This thesis used the methods of differential, cognitive and theoretical psychology to investigate the relationships between pathological narcissism and maladaptive personality and behaviour in general, and to attempt the beginnings of construct validation of M. Scott Peck's proposed `evil' subtype of the DSM -IV Narcissistic Personality Disorder in particular. After a review of theoretical and empirical contributions to the psychology of narcissism, two empirical studies were conducted. In the first, joint self- report survey research using 338 subjects investigated the psychometric structure of narcissism, normal and abnormal personality, and constructs theoretically related to narcissism. Item -level exploratory principal components analysis and confirmatory factor analysis resulted in new subscales for the Narcissistic Personality Inventory. Scale -level exploratory principal components analysis of the combined questionnaires found evidence for a five - factor structure of abnormal personality. One of the five factors was related to narcissistic will to power and low agreeableness; another was related to narcissistic self - love and extraversion. Confirmatory factor analysis of a subset of the data found fair fit for the model.

The second study investigated the relationships between narcissism, compartmentalisation, splitting, attribution style and response to disconfirming feedback. No evidence was found to support the hypothesis that narcissists have compartmentalised self -concepts. Moderate test -retest and alternate -form reliability data were obtained for the card -sort task used to assess compartmentalisation. No evidence was found that narcissists use splitting to translocate unwanted self -aspects onto others. Narcissistic dominance was associated with claiming personal credit for positive outcomes, while narcissistic vulnerability was associated with self -blame for negative events. Multiple regression with interaction terms indicated that the relationship between narcissism and response to disconfirming feedback is a function of other personality traits such as neuroticism, and may have opposite effects in different personalities and circumstances. Zero -order correlations suggested significant roles for splitting, narcissistic grandiosity and idealisation, and 12 NEO -PI -R facet -level variables in response to disconfirming feedback.

Peck's proposed subtype of narcissistic personality disorder is theoretically distinguished from psychopathy by the use of self- deception to defend the conscience against guilt for the consequences of one's actions, rather than the absence of conscience that characterises psychopathy. Self- deception is among the most difficult and controversial topics in philosophical psychology. Theoretical chapters approached the issue with historical and literary examples, argued that splitting may be a fundamental basis of much self -deception, offered a model of malignant narcissism within the Schematic Propositional Associative Analogue Representation Systems (SPAARS) theory, and addressed the issue in the philosophy of science of whether the proposed personality disorder is a natural or social kind.