Keywords philosophy, evil, self-deception, psychopathy, narcissism, sadism Kubarych (2005) first draws on Peck (1983) to suggest a distinction between psychopaths who have no conscience and therefore no need for self-deception, and evil narcissists who use self-deception to keep the emotional consequences of their crimes out of awareness. He then draws on Davidson (1985) to emphasize a parallel between self-deception (or weakness of the warrant) where an irrational belief conflicts with the evidence, and akrasia (or weakness of the will) where an irrational intention is in conflict with one's values. Although self-deception has long been de-scribed and debated (Fingarette 2000; McLaughlin and Rorty 1988; Mele 2001), a cognitive-affective neuroscience of self-deception has become possible only recently. Such an approach includes several strands. First, contemporary information processing constructs (e.g., schemas) have been used to reframe the early insights of writers like Freud and James, and to emphasize how inattention to painful truths provide a shield against anxiety (Goleman 1997). Second, the theoretical framework of evolutionary psychology has been used to emphasize the adaptive advantages of self-deception (Lockard and Delroy 1988). Third, functional brain imaging studies have noted increased activity in executive regions (e.g., prefrontal and anterior cingulate cortices) during attempted deception but not during truthful responding, suggesting that the former requires active effort (Spence et al. 2004). The functional imaging of memory suppression suggests similar processes (Anderson et al. 2004). To date, however, there has been little functional imaging work on self-deception. Similarly, although evil has always been a key concern of philosophical thought, a cognitive-affective neuroscience of evil has emerged only in recent years (Stein 2000). Evil actions are not homogenous, and this area of research again comprises a number of different strands. First, there are actions where the primary aim is not necessarily to hurt, but where harm is in fact inflicted. (A contested, but perhaps useful example, is the battery-raising of animals for human consumption [Singer 2001]). Second, there are actions where the aim is to hurt, and where harm to the other may even be enjoyed. (During normal altruistic punishment, for example, reward circuitry is activated [de Quervain et al. 2004]). Third, there are actions characterized by impulsive aggression, and a relative absence of executive control. (There may be psychobiological distinctions between reactive aggression in response to frustration or threat, and instrumental aggression that is more goal directed [Blair 2004]). What are the implications of these two sets of cognitive-affective neuroscience literature for a philosophical consideration of evil and of narcissism? [End Page 261] It may be useful to differentiate healthy processes (e.g., having sufficient frontal cortical activity to be able to deceive others) from psychopathologic processes (e.g., not having the necessary frontal-amygdala circuitry required for empathic processing of emotions in social interaction). Similarly, it may be useful to differentiate between normal healthy narcissism (Taylor and Brown 1988) and pathologic malignant narcissism (Kernberg 1985), and between normal altruistic punishment (Hamilton 1963) and pathologic sadistic personality (Kaminer and Stein 2001). In addition, it is perhaps the case that some cognitive-affective processes that do not seem to involve psychopathology may nevertheless be seen as evil (e.g., arguably, the processes that allow a human to eat a battery-raised chicken), whereas some cognitive-affective processes that are clearly psychopathologic may not be viewed as evil (e.g., impulsive aggression in the context of a frontal lobe tumor may be viewed as excusable, as "mad" rather than "bad"). Taken together with recent work on the cognitive-affective neuroscience of psychopathy, narcissism, and sadism, these considerations may lead to somewhat different approaches from those taken by Peck and Davidson. In his work on psychopathy, for example, Blair emphasizes a dysfunction in cognitive-affective processing resulting in abnormalities in the weighing up of social emotions (Blair 2004). This account is not inconsistent with a view that in psychopathy there is a disturbance in moral processing that is evil (contra Peck). Furthermore, this...