Moral judgment in adults with autism spectrum disorders

Abstract

The ability of a group of adults with high functioning autism (HFA) or Asperger Syndrome (AS) to distinguish moral, conventional and disgust transgressions was investigated using a set of six transgression scenarios, each of which was followed by questions about permissibility, seriousness, authority contingency and justification. The results showed that although individuals with HFA or AS (HFA/AS) were able to distinguish affect-backed norms from conventional affect-neutral norms along the dimensions of permissibility, seriousness and authority-dependence, they failed to distinguish moral and disgust transgressions along the seriousness dimension and were unable to provide appropriate welfare-based moral justifications. Moreover, they judged conventional and disgust transgressions to be more serious than did the comparison group, and the correlation analysis revealed that the seriousness rating was related to their ToM impairment. We concluded that difficulties providing appropriate moral justifications and evaluating the seriousness of transgressions in individuals with HFA/AS may be explained by an impaired cognitive appraisal system that, while responsive to rule violations, fails to use relevant information about the agent’s intentions and the affective impact of the action outcome in conscious moral reasoning.

Introduction

Autism spectrum disorders (ASDs) are pervasive developmental disorders characterized by abnormal social interaction, verbal and non-verbal communication problems, and restricted interests. Moral reasoning is a key feature of social cognition. Thus, assessing whether individuals with ASDs have intact abilities to express and justify moral judgments is crucial for establishing the nature and the extent of their social impairments. Surprisingly, however, moral cognition has rarely been investigated in individuals with ASDs.

Research on moral cognition in subjects with typical development has focused on two perspectives. The first perspective claims that moral judgment is the product of conscious, effortful and sophisticated reasoning on the basis of explicit abstract principles (Piaget, 1965/1932, Kohlberg, 1981); the second one relies on the assumption that morality takes the form of intuitions, accomplished by rapid, automatic, and unconscious affective responses (Haidt, 2001). In contrast with the ‘conscious reasoning’ perspective (Kohlberg, 1981), Haidt (2001) showed that when people are confronted with moral scenarios they engage in a process called ‘moral dumbfounding’ in which they are unable to articulate sufficient justifications for their confidently expressed moral judgments (Haidt, 2001, Haidt, 2003) and that conscious reasoning only provides post hoc explanations for moral justifications. According to Haidt (2003), amongst moral emotions, prosocial emotions such as empathy, sympathy, concern and compassion, promote morally good behaviour by orienting us to the welfare of society or to the needs of persons other than the agent. In this regard, empathy can be regarded as the capacity to experience other people’s emotions vicariously. Indeed, empathy is a complex process that is made up of two relatively independent components: an affective, phylogenetically early, emotional contagion system and a more advanced, cognitive perspective-taking system (De Waal, 2008). The ability to experience cognitive empathy appears late in ontogenesis: infants show emotional responsiveness to the distress of others without being able to separate their own and the other’s distress (Singer, 2006), and only older children and adolescents may be able to take the other individual’s perspective and display cognitive empathy (Preston & de Waal, 2002). While affective empathy refers to the capacity to experience emotional reactions to the observed experiences of others by automatically activating one’s own representations for the related emotional (or arousal) states, the cognitive notion of empathy emphasizes the ability to engage in the effortful conscious cognitive process of adopting another’s psychological point of view and infer their mental states (Davis, 1994). In other terms, cognitive empathy involves processes, such as perspective taking and ToM (Blair, 2005, Shamay-Tsoory et al., 2009, Singer, 2006). Neuroimaging research further supports the view that these two components are mediated by distinct neural circuits within the frontal cortex: the affective empathy system preponderantly involves the Inferior Frontal Gyrus, whereas the more cognitive system is subserved by the ventromedial prefrontal regions (Shamay-Tsoory et al., 2009).

With respect to empathy dysfunction, the inability to share emotional states with others has been described as one of the most striking clinical features of individuals with ASDs (Kanner, 1943). Although ASDs have often been associated with an impairment in processing and naming facial expressions of emotions (Capps et al., 1992, Grossman et al., 2000, Hobson, 1986 David and Tager-Flusberg, 1997, Yirmiya et al., 1989), Baron-Cohen and collaborators (Baron-Cohen, 1991, Baron-Cohen et al., 1993) have shown that disturbances in understanding others’ affective states arise in people with ASDs when the appreciation of the emotion requires the representation of others’ beliefs, such as surprise or embarrassment (i.e. belief-based emotion), but not for emotions generated by factual events (i.e. reality-based emotions). However, Castelli (2005) showed that children with autism were as able as controls to recognize the six basic emotions from facial expressions (anger, fear, disgust, happiness, sadness, surprise) with different intensity levels. The author suggested that individuals with autism might use compensatory strategies to bypass their deficit in emotion recognition.

According to Blair (1999), while children with autism show psychophysiological responsiveness to others’ distress, they lack the cognitive component of empathy which gives rise to the experienced feelings. In the same direction, those studies that differentiate between different components of empathy have revealed difficulties in cognitive, but not in affective, aspects of empathy in individuals with ASDs (Dziobek et al., 2008, Rogers et al., 2007, Shamay-Tsoory et al., 2002, Yirmiya et al., 1992). Overall, these findings suggest relatively intact emotional empathic reactions to other individuals’ affective states, along with impaired ToM and cognitive empathy.

In individuals with typical development, moral cognition has been primarily studied by assessing the ability to distinguish moral transgressions (e.g. hitting another person) from conventional transgressions (e.g. going to school wearing pyjamas) (for reviews, see Nucci, 2001, Smetana, 1993). Cross cultural studies have shown that 3-years-old children with typical development distinguish between these two types of transgression along a number of dimensions (Nisan, 1987, Smetana and Braeges, 1990, Turiel, 1983). Moral transgressions are considered to be more serious, less permissible and less authority-dependent than conventional transgressions. Furthermore, the justifications of why moral transgressions are wrong tend to make reference to fairness and harm to victims, while in the case of conventional transgressions the explanation statement is usually given in terms of violation of social rules. According to Turiel (1983), the distinction between moral and conventional rules is grounded partly in differences in the affective responses elicited by the protagonist’s actions: transgressions of conventional rules have effects on social order and rarely cause distress to persons, whereas moral transgressions are likely to produce negative emotional effects.

Recently, Nichols, 2002, Nichols, 2004 has elaborated a model according to which moral judgment is based on the interaction of two independent cognitive mechanisms: (a) a Normative Theory, i.e., an internally represented set of norms prohibiting behavior that harms others; (b) an Affective System that confers a special status on these norms, distinguishing them from mere conventional norms. Accordingly, he claimed that the moral/conventional distinction taps a distinction between affect-backed and affect-neutral norms. From this perspective, other affect-provoking transgressions will be distinguished from conventional transgressions. In line with his prediction, Nichols (2002) found that disgust transgressions, which also elicit strong affective reactions, are considered less permissible, more serious and less authority-contingent than conventional transgressions, suggesting that both moral norms and norms prohibiting disgusting behaviours are affect-backed normative rules.

In a previous study, Blair (1996) tested two groups of children with ASDs on the moral/conventional task. The first group passed both the Sally-Ann and the Smarties false belief tests, while the second group failed the two tests. Since both groups were able to draw the distinction between moral and conventional transgressions, Blair concluded that moral judgment is spared in autism and that it is independent of ToM abilities. Leslie and collaborators (2006) replicated Blair’s findings on a sample of children with autism who failed both aforementioned false belief tests. Like Blair, Leslie and colleagues drew the conclusion that basic moral judgment is preserved in autism and that it may function somewhat independently of ToM.

The ability to understand another person’s action from that person’s intentions and desires plays an important role in moral judgments. For example, wrong intentional actions are judged to be worse than similar unintentional ones (Lagnado & Shannon, 2008) and, similarly, attempted but failed harmful acts are judged to be more morally blameworthy than accidentally harmful acts (Young & Saxe, 2009). Developmental studies have shown that children’s motive-based moral reasoning was positively correlated with their false-belief understanding, suggesting a relation between children’s theory of mind and the domain of moral judgment (Baird & Astington, 2004). This close connection between moral judgment and ToM is further supported by recent neuroimaging evidence (Young et al., 2010, Young et al., 2007). Moral judgments require that participants balance evaluations of the actual outcomes against considerations of the agent’s desires, beliefs and intention, which rely upon ToM reasoning.

Difficulties with ToM tasks are largely documented in individuals with ASDs (Baron-Cohen, 1989, Baron-Cohen, 1995, Baron-Cohen et al., 1985, Baron-Cohen et al., 1986, Leslie, 1987). In false belief tasks, the attribution of the agent’s intentions relies on a more competent belief-desire reasoning ability since the default true-belief attribution has to be inhibited to select the appropriate mental content (Friedman and Leslie, 2004, Wimmer and Perner, 1983). However, ToM impairment in ASDs may extend well beyond belief understanding, and include difficulties with the attribution of desires and intentions (Phillips, Baron-Cohen, & Rutter, 1998).

Given the relevance of intention to the determination of moral appreciation, it is somewhat surprising that moral judgment has been found to be substantially intact in people with ASDs (Blair, 1996, Leslie et al., 2006). However, in Blair’s (1996) and Leslie Mallon, and Dicorcia’s (2006) studies, participants were not asked to explain why they judged transgressions to be wrong. This is crucial missing information, since a competent moral judge is not only able to distinguish moral from conventional transgressions, but also to provide appropriate justifications for her judgments. Indeed, according to Grant and collaborators (2005), when asked to justify moral judgments, children with autism gave justifications of poor quality and, more recently, Moran and collaborators (2011) showed that ToM impairments in individuals with high functioning autism affected their moral judgments as they were less willing than adults with typical development to exculpate agents for accidental harm caused on the basis of innocent intentions.

Within the domain of ASDs, high functioning autism (HFA) commonly refers to individuals meeting criteria for autism with normal intellectual functioning and a history of speech and language delay. Asperger Syndrome (AS, DSM-IV, American Psychiatry Association, 2000, ICD-10, 1992) describes individuals with no delayed language function and with normal intellectual abilities. Although individuals with HFA or AS can often solve first-order (e.g. “Sally thinks it’s x, when really it’s y”) and second-order false beliefs tests (e.g. “Sally thinks Mary thinks x, but both Sally and Mary are wrong”) (Bowler, 1992, Dahlgren and Trillingsgaard, 1996, Happé, 1995, Leekam and Prior, 1994), they might fail tasks that require more complex reasoning about others’ mental states, based on detection of sarcasm, irony or bluff (Happé, 1994) or on recognition of Faux Pas (Baron-Cohen et al., 1999, Zalla et al., 2009).

In a previous study, using a more advanced ToM test, the Faux Pas recognition task (Baron-Cohen, O’Riordan, Jones, Stone, & Plaistead, 1999), Zalla and collaborators (2009) have showed that individuals with HFA/AS have difficulty distinguishing intentional from non-intentional behaviors. A faux pas is a particular case of a non-intentional action, since it occurs when a speaker says something that might hurt or be unpleasant to the listener, although the speaker never intended it to do so. Participants were presented with stories describing interpersonal interactions in everyday life situations in which a faux pas occurred, and control stories containing a minor conflict or accident.¹ Here is a typical faux pas story: “Jill had just moved into a new apartment and she had bought new curtains. When Jill had just finished hanging the new curtains, her best friend, Lisa, came over to visit the new apartment. After a tour of the apartment, Lisa said: “Those curtains are horrible. I hope you’re going to get some new ones!”. Obviously, Lisa’s statement reflected her mistaken belief that the curtains had been left by the previous owner, and she did not have the intention to hurt Jill. Hence, a full-fledged understanding of faux pas situations requires the ability to understand the speaker’s state of mind and appreciation of the emotional impact of the statement on the listener. Interestingly, although individuals with HFA/AS generally acknowledged that Lisa said something awkward, they were unable to provide correct justifications of why what Lisa did was awkward, they failed to understand that Lisa had a mistaken belief which was what had caused her faux pas and interpreted Lisa’s statement as caused by her intention to hurt Jill. Typically, individuals with HFA/AS provided explanations in terms of malicious intentions. They judged that the speaker committing the faux pas intended to humiliate and offend the listener and, interestingly, they failed to describe appropriately the emotional impact on the listener. By contrast, for control participants a faux pas is a non-intentional by-product of an intentional act based on some false beliefs. These results revealed that while individuals with HFA/AS are able to detect social rule violations, both their abilities to interpret an action outcome as intentional or accidental and to provide an empathic appreciation of the listener‘s emotional state (knowledge about emotions) were diminished. It is noteworthy, however, that the empathy question in the Faux Pas was not specifically designed to assess either the cognitive or the emotional components of empathy. In a recent study (Zalla & Leboyer, 2011), we showed that, although under certain circumstances, individuals with HFA/AS and people with typical development have similar intuitive judgments of intentionality, over-assignment of praise judgments and the reduced use of folk-psychological concepts in moral judgment in individuals with HFA/AS likely reflect difficulties using intentionality information for moral reasoning.

In the present study, we investigated whether adults with HFA/AS are able to distinguish moral, conventional, and disgust transgressions using the task developed by Nichols (2002). In so doing, we aimed to assess whether impairments in ToM, as measured by the Faux Pas recognition test (Baron-Cohen, Jolliffe, Mortimore, & Robertson, 1997), would affect the ability of individuals with HFA/AS to draw the distinction between these three types of normative transgressions. Participants were given a set of six transgression scenarios, each of which was followed by questions about permissibility, seriousness, authority contingency and justification. If one hypothesizes that ToM is a crucial component of moral cognition, one should expect that individuals with HFA/AS would encounter difficulties in distinguishing types of transgressions and in giving appropriate normative justifications for moral scenarios.