendency to make other’s wr&ve justifications for the transgression stories and tendency to make othrr’swe/&e juqtificationa for the positwe act stories PCL-R Item Lack of remorse/guilt Early behav. problemr Revocation Criminal versatility Total PCL-R score N other’s welf~ve for Ts N other’s wdfure for Ps
Quality of MvC distinction - 0.462 - 0.470 n.s. “.S. - 0.456 0.418 “S.
N other’s
uelfbrefor - 0.433 “.S. - 0.418 - 0.409 - 0.407 I .oo “.S.
Ts
N other’s
we/firefor
Ps
n.5 n.s “.S. “.S. “.F. “.S.
I .oo
Quality of MvC Distinction = quality of the moral/conventional distinction; N other’s we/jbre for Ts = number of other’s welfare justifications for transgression stories; N other’s welfarefor Ps = number of other’s welfare justifications for the positive act stories: Early behav. problems = early behavioural problems; Revocation = revocation of conditional release.
Is the psychopath ‘morally insane’?
749
study, that psychopaths were much less likely than non-psychopathic controls to justify their responses to moral transgression stories with reference to other’s welfare. Thirdly, that the quality of the S’s moral/conventional distinction and the S’s tendency to make reference to other’s welfare when justifying moral transgressions were not only correlated with the total PCL-R score but also with individual items of the PCL-R; in particular, with the item “Lack of Remorse/ Guilt”. Fourthly, that psychopaths and non-psychopathic controls did not differ in either theirjudgements of, orjustifications of, moral and conventional positive acts. In summary, therefore, this study closely replicated the findings of the earlier study for moral and conventional transgressions (Blair, in press) but failed to show any relationship between psychopathy and responding to positive moral and conventional items. It should be noted that while the results for the transgression stories were closely similar to those obtained in the previous study, they would not be expected from an analysis of the developmental literature. The observation of a moral/conventional distinction in normally developing children in adults is a robust phenomenon. It is found across ages (e.g. Nucci, 198 1) and across cultures (e.g. Song et al., 1987) and is developmentally independent of autism (Blair, submitted) and physical abuse (Smetana, Kelly & Twentyman, 1984). The results for the transgression stories are broadly in line with the VIM position. The only indication of a moral/conventional distinction in the psychopaths was for the seriousness question; the psychopaths were scoring the moral transgressions as more serious than the conventional transgressions. However, given that all the Ss were incarcerated because of actions similar to the moral transgressions described (i.e. violence, theft) and not for committing conventional transgressions (i.e. talking in class) this is perhaps less surprising. There are many sources of information for making seriousness judgements, not only the activation of VIM. Thus, as noted in the introduction, some normally developing children have been found at some ages to judge some conventional transgressions as serious as some moral transgressions (Stoddart & Turiel, 1985; Turiel, 1983). It is, as was shown by the Stoddart & Turiel (1985) and Turiel (1983) studies, the S’s responses to judgements such as authorityjmrisdiction, where the rule has been removed, which crucially indicate whether the person is making the moral/conventional distinction. This study showed that the psychopath failed to make the moral/conventional distinction under these conditions. According to Blair’s position, the activation of VIM by moral transgressions results in these transgressions being judged as non-permissible whatever the rule conditions (Blair, in press). It is suggested that conventional transgressions, being solely the products of the rule conditions, should be permissible under conditions where there is no rule prohibiting the act. The non-psychopathic controls did generally judge the conventional transgressions as permissible in the absence of rules while the moral transgressions were still considered to be non-permissible. In contrast, the psychopaths generally considered both types of transgression to be non-permissible in the absence of rules. This result, also obtained in the earlier study (Blair, in press), is not a direct prediction of the VIM account. Indeed, the prediction of the VIM account would be that psychopaths should judge the moral transgressions as permissible in the absence of rules rather than judge the conventional transgressions as non-permissible. Of course, some of the psychopaths were responding like this; the psychopaths were more likely to judge the moral transgressions as permissible in the absence of rules than the non-psychopathic controls. However, the general trend of responding was for the psychopaths to consider the conventional items as non-permissible. But, perhaps this finding is not so surprising. These Ss were all incarcerated and presumably motivated to be released. All wished to demonstrate that the treatments they were receiving were effective. They therefore would be motivated to show how they had learned the rules of society. The psychopaths manifest this desire on the authority jurisdiction criterion judgement, by suggesting that all transgressions are authority independent, Blair (in press) suggested that this was because the psychopaths lack VIM and thus are unable to identify the distinguishing features differentiating moral and conventional transgressions. This inability, coupled with a desire to demonstrate adherence to societal rules, results in their judgement of all the transgressions as authorit?; independent. The non-psychopaths, in contrast, though presumably equally motivated to be released, are incapable of ignoring the distinguishing features of moral and conventional transgressions because of the operation of VIM and thus answer the authori~jurisdiction question appropriately. The significant correlations between the individual PCL-R items, in particular item 6 (“Lack of remorse or guilt”), with the two measures, quality of moral/conventional distinction and tendency to
750
R. J. R. Blair et al
make other’s we&are justifications for the moral transgressions are very interesting. Not only were these results obtained in the previous study (Blair, in press) but this relationship was predicted by the VIM position. According to the VIM model, the individual lacking VIM will fail to show arousal to distress cues and will thus fail to develop the moral emotions (e.g. remorse, guilt, sympathy). In line with this prediction, studies have shown that psychopaths do not show arousal responses to the distress of others (Aniskiewicz, 1979; House & Milligan, 1976). It is interesting in this context to consider the distinction that has been made between primary and secondary psychopaths (e.g. Fagan & Lira, 1980; Hare, 1975; Lykken, 1957; Mealey, in press). Primary psychopaths are characterised by their lack of guilt. Secondary psychopaths are characterised by being more likely to experience guilt (e.g. Gudjonsson & Roberts, 1983; Hare, 1975). The VIM model is meant as a model of the development of the primary psychopath. This model suggests that the lack of guilt in the primary psychopath is due to the dysfunction within VIM (Blair, in press). This model predicts therefore that there should be a strong relationship between reported feelings of an absence of guilt and failure of the moral/conventional distinction. This relationship was observed in the present study. The VIM model implies, given the reports of guilt in secondary psychopaths (e.g. Gudjonsson & Roberts, 1983; Hare, 1975), that secondary psychopaths have an intact VIM. The model therefore predicts that secondary psychopaths, such as those in the Gudjonsson and Roberts (1983) study, should show the moral/conventional distinction. Indeed, it is quite possible, given the fact that primary/secondary psychopaths were not distinguished in the present study, that some of the psychopathic Ss were secondary psychopaths. This may be considered even more possible as, according to their files, three of the psychopaths were reporting guilt over their actions. Interestingly, these three Ss were three of the five psychopaths who made a clear moral/conventional distinction as defined by their performance on the authority ,jurisdiction question (see Table 4). As regards the positive act stories, the lack of a moral/conventional distinction in both groups is surprising. This result is certainly in contrast to that obtained by Smetana et al. (1983). However, the difference in results may be due to differences in methodology; Smetana et al. (1983) asked the Ss to rank the stories in order of seriousness, forcing the Ss to compare one positive act with another. In contrast, in the present study, the Ss were asked to assign a seriousness score to each of the positive act stories; there is thus less of a necessity of comparing one story with another. Though it should be noted that while almost all of the Ss were performing at ceiling on the permissibility judgement, only 5 (4 psychopaths and 1 control) were performing at ceiling on the seriousness judgement. Indeed, the lack of a moral/conventional distinction for the seriousness judgement for the positive acts was due to the fact that a substantial minority of the Ss (N = 5 in both groups) actually approved of the positive conventions more than they approved of the positive moral acts. Unfortunately, however, the lack of a moral/conventional distinction for either group makes the lack of group differences to the judgement questions more difficult to interpret. As stated in the introduction there are two alternative hypotheses concerning whether the Violence Inhibition Mechanism might be involved in the development of approval for moral positive acts. First, the Violence Inhibition Mechanism might be crucial for the development of approval for moral positive acts. According to the VIM position, activation of VIM by the distress of another should activate an arousal response which is aversively experienced. Certainly, this has been found empirically to be the case (e.g. Bandura & Rosenthal, 1966). Morally positive acts (e.g. helping the distressed individual) might be approved of because they remove the negatively valued state of someone else’s distress. If this position is correct, a moral/conventional distinction in positive acts would be expected in the controls but not in the psychopaths. However, this prediction was not confirmed. Secondly, the VIM may play very little role in the development of approval for moral positive acts. Other factors may be more important in generating approval ratings for moral positive acts; perhaps, social modelling or a positively reinforcing emotional response to smiling, gratified faces. VIM may only be involved in motivating an individual to help in a given situation; certainly, there have been frequent demonstrations that inducing empathy with a distressed other increases helping for that other (Batson, Fultz & Schoenrade, 1987; Krebs, 1975). The clearest indication that VIM may not be involved in the development of approval for moral positive acts is provided by the data on the S’s justifications of these positive acts. While psychopaths make significantly less reference to other’s welfare when justifying moral transgressions than non-psychopathic controls, there is no significant difference in the amount of reference to other’s welfare the two groups make when justifying moral
Is the psychopath ‘morally insane’?
751
positive acts. It is only when the other’s welfare justifications of the two groups for the positive acts are broken down according to their focus (on the other’s present pain or the other’s future pleasure) that a significant group difference again emerges; as was the case for the transgressions stories, the psychopaths are less likely to refer to the pain of others than the controls. It thus appears that while VIM is a developmental prerequisite for reference to another’s pain when reasoning about moral transgressions it is relatively unimportant for generating reference to another’s future pleasure when reasoning about moral positive acts. This impression is strengthened when it is considered that there is a strong correlation between an individual’s tendency to make reference to another’s pain when reasoning about a moral transgression and their quality of moral/conventional distinction on the a&ho&y jurisdicfion justification. There is no significant correlation between either of these measures and tendency to make reference to another’s welfare when reasoning about a positive act. What does this study imply about the morality of the psychopath? First, it is clear that psychopaths do represent societal rules. Indeed, the main empirical anomaly of this study is the finding that psychopaths judged conventional transgressions in the same way that normally developing Ss judge moral transgressions. Secondly, given the present and previous (Blair, in press) findings, it does appear that the psychopath fails to make the moral/conventional distinction for transgressions. The psychopath does not seem to represent the specific nature of moral transgressions, that they have victims, even though the moral/conventional distinction is made in normally developing children by the age of four (Smetana & Braeges, 1990). Thirdly, the status of the positive act moral/conventional distinction in psychopaths remains unclear. In the present study, against predictions and probably due to methodological changes, both groups failed to make the positive act moral/conventional distinction. However, it is clear that when psychopaths are justifying both moral positive acts and moral transgressions, they are less likely to consider another’s pain. AcknowledgemerIts-James Blair was supported by Wellcome Mental Health Training Fellowship 37 132/2/92/2/l .4Q. We thank the staff at Broadmoor Special Hospital and at Wormwood Scrubs H.M.P. We are also grateful to Professor J. Morton, Dr L. Cipolotti, Dr Tidmarsh and Professor. R. Blackburn for their comments on earlier drafts of this paper.
REFERENCES American Psychological Association. ( I987).l)i(r~nosri(, LI!Z~Strrfisticcrl Mnnuul cf Menral Disorders, 3rd revised edition (DSM-III-R). Washington, DC: American Psychological Association. Aniskiewicz, A. S. (1979). Autonomic components of vicarious conditioning and psychopathy. Journal of Clinical Psychology, 35, 60-67. Bandura, A. & Rosenthal. T. I. (1966). Vicarious classical conditioning as a function of arousal level. Journal of Personality and Sociul Psychology, 3, 54-62. Batson, CD., Fultz, J. & Schoenrade, P. A. (I 987). Adults’ emotional reactions to the distress of others. In Eisenberg, N. & Strayer, J. (Eds), Emparhy cmd its Development, (pp. 163-l 85). Cambridge: Cambridge University Press. Blackbum, R. ( 1988). On moral judgementn and personality disorders: The myth of psychopathic personality revisited. British Journal of Ps?chitrtry. 153. 505-S 12. Blair, R. J. R. (in press). A cognitive developmental approach to morality: Investigating the psychopath. Cognition. Blair, R. J. R. (submitted). Morality in the autistic child. Blasi, A. (1980). Bridging moral cognition and moral action: A critical review of the literature. Psychological Bulletin, 88, l-45. Campagna, A. F. & Harter. S. ( 1975). Moral judgementa in sociopathic and normal children. Journul of Personalityand Social Psychology, 3 I, 199-205. Colby, A. & Kohlberg. L. (1987). The measurement of mor~d judgement. New York: Cambridge University Press. Davidson, P., Turiel, E. & Black, A. (1983). The effect of stimulus familiarity on the use and criteria of justifications in children’s social reasoning. Brifkh Jourmrl ofDerelopmentcr1 Psychology, I, 49-65. EibI-Eibesfeldt, I. (I 970). Efhology. The biology ofbehcrviour. New York: Holt, Rinehart and Winston. Fagan, T. J. & Lira, F. T. (1980). The primary and secondary sociopathic personality: Differences in frequency and severity of antisocial behaviours. Journal ofAbnormn1 Ps~c~hology, 89, 493496. Fodor, E. M. (I 973) Moral development and parent behaviour antecedents in adolescent psychopaths. Journnl of Generic Psychology, 122. 3743. Gudjonsson, G. H. & Roberts, J. C. ( 1983). Guilt and self-concept in secondary psychopaths. Personality and Individual Differences, 4, 65-70. Hare, R. D. (1975). Psychopathy. In Venables. P. H. & Christie, M. J. (Eds), Research in P.\ychophysiology (pp. 325-348). New York: Wiley. Hare, R. D. (1985). Scoring manual for the psychopathy checklist. Unpublished manuscript. Hare, R. D. (1991). The Hove psvchopufhy checklist-rrtised. Toronto, Ontario: Multi-Health Systems. Henderson, D. (195.5). The classification and treatment of psychopathic states. Brifish Journal of Delinquency, 6, 5-14. House, T. H. & Milligan, W. L. (I 976). Autonomic responses to modeled distress in prison psychopaths. JournnlofPersonalify and Socictl Psychology, 34, 556-560. Jurkovic, G. J. & Prentice, P. M. (I 977). Relation of moral and cognitive development to dimensions ofjuvenile delinquency. Journnl of Abnormal Ps.~holo~~. 86, 4 14420.
R. J. R. Blair et al.
752
Karpman, B. (1941). On the need for separating psychopathy into two distinct types: The symptomatic and the idiopathic. Journal of Criminal Psychopathy, 3, 112-I 31. Karpman, B. (I 948). The myth of the psychopathic personality. Journal of Psychiatry, 104, 523-534. Kohlberg, L. (1969). Stage and sequence: The cognitive-developmental approach to socialization. In Goslin, D. A. (Ed.), Handbook of socialisation theory and research. Chicago: Rand McNally. Krebs, D. L. (1975). Empathy and altruism. Journal of Personality and Social Psychology, 32. I 134-l 146. Lee, M. & Prentice, N. M. (1988). Interrelations of empathy, cognition, and moral reasoning with dimensions of juvenile delinquency. Journal of Abnormal Child Psychology, 16, 127-139. Light, P. (1993). Developing psychologies. In Bennett, M. (Ed.), The child aspsychologist: An introduction to the development ofsocial cognition, (pp. 191-200). New York: Harvester Wheatsheaf. Lorenz, K. (198I). The foundation of ethology. New York: Springer. Lunney, G. H. (I 970). Using analysis of variance with a dichotomous dependent variable: An empirical study. Journal of Educational
Measurement,
7, 263.
Lykken. D. T. (1957). A study of anxiety in the sociopathic personality. Journal of Abnormal and Social Psychology, 55, 6=nlO. Mandler, G. (1984). Mind and body. New York: Norton. Maughs, S. (I 941). A concept of psychopathy and psychopathic personality: Its evolution and historical development. Journal of Criminal Psychopathology, 2, 329-356. Mealey, L. (in press). The sociobiology of sociopathy: An integrated evolutionary model. Bruin Behavioral Sciences. Nucci, L. (198 I). Conceptions of personal issues: A domain distinct from moral or societal concepts. Child Development, 52. 114-121.
Nucci, L. P., Turiel, E. & Encarnacion-Gawrych, G. E. (1983). Children’s social interactions and social concepts: Analysis of morality and convention in the Virgin Islands. Journal of Cross-Cultural Psychology. 4. 469-487. Rest, J. R. (1983). Morality. In Mussen, P. H. (Eds), Handbook of child psychology. New York: John Wiley. Shweder, R. A., Mahapatra, M. & Miller, J. G. (1987). Culture and moral development. In Kagan, J. & Lamb, S. (Eds), The emergence of morality in young children, (pp. l-83). Chicago: University of Chicago Press. Smetana, J. G. (1981). Preschool children’s conceptions of moral and social rules. Child Development. 52, 1333-1336. Smetana. J. G. (1982). Concepts of self and morality: Women ‘s reasoning about abortion. New York: Praeger. Smetana, J. G. (1985). Preschool children’s conceptions of transgressions: Effects of varying moral and conventional domain-related attributes. Developmental Psychology, 2/, 18-29. Smetana, J. G. (1993). Understanding of social rules. In Bennett, M. (Ed.), The child as psychologist. London: Harvester Wheatsheaf. Smetana, J. G. & Braeges, J. L. (1990). The development of toddlers’ moral and conventional judgements. Merrill-Palmer Quarterly,
36, 329-346.
Smetana, J. G., Bridgeman, D. L. & Turiel, E. (1983). Differentiation of domains and prosocial behaviour. In Bridgeman, D. L. (Ed.), The nature of prosocial development: Inrerdisciplinaty theories and strategies. New York: Academic Press. Smetana, J.G., Kelly, M. & Twentyman, C.T. (1984). Abused, neglected and non-maltreated children’s conceptions of moral and social-conventional transgressions. Child Development, 55, 277-287. Song, M., Smetana, J. G. & Kim, S. Y. (1987). Korean children’s conceptions of moral and conventional transgressions, Developmental
Psychology,
23, 511-582.
Stoddart, T. & Turiel, E. (1985). Children’s concepts of cross-gender activities. Trevathan, S. &Walker, L. J. (1989). Hypothetical versus real-life moral reasoning and Psychopathology, 1, 9 l-103. Turiel, E. (1983). The development of social knowledge:
Child
Development,
among psychopathic
56. 1241-1252.
and delinquent
youth.
Development
Morality and convention. Cambridge: Cambridge University Press. Turiel, E., Killen, M. & Helwig, C. C. (1987). Morality: Its structure, functions, and vagaries. In Kagan, J. & Lamb, S. (Eds), The emergence of morality in young children, (pp. 155-245). Chicago: University of Chicago Press. Wong, S. (1988). Is Hare’s Psychopathy Checklist reliable without the interview? Psychological Reports, 62, 93 l-934.