Frontline Learning Research Special Issue Vol 8,
No. 5 (2020) 92 - 104
ISSN 2295-3159
aLudwigs-Maximilians-Universität
München, Germany
Doi: https.www.doi.org/10.14786/flr.v8i5.689
Research on the role of moral emotions in moral judgment, both in hypothetical dilemmata and in real-life moral decision making, has focused on preschool and elementary school age, with few studies spanning a larger age range, into adolescence and adulthood. The present special issue addresses a neglected area, the development of moral emotions and moral motivation in adolescence and adulthood. The focus is on the “Happy Victimizer Phenomenon”, a pattern of emotion attributions to a moral transgressor that has been primarily observed in childhood, but that does not seem to disappear with age. We will begin by briefly reviewing 30 years of developmental research on moral emotion attribution and the “happy victimizer phenomenon”. This review will be followed by a discussion of the present papers.
While research on early moral development has emphasized, over the last 30 years, that traditional descriptions of the young child as “pre-moral”, unaware of the nature of moral rules, and unable to take an agent’s intentions into account when evaluating his or her actions, were fundamentally wrong, there is reason to believe that young children’s understanding of moral emotions differs in important ways from older children’s and adults’, and that this has consequences for their understanding of moral agency and their moral motivation. In the first systematic study of moral emotion understanding in children, Nunner-Winkler and Sodian (1988) found marked developmental change between the ages of 4 and 8 years, in children’s attributions of emotions to moral wrongdoers: When asked, for instance, how a child who had pushed another child from the swing, would “feel” after having committed this transgression, 74% of 4-year-olds, 40% of 6-year-olds, but only 10% of 8-year-olds attributed positive emotions, rather than moral emotions such as guilt, shame or empathy with the victim to the transgressor. This was not due to young children’s lack of knowledge about moral or empathic emotions in general. They attributed empathic emotions to a bystander who witnessed a harmful act, but they appeared to base their emotion attributions to a transgressor exclusively on the information about this agent’s desires. This developmental trend in emotion attributions, with negative or mixed feelings being attributed to the victimizer by the majority of the children from around the age of 7 years, has proved stable in subsequent research (see Arsenio, 2014; Krettenauer, Malti, & Sokol, 2008, for reviews). Even with rigorous probing for opposite valence emotions most 4-year-olds continued to anticipate positive emotions in the victimizer (Arsenio & Kramer, 1992). In contrast, when asked how they would feel themselves if they had committed a transgression, even young children attributed negative emotions more often to themselves than to another person. However, more than 50% of the 5- to 6-year-olds still responded with positive emotions when answering for themselves (Keller, Lourenco, Malti, & Saalbach, 2003). A recent study by Gummerum, Lopez-Perez, Ambrona, Rodriguez-Cano, Dellaria, Smith, & Wilson (2016) found that executive function (inhibition) and counterfactual reasoning were associated with the happy victimizer phenomenon, indicating that a tendency to respond impulsively and a limited ability to consider hypothetical alternatives may contribute to the happy victimizer response pattern. However, this response pattern can not be reduced to executive demands of the tasks.
Theoretical explanations of the happy victimizer phenomenon in children have accounted for the sharp developmental trend by interpreting it in the broader context of children’s developing understanding of the mind. From the point of view of children’s reasoning about desires (as part of their theory of mind), the HV emotion attribution pattern indicates an understanding of the subjectivity of desires: While 3-year-olds often make their emotion attributions dependent on the objective valence of an action outcome, 4-year-olds consistently attribute positive emotions to an agent when his desire was fulfilled, and negative emotions when this was not the case (Yuill, Perner, Pearson, Peerbhoy, & van den Ende, 1996). Early cognitive interpretations of the happy victimizer pattern (Nunner-Winkler & Sodian, 1988 ; Yuill et al., 1996) have argued that young children, when forming a first understanding of the subjectivity of desires, infer emotional reactions exclusively from the correspondence between desires and action outcomes. Only later, they begin to take other relevant determinants of emotions into account, such as the violation of moral norms or the suffering of the victim.
Different explanations have been proposed for why it is difficult for young children to integrate their knowledge about moral norms with their attribution of emotions to a victimizer. One reason may lie in young children’s failure to understand emotional conflict (Arsenio & Kramer, 1992): Victimizers’ emotional reactions are likely to be deeply ambivalent, with happiness over the satisfaction of one’s wicked desires being intermixed with sadness over/ empathic concern for the victim’s suffering and/or regret and shame with regard to one’s own behaviour. Research on children’s concepts of emotions has shown that ambivalent or conflicting emotions are generally conceptualized late in development, only around the age of 8 years (Pons & Harris, 2005), even though even infants appear to experience emotional ambivalence, for instance when reacting to their mother’s return after a brief separation in the Strange Situation. Similarly, self-evaluative moral emotions (shame and guilt) have been demonstrated in 2-year-olds (e.g., Vaish, Carpenter, & Tomasello, 2016), with effects on prosocial behaviour in the case of guilt. There is a long developmental lag between children’s first experience of complex emotions and emotional conflict and their conceptual understanding of such emotions and their antecedents.
Other theorists have interpreted children’s developing understanding of moral emotions within the broader framework of developmental change in their understanding of agency (see Krettenauer et al., 2008). Harris (1989) argued that an understanding of moral emotions requires a shift from seeing people as agents to seeing them as observers of their own agency. The process of internalizing an external audience allows children to view their own and others’ actions from a third-person perspective, and this perspective is necessary to integrate self-evaluative emotions with simple emotional reactions to a desire-action match or mismatch. Another proposal (Krettenauer et al., 2008; Sokol, 2004; Sokol & Chandler, 2003) focuses on the development of an understanding of autonomous agency and the free will. Young children may not understand that an agent who acts based on his desire has a choice to act according to this desire or not. Perner (1991) argued that children may progress from a non-representational to a representational understanding of desires. A representational understanding is required to comprehend, for instance, that agents can change their desires, and thus are not helplessly exposed to the frustration of seeing their desired course of action thwarted. Such a representational understanding of the mind is reached around the age of 4 years. Krettenauer et al. (2008), however, argue that a higher level of understanding of the interpretive mind is necessary to achieve an understanding of autonomous agency that allows for a deliberate decision between fulfilling selfish desires and acting in accordance with moral norms. Such an interpretive understanding of agency involves a rudimentary understanding of interpretive frameworks. Developmental research on understanding social stereotypes (Pillow & Henrichon, 1996) and interpretive processes (Carpendale & Chandler, 1996) has shown that a first understanding of interpretive frameworks emerges around the age of 7 years. An understanding of autonomous agency and the deliberate choice between moral and immoral courses of action is assumed to require a notion of the interpretive mind that enables agents to weigh alternatives based on rational argument. Consistent with this theory, strong correlations were found between children’s performance on the happy victimizer task and independent measures of an interpretive theory of mind (Sokol, 2004; Sokol, Chandler, & Jones, 2004). These associations could not be accounted for by the more basic ability to represent two different aspects of a situation or event at the same time (Sokol, 2004). In contrast, the relation of HV emotion attributions and counterfactual reasoning found by Gummerum et al. (2016) is consistent with the idea that conceiving of alternative interpretations of a given set of phenomena is linked to moral emotion understanding via a conception of autonomous agency.
While there is a clear age trend towards increasingly morally oriented emotion attributions in childhood, several studies indicate that the Happy Victimizer pattern does not disappear in middle childhood (see Keller et al., 2003). Murgatroyd and Robinson (1993) introduced a variation of the original paradigm in which an authority figure witnessed the transgression. Under these conditions, the attribution of fear (of sanctions) was more frequent than the attribution of sadness or moral emotions. When the witness mistakenly displayed a positive reaction (e.g., ‘good of you to pick up litter’, when a student crumpled up a classmate’s homework), up to 40% of adolescents and young adults (college students) attributed happy emotions to a transgressor who gained some personal benefit through his transgression (Murgatroyd & Robinson, 1997). About 30% of young adults failed to attribute any moral emotions, even in an open-ended format, indicating a strong dependency of emotion attributions in adolescence on the witness, rather than on moral norms.
Without the influence of an onlooker, however, only a very small proportion of adolescents appear to show a “happy victimizer pattern”. A study by Krettenauer and Eichler (2006), using the standard assessment procedure, found almost no evidence for a happy victimizer pattern in 13- to 19-year-olds: Less than 10% of the participants anticipated “no bad” feelings following transgressions of varying severity (giving false testimony to get a desired job; absconding from a traffic accident because of being drunk; not returning a found wallet; stealing from a flea market salesman). Moreover, intensity of emotion ratings, using a 6-point scale (extending from “not bad” to “extremely bad”), with age became increasingly consistent with moral judgment and justification, as well as with certainty ratings (‘How sure are you about your moral judgment’?), thus indicating an increasing integration of norm orientation and emotion attribution in adolescence.
A long-term longitudinal study of moral reasoning and moral motivation was carried out by Nunner-Winkler (1999; 2009), as part of the project LOGIC (Weinert & Schneider, 1999; Schneider & Bullock, 2009), a longitudinal study of originally N=200 3-4 year old children, N=172 participants at age 18 and N=153 at age 23 were presented with three moral conflicts (breaking a promise given to the first customer in a market transaction; lying for a career; keeping a lost good) and one moral dilemma (allowing the blaming of an innocent person to favour one’s own friend). They were asked to specify and justify what they themselves (or the same sex protagonist in the career story) would do and how they (or the protagonist) would feel in the role of the agent and the counter role of the victim of another’s transgression. Answers were scored as reflecting moral versus pragmatic concerns using the following criteria: reference to moral principles or harm done vs. personal profit; asymmetries in emotions ascribed to self in the role of agent vs. victim (e.g., happy as transgressor vs. indignant as victim); intensification or mitigation of action decisions or emotion attributions (e.g., “I would never do such a thing” vs. “I’d probably not do that, I assume”). Based on the combination of these criteria strength of moral motivation was rated as high, middle or low according to the predominance of moral vs pragmatic aspects expressed. 25.6% of the participants were classified as low and 40.3% as high at the age of 18 years, and 18.3% as low and 47.1% as high at age the age of 23 years. A slightly higher percentage of low moral motivation (35%) was obtained with the same vignettes and the same coding procedure in a cross-sectional study of N=203 15-16-year old German students (drawn from former East and West Germany and higher and lower educational tracks (Nunner-Winkler, Meyer-Nikele, & Wohlrab, 2006).
The longitudinal data showed moderate stability of moral motivation between the ages of 18 and 23 years (r=.36). There was, however, no predictive relation between childhood and adolescence: On the contrary, moral motivation at the age of 9 years was negatively correlated with moral motivation at the age of 18 years (r=-.21). This may be partly due to the difference in assessment methods. While the assessment of moral motivation in childhood was exclusively based on moral emotion attribution, an aggregate score of moral reasoning and emotion attribution was used in adolescence and adulthood. Alternatively, the low stability of moral emotion attributions in childhood years might reflect differences in speed of children’s socio-cognitive development, whereas in adolescence the relative importance attributed to moral versus non-moral values becomes more influential and value orientations are more liable to change in response to varying circumstances and social context factors.
Identity formation plays a crucial role in adolescence. The impact processes of identity formation exert on adolescents' moral commitment is convincingly illustrated by findings on the connection between moral motivation and sex role identification (Nunner-Winkler 2009, Nunner-Winkler et al., 2006). No difference in strength of moral motivation was found among participants with low gender identification. However, among highly gender-identified participants there were significantly more males with low moral motivation as well as male LOGIC participants who had experienced a decrease in moral motivation between ages 9 and 23. This finding reflects the fact that participants' male stereotypes were predominantly morally aversive (e.g. assertive, unwilling to admit to weakness, shrewd).
If moral emotion attribution is, in fact, a valid indicator of the strength of moral motivation, then there should be a relationship between moral emotion attributions in hypothetical scenarios and real-world prosocial and antisocial behaviours. In the longitudinal study LOGIC, 7-year-olds’ amoral emotion attributions combined with low values in shyness predicted egotistic ruthless behaviour in a lab-assessment of sharing behaviours (Asendorpf & Nunner-Winkler, 1992). Further, moral emotion attributions in adolescence and adulthood (but not in childhood) predicted antisocial behaviours at age 23, independently of conscientiousness and agreeableness. Interestingly, conscientiousness at age 12 contributed to the development of moral emotion attributions at age 18 and these, in turn, predicted change in conscientiousness at the ages of 18 and 23 years (Krettenauer, Asendorpf, & Nunner-Winkler, 2013). Similarly, Krettenauer and Eichler (2006) reported a correlation between adolescents’ strength of self-attributed moral emotions in hypothetical situations and their self-reported real-life delinquent behaviours even when social desirability was controlled for.
Consistent with these findings, a meta-analysis of 42 cross-sectional studies of moral emotion attribution and pro- and/or antisocial behaviours (with 80000 participants aged 4 to 20; Malti and Krettenauer, 2012) yielded evidence for significant associations between emotion attributions and social behaviour in all age groups under study, not restricted to adolescence. There were small size relations of emotion attribution with prosocial behaviour and moderate size relations of emotion attributions with antisocial behaviour, emphasizing the role of moral emotion attributions such as guilt, regret (or sadness) in morally relevant real-world behaviours. Age did not moderate the relation between moral emotion attribution and social behaviour. Thus, despite developmental change in children’s understanding of moral emotions, moral emotion attributions appear to reflect at all ages individual differences in moral motivation. Self-attributed moral emotions were more strongly related to antisocial behaviour than other-attributed emotions. Furthermore, measures of the intensity of moral emotions showed larger effect sizes than mere positive or negative ratings. In sum, theoretical views of the important role of moral emotion attribution in moral development are supported impressively by research on the relation of emotion attribution and social behaviour.
Despite its significance for moral choices of real-world importance, moral emotion attribution in adolescence and adulthood is an underresearched area. While most studies have focused on developmental change in childhood, sometimes including an adolescent or adult comparison group, systematic research into moral emotion attribution in adolescence and adulthood is rare. The present papers are the first to specifically address this issue, with a focus on the “happy victimizer” pattern, its frequency, transsituational stability and theoretical significance beyond childhood.
Paper 1 “Happy Victimizing in Adolescence and Adulthood – Empirical Findings and Further Perspectives” broadly investigated the incidence and transsituational stability /variation of the HV pattern in adolescence and adults in four empirical studies.
In Study 1 a large, representative, standardized sample of 4th, 7th, and 9th grade students (N= 4935) gave written responses to test questions concerning two vignettes on moral transgressions (breaking a contract / keeping money instead of returning it to the owner). The test questions were “What would you (participant) do in this situation? Why would you do it? How would you feel?” There was an increase in Happy Victimizer patterns from 4th to 9th grade, but only a minority of 9th graders (27% / 12%) showed the HV pattern. The responses to the two vignettes were not correlated, i.e., the responses appeared to be situation specific. Rule enforcement was apparently not assessed, i.e., we do not know whether participants who said they would transgress saw the rule as personally binding.
Study 2 introduced a situation of passive moral temptation (getting too much change money) and asked what the protagonist should do (instead of will do). N=331 14-year-old secondary students participated. Only 72% enforced the rule, i.e., they said, it was not okay that the protagonist kept the money. 10.9% showed the HV pattern, judging the protagonist’s transgression as wrong while attributing positive emotions to the protagonist. The predominant justification for emotion attributions in all these patterns was hedonism.
Study 3 addressed moral emotion attributions in adults. N=271 students of economics and business education gave written responses to vignettes about morally relevant decisions in business contexts. Participants were asked to decide about a course of action for themselves and to justify their choice, as well as to judge how they would feel. HV patterns emerged in 31% to 49% of the participants. Depending on the context, up to half of the participants who chose to transgress were anticipating happy emotions. Correlations among the scenarios ranged from .17 to .35, indicating substantial cross-situational variation. There was no assessment of rule validity.
Study 4 included a fuller assessment of rule understanding and adults’ own decisions relative to a perceived norm. N=233 students judged two situations of temptation. 35% approved of the transgression in the “Start-up” / 77% in the “Jana” story. Emotion attributions showed 29% / 10% positive emotion attributions following transgression choices for self. It is unclear whether these reflect HV patterns in the strict sense, since many of the trangressors may not have acknowledged the rule as personally binding. Again, the findings indicate situation specificity of moral emotion attribution.
Commentary
Paper 1 clearly demonstrated the incidence of the HV phenomenon beyond childhood. Moreover, situational factors influencing HV responses, and variations in response styles emerged.
Incidence of HV beyond childhood. The findings are clear: HV can still be found in later development. The present studies yielded between 10% and 50% HV-like responses in all age groups, with a tendency for a higher proportion of HV responses in adults than in adolescents. However, developmental trends can not be inferred from these comparisons across studies, since different situational types of transgressions and different questioning procedures were used in the different studies.
Consistency across situations. The frequency of HV like answers varied considerably between vignettes even when the same age group was tested and the same measurement used. Moreover, there was also substantial intra-individual variation across situations.
This variation is hard to interpret, since the vignettes used varied in several dimensions. More HV responses might be expected if the transgression arose from a passive temptation than from an intentional plan (e. g. study 3: keeping change vs. TV sale), if the damaged party is a large anonymous company rather than a concrete individual (e. g. study 3 travel cost vs keeping change), if the situation makes it easy to justify the transgression or to put the blame on the victim (e. g. study 1: the first customer had beaten down the price, should have brought the money along in the first place cf. Paper 2 disengagement strategies), and if the damage caused is low. Moreover, different assessment procedures were used. Study 1 asked for a justified action decision and an emotion ascriptions to self (What would you do? Why? How would you feel? Why?) HV was coded when participants expected to feel good for hedonistic reasons about the profit gained by an immoral action decision. Study 2 and study 4 first requested 'deontic judgments' (what should protagonist do? Why? How does s/he feel? Why), and 'self judgement' (what would you do? Why? How would you feel? Why?). Then they presented a hypothetical wrongdoer and asked for a moral judgment and an emotion ascription ('Is it ok or not ok that protagonist transgressed? Why? How does protagonist feel? Why?).
As documented by Keller et al. (2003), HV responses are ascribed more often to hypothetical wrongdoers than to self. In contrast, the present study 4 found “significantly more morally appropriate responses to the the perpetrator in the classical HV situation than in the self judgement situation” (p-14). This finding may be due to the fact that many participants judged the transgression to be ok (35% in the start-up, 73% in the “keeping change” situation). As the present authors acknowledge, a positive emotion attribution to a transgressor does not, strictly speaking, conform to the HV pattern, if the rule that was transgressed is not seen as personally binding. Given that there is no longer a collective consensus on a hierarchy of values in modern pluralistic societies, people may differ in how much importance they attribute to money, professional standing, moral integrity. Thus, the degree of temptation experienced will differ not only across stories but also across individuals. Some people attribute very high / very low importance to moral integrity. These persons can be expected to consistently attribute negative / positive emotions to wrongdoers. Most people, however, will evaluate the costs or gains of transgressing according to context features and their individual value preferences. Thus, moral emotion attribution in adulthood cannot be interpreted independently of adult participants’ evaluation of moral rules and value systems.
Response patterns. The present authors found several new response types. Among participants judging a norm as valid there are happy and unhappy victimizers (transgress and feel good / bad), as well as moralists (conform and feel good / bad). Among participants who did not accept the norm as binding they found – depending on the test question used - “deontic happy transgressors”, “happy transgressors self”, and “happy transgressors misdeed”. According to the authors these new patterns reflect the openness of the situation in which an intention has yet to be formed, whereas in the Classical HV paradigm a moral judgment is requested after the transgression has already been committed (p.10). Note however, that the original HV paradigm was just as open with respect to moral judgment which was requested in the situation of temptation. Rather, the new response patterns may reflect the fact that the norms presented in the present research were by no means regarded as personally binding by all participants, whereas this was the case for the norms used in the classical HV research with children.
The new patterns documented raise the question as to their meaning. It is not clear, for instance, whether unhappy victimizing (UV) indeed stands for true regret or merely serves impression management in the testing situation. Further, one may wonder whether Happy Moralizing versus Unhappy Moralizing may indicate high versus low moral commitment. Research on moral exemplars (persons who all through their lives remained true to their moral convictions despite grave costs incurred) has not reported any special positive feelings about having acted morally right. For them doing what is right was a matter of course. They did, however, mention feelings of regret or guilt about having inadvertently caused harm to their family members (e. g., children suffered when for moral reasons they gave up well paying jobs) (Colby & Damon,1992). Such considerations may also play a role in adults' attribution of mixed emotions (paper 3).
It should also be noted that victimizing responses may be affected by the authors’ “deontic judgments”: ‘What should the protagonist do’? These request action recommendations. As has been argued with respect to Kohlberg’s description of children’s pre-conventional stage, these questions confound the cognitive and the motivational dimension (Nunner-Winkler, 1999). Thus, participants might interpret ‘should’ in the test question in prudential rather than in moral terms. In other words, they may recommend that the protagonist should do not what is right but what benefits him most.
In sum, Paper 1 extensively documents the existence of HV responses beyond childhood, consistent with previous research. Thus, HV is not a transitional phenomenon of childhood cognitive development and needs to be interpreted in a broader context of moral judgment and moral motivation. Further, the studies document high response variability across individuals and situations - an innovative finding that can not yet be conclusively interpreted. Finally, by describing new response patterns, the studies open a whole set of promising research questions.
Paper 2 “An action-theoretical approach to the Happy Victimizer pattern – exploring the role of moral disengagement strategies on the way to action” advances an action theoretical approach to moral decision making which focuses on the phase of forming an intention: Moral distancing strategies (MDS) are conceptualized as cognitive control- or volitional strategies which support an individual’s dealing with an inner conflict or ambivalence. The perceived moral intensity of a situation may impact the process of forming an intention and acting. Situations of high moral intensity may evoke the conscious, reflective application of cognitive control strategies, whereas more automatic processes of moral motivation may dominate in situations of low moral intensity.
The research questions concerned the extent to which adults apply victimizing strategies in situations of low moral intensity, the situation specificity of victimizing decisions, the application of MDS in situations of low moral intensity, and the situation specificity of the application of MDS. N=587 university students (economics and teacher students) were tested in hypothetical low intensity moral transgression situations. Participants were asked what they would do and how they would feel. Justifications for the decisions were coded for MDS. An HV pattern was observed in 11% to 19% of the participants. Among the other response patterns, the Happy Moralizer was the most frequent pattern. There was moderate transsituational consistency. MDS were analyzed for different types of strategies. Some differences emerged descriptively between the MDS associated with HV and UV response patterns.
Commentary
Paper 2 replicates core results presented in paper 1: HV responding occurs beyond childhood and responses vary across situations and – to a moderate degree – between individuals. Besides, paper 2 adds a very interesting expansion in the HV paradigm – the study of MDS in situations of moral temptation. The present findings indicate that MDS vary across situations but that there are also some person-specific preferences for particular types of MDS. The data prompt some speculations which – in addition to the rich research program suggested by the authors – might be followed up in future studies. In criminal law, two types of defenses are distinguished– justification (the act is justifiable or at least not that reprehensible, e.g., served higher moral ends, was not so bad) and excuse (the act is bad but the agent is pardonable, e.g., acted under constraint). The 'Travel costs' story draws more and allows for more justifications (MDS types 2 and 3 presenting seemingly fair solutions, e.g., paying the friend, belittling the harm affecting a large company). The 'Change' story produces more excuses – justifications are hard to come up with (MDS types 6 and 7). Despite the low numbers involved the finding of differences in MDS used by HVs and UVs is suggestive. HVs more often use denial of the wrongness of the act or of own responsibility (travel cost – MD 2: trivializing the behaviour, change MD type 7 – putting the blame on the victim), UVs, in contrast, (at least implicitly) acknowledge the wrongness of the act but appeal to stressful circumstances (travel cost/ change: MD type 6 pointing to own needs). If this finding were robust, UV might indicate higher moral commitment: Participants admit having done wrong and hope for lenience.
Paper 3 “Happy Victimizing in Emerging Adulthood: Reconstruction of a Developmental Phenomenon?” deepens our understanding of moral emotion attribution in adulthood. N=285 pre-service teachers received a paper & pencil questionnaire consisting of 3 vignettes (Change Money; The Motorbike; Lying to a Customer). Moral rule understanding was assessed by a “deontic” judgment plus justification, extended emotion ascription (choice between positive, negative and mixed emotions, specification of the chosen emotion(s) and justification). Finally, a self-judgment was followed by an emotion attribution. The vignettes were followed by an assessment of the “moral self”. The results showed a “pure” HV pattern only in 2.2% of the participants, however, a mixed emotion attribution in roughly 40%/ 49%/27%, depending on the story. Most justifications for the mixed emotions HV pattern were hedonistic. Several types of hedonistic justifications emerged, the protagonist being “happy” for enjoying a profit, e.g., having money, or the protagonist being “justifiably” happy because the victim was “stupid”. Some of the mixed emotion attributions contained a hedonistic justification for the positive, and a moral (e.g., bad conscience) justification for the negative emotion (unfortunately, no detailed analysis of these mixed patterns is provided). The authors provide a further analysis of the positive and negative emotions by levels of complexity, showing that in the case of negative emotions the large majority was unspecific (“bad, uncomfortable”), but roughly 30% of the participants explicitly mentioned a bad conscience. The analysis of the moral self scales revealed that participants categorized as HV ascribed more importance to honesty and truthfulness to themselves than participants categorized as no-HV.
Commentary
The study showed that “pure” HV responses occur only very rarely, when adults are offered the option of emotional ambivalence: Most adults who see a moral rule as binding and expect a protagonist to transgress expect him or her to experience mixed emotions. Ambivalent emotions can vary in level of complexity and in terms of the recognition of moral concerns. Justifications need to be analyzed more deeply. It is possible that participants, who expect emotional ambivalence, may be more strongly morally motivated/ and or more advanced in their moral understanding than participants who expect a single emotional reaction. It is difficult, however, to use the mixed emotion pattern as a straight measure of moral motivation, since the emotion attribution question was originally designed to yield a spontaneous decision for one or the other emotional valence, which is supposed to be an indicator of the strength of moral motivation. Another problem of interpretation arises from the substantial proportion of participants who said the rule should be transgressed. These cannot be included in an analysis of HV patterns. The finding points to a problem associated with the use of situations of low moral intensity. If the rule is not seen as binding, then, strictly speaking, the instrument is not suitable for measuring moral motivation.
An interesting finding is the increase in mixed HV responses if the 'deontic' judgment is added to the classical HV judgment (Tables 1 and 3). The 'should' question and the 'is wrong' question obviously tap different meanings: The first one expresses an action recommendation which may comprise not only moral considerations but also prudential concerns (evaluation of potential costs, risks and gains). The second question unequivocally requests a moral judgment – it is the one that more adequately could be labeled 'deontic'.
Another interesting finding is the higher importance that HVs, compared to non HVs, ascribed to honesty and truthfulness. They in fact do openly admit to their lack of moral concerns. After all, even young children know that a remorseful wrongdoer is better than a joyful one. Conversely, HVs’ straightforwardness suggests it might be worthwhile to more carefully explore the potential influence social desirability might have on moral emotion attributions.
Paper 4 “How to explain the Happy Victimizer in Adulthood” addresses the HV phenomenon from the point of view of rational choice theory.
N=481 university students (economists vs. teacher students) were tested with versions of the Prisoner’s Dilemma (Wall street game; Community game). Participants were given a choice between strategies (cooperate vs. defect), asked to explain their choice, rate how they felt about the decision, and explain their feelings. The cooperation strategy was less frequent in economists than in teacher students. There were between 55% (economists) and 31% (teacher students) happy victimizers, and conversely between 36% and 67% happy moralists.
A further analysis distinguished between strategic moralists and HV in the strict sense. The strategic moralists reason that the situational constraints of the dilemma made it necessary for them to pursue their own interests. If strategic moralists are separated, there is a proportion of 47% “pure” HV among economics students and of 27% among teacher students. Thus, the study demonstrates a high proportion of HV or HV-like patterns in a Prisoner’s Dilemma situation, with marked intra- and interindividual differences, reflecting situational determinants, as well as (possibly) disciplinary variation in strategic thinking and moral judgment.
Commentary
Minnameier follows the core claim of rational choice theory: Agents will always try to maximize utility by pursuing their interests – be these selfish or other-regarding. According to the author, economists' higher rate of defection results from their realizing that – without social sanctions – the conflict of interests cannot be solved. In such situations it is morally justified that agents pursue their own interests as long as they accept others doing so, too. Thus, in line with this Neo-Kohlbergian “stage 2A” moral principle, defecting is morally acceptable behaviour.
However, modern functionally differentiated societies comprise morally neutral domains which allow pursuing one's interests as long as the more global moral/legal framework is respected. PD behaviour belongs to this sphere, is non-moral behaviour. Therefore, Minnameier is right in claiming that defecting does not necessarily indicate a lack of moral motivation. Nevertheless, cooperation might indicate special moral commitment – the willingness to follow moral convictions even when this is not mandatory and no sanctions impend. Thus, the ‘naive’ understanding of cooperation that many teachers displayed might not - as Minnameier asserts - reflect a cognitive deficiency but rather “habitual action schemes” (p.4) that might be characteristic of persons with high moral motivation. These would be based on fundamental premises – basic trust, appreciation of equality, and fairness. Such moral attitudes are not contrary to rationality as Minnameier argues. This claim is supported by game theory (cooperation in the tit-for-tat strategy yields the best results), by analyses of political systems (e. g. an analysis of 20 regional Italian governments found efficiency strongly correlated with citizens’ solidarity, tolerance, trust, Putnam et al., 1993) and by studies on economic growth (e. g. in 160 countries a strong correlation was found between prosperity and residents’ happiness with absence of corruption: Delhey, 2002).In fact, Minnameier’s own data confirm the claim that cooperation based on trust is advantageous. On average, economists will end up with 20 Euros, given that most other economists will defect whereas teacher students will end up with 50 Euros given that most others will cooperate. Yet, irrespective of the long-term advantage of cooperation, morally motivated people might appreciate equal-distribution principles as an intrinsic value. After all, the idea of equality is constitutive for a secularized moral understanding. Also, those who care about morality will disapprove of unfairness and in PDs the greater profit results from exploiting the trust advanced by the partner.
The present papers contribute importantly to a fuller understanding of the HV phenomenon. Despite large variation in the incidence of HV-type responses in adolescents and adults, depending on the task format, all studies documented the classic HV pattern (the participant regards a rule as binding and nevertheless attributes exclusively positive emotions to the transgressor) in a small minority of adults (up to 10%). There was no clear age-trend from adolescence to adulthood. Given the association of HV-responding with antisocial action (Malti & Krettenauer, 2012), this minority may be at risk for deviant behaviours. It is possible that the HV pattern in the strict sense is also a sign of cognitive immaturity in a small minority of adolescents and adults who fail to develop an interpretive Theory of Mind, including an understanding of humans as autonomous agents, whose actions are subject to interpretation and evaluation.
Depending on task format and context, up to 50% of HV-like response patterns were observed in adults. Situations of low-intensity moral demands that allow for moral distancing strategies, especially situations of competition that are ambiguous with respect to norms of competitive “fair play” versus norms of cooperation, elicit a high proportion of HV-like responses. Generally, emotion attribution in situations of norm violation cannot be interpreted without evidence on participants’ construal of the norms in question. If participants do not regard the norm as binding, and justify this judgment with pragmatic reasons, then an emotion attribution that focuses solely on the agent’s desires is to be expected. Since there is considerable variation in norm understanding among adolescents and adults, it appears that a compound score of moral motivation as used by Nunner-Winkler (2009), including both norm evaluation and emotion attribution, may provide a better indicator of moral motivation than emotion attribution per se.
With respect to emotion attribution, paper 3 clearly indicates that adults will generally be aware of emotional ambivalence and choose mixed emotions if this option is provided. It is hard to interpret the choice of emotional ambivalence in terms of moral motivation without an in-depth assessment of the emotional qualities anticipated by the participants, and the justification for these attributions. In contrast, the original rationale for using emotion attribution as an indicator of moral motivation was that the spontaneous choice of a specific emotional valence may reflect dispositional tendencies that may also contribute to real-world moral decision making. Thus, offering the choice of emotional ambivalence may make emotion attribution invalid as an indicator of moral motivation.
The present studies revealed a lack of consistency in moral emotion attributions in adulthood that had not been noted before. Responses varied greatly across situations. It appears that the determinants of such diversities are not yet fully understood and deserve further explorations. Among the dimensions that should be further explored are the impact of positive vs. negative duties (e.g. breaking a promise vs not returning a lost good), the severity of the harm done (e.g. not returning change vs giving false testimony at court or absconding from an accident); the status of the victim (e.g. personal friend, unknown individual, anonymous organization), the social context of the transgression (action takes place secretly or in public, a (dis)approving authority is or is not present).
Furthermore, the present authors used a variety of modifications of the interview procedures. The results document the unexpectedly large influence details of the wording of questions and the order of their presentation may have on participants' reactions. Further systematic research on methodological variations seems warranted. In the classical HV paradigm, for instance, justified moral judgments were requested in the situation of temptation and emotion attributions after the protagonist had transgressed. In contrast, some recent studies present the transgression before requesting its evaluation or do not request a moral judgment at all, assuming broad agreement on the immorality of the act presented. Heinrichs et al. (paper 1, study 2) start off with ‘deontic’ and ‘self judgements’. Krettenauer and Eichler (2006) added an epistemic judgment. Response formats may have effects as well. Some studies allowed for open answers, others presented 2- to 6-point Likert scales on which specific emotions (happy, ok, scared, sad, good, bad, not bad, a little bad, moderately bad, bad, very bad, extremely bad) had to be rated.
Maybe most importantly, the present authors have succeeded in linking up the HV phenomenon with other research areas. Thus, they have taken up the recent debate on moral disengagement and analysed participants' strategies of vindicating deviant action decisions. And they have connected the study of moral emotions with behavioural economics thus opening up a way of experimentally testing whether emotion attributions to wrongdoers may predict real life behaviour, i.e. to what extent a motivational interpretation of HV patterns is warranted.
Finally, the present findings raise a number of issues relevant to moral education. If emotion attribution becomes increasingly situation-specific and contextualized in adolescence and adulthood, then what needs to be supported in education is the establishment of contexts that are conducive to the growth of moral motivation. It would, however, be a fatal mistake to focus interventions on moral emotions. Such educational efforts might merely promote conformist public reactions. Attempts at improving the moral climate in classrooms appear to be more promising.
A good way – also suggested by the present authors - is the establishment of just communities. Studies have shown: In such democratic contexts, participants feel more bound to norms given they freely agreed to them, they feel more responsible for class mates and are more ready to support and help them (Oser & Althof, 1992; Higgins, Power, & Kohlberg, 1984).
Beate Sodian was supported by DFG grant SO 213/27-3.
1 Nunner-Winkler & Sodian (1988) have recently been labelled as proponents of a purely motivational interpretation of the HV phenomenon (Gummerum et al, 2016; Krettenauer et al., 2008). This is strange since they initially discussed their findings in terms of emotion- and desire-understanding, i.e., with respect to young children’s emotion concepts and Theory of Mind. There has never been a debate on whether the HV-phenomenon had either cognitive or motivational significance. Moral emotion attributions in young children may reflect an immature understanding of the determinants of emotions, and this understanding may undergo conceptual change. At the same time, the level of understanding of moral emotions that a child has may gain motivational significance in real-life moral conflict situations.
Arsenio, W. (2014). Moral emotion attributions and aggression. In
M. Killen & J. Smetana (Eds.). Handbook of moral
development (2 nd edition) (pp.235-255). New
York: Psychology Press.
Arsenio, W. F., & Kramer, R. (1992). Victimizers and their
victims: Children's conceptions of the mixed emotional
consequences of moral transgressions. Child Development,
63(4), 915-927.
Asendorpf, J.B. & Nunner-Winkler, G. (1992). Children's moral
motive strength and temperamental inhibition reduce their immoral
behaviors in real moral conflicts. Child Development, 61,
1223-1235.
Carpendale, J. I., & Chandler, M. J. (1996). On the
distinction between false belief understanding and subscribing to
an interpretive theory of mind. Child Development, 67(4),
1686-1706.
Colby, A. & Damon, W. (1992) Some do care. Contemporary
lives of moral commitment. New York: Free Press.
Delhey, J. (2002) Korruption in Bewerberländern zur Europäischen
Union. Soziale Welt, 53, 345-366.
Gummerum, M., López-Pérez, B., Ambrona, T., Rodríguez-Cano, S.,
Dellaria, G., Smith, G., & Wilson, E. (2016). Children's Moral
Emotion Attribution in the Happy Victimizer Task: The Role of
Response Format. Journal of Genetic Psychology, 177(1),
1-16.
Harris, P. L. (1989). Children and emotion: The development
of psychological understanding . Basil Blackwell.
Higgins, A., Power, C., Kohlberg, L. (1989).The relationship of
moral atmosphere to judgements of responsibility. In: W.M.
Kurtines, & L. Gewirtz (Eds.) Morality, moral behavior,
and moral development. Basic issues in theory and research
. (pp 74- 106) New York: Wiley.
Keller, M., Lourenco, O., Malti, T., Saalbach, H. (2003).The
multifaceted phenomenon of 'happy victimizers': A cross-cultural
comparison of moral emotions. British Journal of
Developmental Psychology 21, 1‒18.
Krettenauer, T., Asendorpf, J. B., & Nunner-Winkler, G.
(2013). Moral emotion attributions and personality traits as
long-term predictors of antisocial conduct in early adulthood:
Findings from a 20-year longitudinal study. International
Journal of Behavioral Development, 37(3),
192-201.
Krettenauer, T., & Eichler, D. (2006). Adolescents'
self‐attributed moral emotions following a moral transgression:
Relations with delinquency, confidence in moral judgment and age.
British Journal of Developmental Psychology, 24(3),
489-506.
Krettenauer, T., Malti, T., Sokol, B. W. (2008). The Development
of Moral Emotion Expectancies and the Happy Victimizer Phenomenon:
A Critical Review and Application. European Journal of
Developmental Science, 2/3, 221‒235.
Malti, T., & Krettenauer, T. (2013). The relation of moral
emotion attributions to prosocial and antisocial behavior: A
meta‐analysis. Child Development, 84(2),
397-412.
Murgatroyd, S. J., & Robinson, E. J. (1993). Children's
judgements of emotion following moral transgression. International
Journal of Behavioral Development, 16(1), 93-111.
Murgatroyd, S. J., & Robinson, E. J. (1997). Children's and
adults' attributions of emotion to a wrongdoer: The influence of
the onlooker's reaction. Cognition & Emotion, 11(1),
83-101.
Nunner-Winkler, G. (1999) Development of Moral Understanding and
Moral Motivation. In: F.E. Weinert, & W. Schneider (Eds.) Individual
Development from 3 to 12. (pp- 253-290) Cambridge UK:
Cambridge University Press.
Nunner-Winkler, G. (2009). Moral Motivation from Childhood to
Early Adulthood. In: W. Schneider & M.Bullock (Eds.) Human
Development from Early Childhood to Early Adulthood. (pp.
91-118) New York: Taylor &Francis.
Nunner-Winkler, G., Meyer-Nikele, M. Wohlrab, D. (2006).
Integration durch Moral. Moralische Motivation und Ziviltugenden
Jugendlicher . Kap. 4 Erhebung und Validierung
moralischer Motivation (pp.65-81). Wiesbaden: Verlag für
Sozialwissenschaften.
Nunner-Winkler, G., Meyer-Nikele, M., & Wohlrab, D. (2007)
Gender and Moral Motivation. Merrill Palmer Quarterly, 53,
26-52.
Nunner-Winkler, G., & Sodian, B. (1988). Children's
understanding of moral emotions. Child Development, 59,
1323-1338.
Oser, F. & Althof, W. (1992) Moralische Selbstbestimmung.
Stuttgart: Klett Cotta.
Perner, J. (1991). Understanding the representational mind.
Cambridge, MA. MIT Press.
Pillow, B. H., & Henrichon, A. J. (1996). There's more to the
picture than meets the eye: Young children's difficulty
understanding biased interpretation. Child Development,
67(3), 803-819.
Pons, F., & Harris, P.L. (2005). Longitudinal change and
longitudinal stability of individual differences in children's
emotion understanding. Cognition & Emotion, 19(8),
1158-1174.
Putnam, R.D., Leonardi. R. & Nanett, R.Y. (1993) Making
Democracy Work.Civic Traditions in Modern Italy. Princeton,
N.J.: Princeton UP.
Schneider, W., & Bullock, M. (Eds.). (2010). Human
development from early childhood to early adulthood: Findings
from a 20 year longitudinal study . Psychology Press.
Sokol, B. W. (2004 ). Children’s conceptions of agency and
morality: Making sense of the happy victimizer phenomenon
. (Unpublished doctoral dissertation). University of British
Columbia, Vancouver, Canada.
Sokol, B.W., Chandler, M.J. (2003). Taking agency seriously in the
theories-of-mind enterprise: Exploring children's understanding of
interpretation and intention.British Journal of Educational
Psychology Monograph Series II, Number 2 ‒
Development and Motivation, 125‒136.
Sokol, B. W., Chandler, M. J., & Jones, C. (2004). From
mechanical to autonomous agency: The relationship between
children's moral judgments and their developing theories of mind.
New directions for child and adolescent development, 2004
(103), 19-36.
Vaish, A., Carpenter, M., & Tomasello, M. (2009). Sympathy
through affective perspective taking and its relation to prosocial
behavior in toddlers. Developmental Psychology, 45(2),
534.
Weinert, F. E., & Schneider, W. (Eds.). (1999).
Individual development from 3 to 12: Findings from the Munich
Longitudinal Study . Cambridge University Press.
Yuill, N., Perner, J., Pearson, A., Peerbhoy, D., & van den
Ende, J. (1996). Children's changing understanding of wicked
desires: From objective to subjective and moral. British
Journal of Developmental Psychology, 14(4),
457-475.