Moral Emotions and Moral Behavior

Abstract

Moral emotions represent a key element of our human moral apparatus, influencing the link between moral standards and moral behavior. This chapter reviews current theory and research on moral emotions. We first focus on a triad of negatively valenced “self-conscious” emotions—shame, guilt, and embarrassment. As in previous decades, much research remains focused on shame and guilt. We review current thinking on the distinction between shame and guilt, and the relative advantages and disadvantages of these two moral emotions. Several new areas of research are highlighted: research on the domain-specific phenomenon of body shame, styles of coping with shame, psychobiological aspects of shame, the link between childhood abuse and later proneness to shame, and the phenomena of vicarious or “collective” experiences of shame and guilt. In recent years, the concept of moral emotions has been expanded to include several positive emotions—elevation, gratitude, and the sometimes morally relevant experience of pride. Finally, we discuss briefly a morally relevant emotional process—other-oriented empathy.

OVERVIEW

What confluence of factors foster a moral life lived to the benefit of self and others? This review summarizes current theory and research on moral emotions, offering a framework for thinking about the ways in which morally relevant emotions may moderate the link between moral standards and moral decisions, and ultimately moral behavior.

Living a moral, constructive life is defined by a weighted sum of countless individual, morally relevant behaviors enacted day in and day out (plus an occasional particularly self-defining moment). As imperfect human beings, however, our behavior does not always bear a one-to-one correspondence to our moral standards.

Many potential explanations exist for the discrepancy between behavioral decisions (intentions) and actual behavior in both moral and nonmoral domains. Historically, much social psychological theory and research was devoted to understanding the imperfect link between intentions (e.g., moral decisions) and behavior. Field theory, the very foundation of social psychology, highlights the variability of individual behavior as a function of situational context (Lewin 1943); interpersonal negotiation can undermine the link between intention and behavior (DeVisser & Smith 2004); and diffusion of responsibility can undermine one’s ability to act on deeply held beliefs (see, e.g., Latane & Darley 1968). Ajzen’s (1991) theory of planned behavior offers a well-integrated model of the ways in which attitudes, norms, and perceived control feed into behavioral intentions and subsequent behavior.

As with the link between intentions and behaviors in general, the link between moral intentions and moral behaviors is likewise an important issue. However, owing to space limitations, this chapter focuses on the processes further upstream from intentions: the less widely studied factors that strengthen (or disrupt) linkages between moral standards and moral intentions (which we refer to throughout this article as moral decisions), and thus moral behaviors. In our view, the link between moral standards and moral decisions and/or moral behavior is influenced in important ways by moral emotions.

Moral standards represent an individual’s knowledge and internalization of moral norms and conventions. People’s moral standards are dictated in part by universal moral laws, and in part by culturally specific proscriptions. The current review emphasizes cognitive and emotional processes relevant to the more cross-culturally invariant moral standards. Of primary interest are prohibitions against behaviors likely to have negative consequences for the well-being of others and for which there is broad social consensus that such behaviors are “wrong” (e.g., interpersonal violence, criminal behavior, lying, cheating, stealing).

Naturally, people do, on occasion, lie, cheat, and steal, even though they know such behavior is deemed wrong by moral and societal norms. Individual differences in people’s anticipation of and experience of moral emotions likely play key roles in determining actual moral choices and behavior in real-life contexts.

Moral emotions represent an important but often overlooked element of our human moral apparatus. Moral emotions may be critically important in understanding people’s behavioral adherence (or lack of adherence) to their moral standards. Haidt (2003) defines moral emotions as those “that are linked to the interests or welfare either of society as a whole or at least of persons other than the judge or agent” (p. 276). Moral emotions provide the motivational force—the power and energy— to do good and to avoid doing bad (Kroll & Egan 2004).

In this article, we focus on a triad of morally relevant, negatively valenced “self-conscious” emotions—shame, guilt, and embarrassment. We also consider several positively valenced moral emotions—elevation, gratitude, and the sometimes morally relevant experience of pride. In addition, we discuss briefly a morally relevant emotional process— empathy.

SELF-CONSCIOUS EMOTIONS: ANTICIPATORY AND CONSEQUENTIAL REACTIONS TO THE SELF

Shame, guilt, embarrassment, and pride are members of a family of “self-conscious emotions” that are evoked by self-reflection and self-evaluation. This self-evaluation may be implicit or explicit, consciously experienced or transpiring beneath the radar of our awareness. But importantly, the self is the object of these self-conscious emotions.

As the self reflects upon the self, moral self-conscious emotions provide immediate punishment (or reinforcement) of behavior. In effect, shame, guilt, embarrassment, and pride function as an emotional moral barometer, providing immediate and salient feedback on our social and moral acceptability. When we sin, transgress, or err, aversive feelings of shame, guilt, or embarrassment are likely to ensue. When we “do the right thing,” positive feelings of pride and self-approval are likely to result.

Moreover, actual behavior is not necessary for the press of moral emotions to have effect. People can anticipate their likely emotional reactions (e.g., guilt versus pride/self-approval) as they consider behavioral alternatives. Thus, the self-conscious moral emotions can exert a strong influence on moral choice and behavior by providing critical feedback regarding both anticipated behavior (feedback in the form of anticipatory shame, guilt, or pride) and actual behavior (feedback in the form of consequential shame, guilt, or pride). In our view, people’s anticipatory emotional reactions are typically inferred based on history—that is, based on their past consequential emotions in reaction to similar actual behaviors and events.

Thus far, we have been discussing situation-specific experiences of consequential and anticipatory feelings of shame, guilt, embarrassment, and pride. In the realm of moral emotions, researchers are also interested in dispositional tendencies to experience these self-conscious emotions (e.g., shame-proneness, guilt-proneness). An emotion disposition is defined as the propensity to experience that emotion across a range of situations (Tangney 1990). From this perspective, shame-prone individuals would be more susceptible to both anticipatory and consequential experiences of shame, relative to their less shame-prone peers. That is, a shame-prone person would be inclined to anticipate shame in response to a range of potential behaviors and outcomes. In turn, such an individual also would be inclined to experience shame as a consequence of actual failures and transgressions.

Shame and Guilt

The vast majority of research on moral emotions has focused on two negatively valanced, self-conscious emotions—shame and guilt.Many individuals, including clinicians, researchers, and lay people, use the terms “shame” and “guilt” synonymously. Nonetheless, a number of attempts have been made to differentiate between shame and guilt over the years.

What’s the difference between shame and guilt?

Attempts to differentiate between shame and guilt fall into three categories: (a) a distinction based on types of eliciting events, (b) a distinction based on the public versus private nature of the transgression, and (c) a distinction based on the degree to which the person construes the emotion-eliciting event as a failure of self or behavior.

Research indicates that type of event has surprisingly little to do with the distinction between shame and guilt. Analyses of personal shame and guilt experiences provided by children and adults revealed few, if any, “classic” shame-inducing or guilt-inducing situations (Keltner&Buswell 1996, Tangney 1992, Tangney et al. 1994, Tracy & Robins 2006). Most types of events (e.g., lying, cheating, stealing, failing to help another, disobeying parents) are cited by some people in connection with feelings of shame and by other people in connection with guilt. Some researchers claim that shame is evoked by a broader range of situations including both moral and nonmoral failures and transgressions, whereas guilt is more specifically linked to transgressions in the moral realm (Ferguson et al. 1991, Sabini & Silver 1997, Smith et al. 2002). In our view (Tangney et al. 2006b), like its sibling guilt, shame qualifies as a predominantly moral emotion, once one moves beyond narrowly conceptualizing the domain of morality in terms of the ethic of autonomy (Shweder et al. 1997). Of the “Big Three” ethics of morality—autonomy, community, and divinity (Shweder et al. 1997)—shame may be more closely tied to violations of the ethics of community (e.g., violations of the social order) and divinity (e.g., actions that remind us of our animal nature), but violations of particular ethics do not bear a one-to-one correspondence to particular situations or events. As demonstrated by Shweder et al. (1997), most failures and transgressions are experienced as relevant to a mix of moral ethics. In short, from this broader cultural perspective, shame and guilt are emotions each primarily evoked by moral lapses.

Another frequently cited distinction between shame and guilt focuses on the public versus private nature of transgressions (e.g., Benedict 1946). From this perspective, shame is viewed as the more “public” emotion arising from public exposure and disapproval of some shortcoming or transgression. Guilt, on the other hand, is conceived as a more “private” experience arising from self-generated pangs of conscience. As it turns out, empirical research has failed to support this public/private distinction in terms of the actual structure of the emotion-eliciting situation (Tangney et al. 1994, 1996a). For example, a systematic analysis of the social context of personal shame- and guilt-eliciting events described by several hundred children and adults (Tangney et al. 1994) indicated that shame and guilt are equally likely to be experienced in the presence of others. Solitary shame experiences were about as common as solitary guilt experiences. Even more to the point, the frequency with which others were aware of the respondents’ behavior did not vary as a function of shame and guilt, in direct contradiction to the public/private distinction. Similarly, in a study of personal emotion narratives, Tracy & Robins (2006) found that, relative to guilt, shame was elicited somewhat more frequently by achievement events and personal events, which are each more private than relational and familial events.

Where does the notion that shame is a more public emotion come from? Although shame- and guilt-inducing situations are equally public (in terms of the likelihood that others are present and aware of the failure or transgression) and equally likely to involve interpersonal concerns, there appear to be systematic differences in the nature of those interpersonal concerns. Tangney et al. (1994) found that when describing shame-inducing situations, respondents expressed more concern with others’ evaluations of the self. In contrast, when describing guilt experiences, respondents were more concerned with their effect on others. This difference in “egocentric” versus “other-oriented” concerns isn’t surprising given that shame involves a focus on the self, whereas guilt relates to a specific behavior. A shamed person who is focusing on negative self-evaluations would naturally be drawn to a concern over others’ evaluations. It’s a short leap from thinking what a horrible person one is to thinking about how one might be evaluated by others. On the other hand, a person experiencing guilt is already relatively “decentered”—focusing on a negative behavior somewhat separate from the self. In focusing on a bad behavior, rather than a bad self, a person in the middle of a guilt experience is more likely to recognize (and have concerns about) the effects of that behavior on others rather than on others’ evaluations. Several subsequent studies (Smith et al. 2002) provide ample evidence that shame is associated with such concerns. For example, participants primed to focus on public exposure of a moral transgression attributed equivalent levels of shame and guilt to story protagonists, but when the public versus private dimension was not highlighted, participants attributed less shame (guilt was uniformly high across conditions). However, taken together, Smith et al.’s findings are consistent with the notion that people focus on others’ evaluations because they are feeling shame, not vice versa. When participants were asked to think of a situation in which they had felt bad because an inferior aspect of themselves “was revealed or publicly exposed to another person or to other people” (p. 154; emphasis added), the majority spontaneously described the resulting feeling as one of embarrassment—only 6.7% identified the feeling as shame (twice as many identified the feeling as guilt). Similarly, in the moral condition (feeling bad because “something wrong” that they did was exposed), the modal emotion term was embarrassment—three times more common than shame (which was no more frequent than guilt). In short, when experiencing shame, people may feel more exposed—more aware of others’ disapproval—but the reality is that situations causing both shame and guilt are typically social in nature. More often than not, our failures and transgressions do not escape the notice of others.

The currently most dominant basis for distinguishing between shame and guilt—focus on self versus behavior—was first proposed by Helen Block Lewis (1971) and more recently elaborated by Tracy & Robins’s (2004a) appraisal-based model of self-conscious emotions. According to Lewis (1971), shame involves a negative evaluation of the global self; guilt involves a negative evaluation of a specific behavior. Although this distinction may, at first glance, appear rather subtle, empirical research supports that this differential emphasis on self (“I did that horrible thing”) versus behavior (“I did that horrible thing”) sets the stage for very different emotional experiences and very different patterns of motivations and subsequent behavior.

Both shame and guilt are negative emotions and as such, both can cause intrapsychic pain. Nonetheless, shame is considered the more painful emotion because one’s core self—not simply one’s behavior—is at stake. Feelings of shame are typically accompanied by a sense of shrinking or of “being small” and by a sense of worthlessness and power-lessness. Shamed people also feel exposed. Although shame does not necessarily involve an actual observing audience present to witness one’s shortcomings, there is often the imagery of how one’s defective self would appear to others. Lewis (1971) described a split in self-functioning in which the self is both agent and object of observation and disapproval. Guilt, on the other hand, is typically a less devastating, less painful experience because the object of condemnation is a specific behavior, not the entire self. Rather than needing to defend the exposed core of one’s identity, people in the throes of guilt are drawn to consider their behavior and its consequences. This focus leads to tension, remorse, and regret over the “bad thing done.”

Empirical support for Lewis’s (1971) distinction between shame and guilt comes from a range of experimental and correlational studies employing a range of methods including qualitative case study analyses, content analyses of shame and guilt narratives, participants’ quantitative ratings of personal shame and guilt experiences, analyses of attributions associated with shame and guilt, and analyses of participants’ counterfactual thinking (for a review, see Tangney & Dearing 2002). Most recently, for example, Tracy & Robins (2006) employed both experimental and correlational methods showing that internal, stable, uncontrollable attributions for failure were positively related to shame, whereas internal, unstable, controllable attributions for failure were positively related to guilt.

Shame and guilt are not equally “moral” emotions

One of the consistent themes emerging from empirical research is that shame and guilt are not equally “moral” emotions. On balance, guilt appears to be the more adaptive emotion, benefiting individuals and their relationships in a variety of ways (Baumeister et al. 1994, 1995a,b; Tangney 1991, 1995a,b), but there is growing evidence that shame is a moral emotion that can easily go awry (Tangney 1991, 1995a,b; Tangney et al. 1996b).

In this section, we summarize research in five areas that illustrates the adaptive functions of guilt, in contrast to the hidden costs of shame. Specifically, we focus on the differential relationship of shame and guilt to motivation (hiding versus amending), other-oriented empathy, anger and aggression, psychological symptoms, and deterrence of transgression and other risky, socially undesirable behavior.

Hiding versus amending

Research consistently shows that shame and guilt lead to contrasting motivations or “action tendencies” (Ketelaar & Au 2003, Lewis 1971, Lindsay-Hartz 1984, Tangney 1993, Tangney et al. 1996a, Wallbott & Scherer 1995, Wicker et al. 1983). On the one hand, shame corresponds with attempts to deny, hide, or escape the shame-inducing situation. Physiological research has linked the shame experience with elevated levels of proinflammatory cytokine and cortisol (Dickerson et al. 2004a), which can trigger postural signs of deference and self-concealment (see New Directions in Research on Shame and Guilt: Physiological Correlates of Shame). Guilt, on the other hand, corresponds with reparative actions including confessions, apologies, and undoing the consequences of the behavior. On the whole, empirical evidence evaluating the action tendencies of people experiencing shame and guilt suggests that guilt promotes constructive, proactive pursuits, whereas shame promotes defensiveness, interpersonal separation, and distance.

Other-oriented empathy versus self-oriented distress

Second, shame and guilt are differentially related to empathy. Specifically, guilt goes hand in hand with other-oriented empathy. Feelings of shame, in contrast, apparently disrupt individuals’ ability to form empathic connections with others. This differential relationship of shame and guilt to empathy is apparent both at the level of emotion disposition and at the level of emotional state. Research on emotional dispositions (Joireman 2004; Leith & Baumeister 1998; Tangney 1991, 1995b; Tangney & Dearing 2002) demonstrates that guilt-proneness consistently correlates with measures of perspective-taking and empathic concern. In contrast, shame-proneness is (depending on assessment method) negatively or negligibly correlated with other-oriented empathy and positively linked with the tendency to focus egocentrically on one’s own distress. Similar findings arise in research on emotional states—feelings of shame and guilt “in the moment.” In describing personal experiences of guilt, people convey greater empathy for others than when describing shame experiences (Leith & Baumeister 1998, Tangney et al. 1994). Marschall (1996) found that people induced to feel shame subsequently reported less empathy for a disabled student, especially among low-shame-prone individuals.

Why might shame, but not guilt, interfere with other-oriented empathy? Shame’s inherently egocentric focus on the “bad self” (as opposed to the bad behavior) derails the empathic process. Individuals in the throes of shame turn tightly inward, and are thus less able to focus cognitive and emotional resources on the harmed other (Tangney et al. 1994). In contrast, people experiencing guilt are specifically focused on the bad behavior, which in turn highlights the negative consequences experienced by others, thereby fostering an empathic response and motivating people to “right the wrong.”

Constructive versus destructive reactions to anger

Third, research indicates a robust link between shame and anger, again observed at both the dispositional and state levels. In her earlier clinical case studies, Helen Block Lewis (1971) observed the peculiar dynamic between shame and anger (or humiliated fury), noting that clients’ feelings of shame often preceded expressions of anger and hostility in the therapy room. More recent empirical research has supported her claim. Across individuals of all ages, proneness to shame is positively correlated with anger, hostility, and the propensity to blame factors beyond the self for one’s misfortunes (Andrews et al. 2000, Bennett, et al. 2005, Harper & Arias 2004, Paulhus et al. 2004, Tangney & Dearing 2002).

In fact, compared with those who are not shame-prone, shame-prone individuals are more likely to engage in externalization of blame, experience intense anger, and express that anger in destructive ways, including direct physical, verbal, and symbolic aggression, indirect aggression (e.g., harming something important to the target, talking behind the target’s back), all manner of displaced aggression, self-directed aggression, and anger held in (a ruminative unexpressed anger). Finally, shame-prone individuals report awareness that their anger typically results in negative long-term consequences for both themselves and for their relationships with others.

Guilt-proneness, in contrast, is consistently associated with a more constructive constellation of emotions, cognitions, and behaviors. For example, proneness to “shame-free” guilt is positively correlated with constructive intentions in the wake of wrongdoing and consequent constructive behaviors (e.g., nonhostile discussion, direct corrective action). Compared with their nonguilt-prone peers, guilt-prone individuals are less likely to engage in direct, indirect, and displaced aggression when angered. And they report positive long-term consequences to their anger (Tangney et al. 1996a). Consistent with these findings, Harper et al. (2005) recently evaluated the link between shame-proneness and perpetration of psychological abuse in the dating relationships by heterosexual college men. Shame proneness was significantly correlated with perpetration of psychological abuse, and men’s anger mediated this relationship.

Shame and anger have been similarly linked at the situational level, too (Tangney et al. 1996a, Wicker et al. 1983). For example, in a study of anger episodes among romantically involved couples, shamed partners were significantly more angry, more likely to engage in aggressive behavior, and less likely to elicit conciliatory behavior from their perpetrating significant other (Tangney 1995b). Taken together, the results provide a powerful empirical example of the shame-rage spiral described by Lewis (1971) and Scheff (1987), with (a) partner shame leading to feelings of rage, (b) and destructive retaliation, (c) which then sets into motion anger and resentment in the perpetrator, (d ) as well as expressions of blame and retaliation in kind, (e) which is then likely to further shame the initially shamed partner, and so forth—without any constructive resolution in sight.

Recently, Stuewig et al. (2006) examined mediators of the link between moral emotions and aggression in four samples. We theorized that negative feelings associated with shame lead to externalization of blame, which in turn leads shame-prone people to react aggressively. Guilt, on the other hand, should facilitate empathic processes, thus reducing outward directed aggression. As anticipated, we found that across all samples, externalization of blame mediated the relationship between shame-proneness and both verbal and physical aggression. Guilt-proneness, on the other hand, continued to show a direct inverse relationship to aggression in three of the four samples. In addition, the link between guilt and low aggression was partially mediated through other-oriented empathy and a propensity to take responsibility.

In short, shame and anger go hand in hand. Desperate to escape painful feelings of shame, shamed individuals are apt to turn the tables defensively, externalizing blame and anger outward onto a convenient scapegoat. Blaming others may help individuals regain some sense of control and superiority in their life, but the long-term costs are often steep. Friends, coworkers, and loved ones are apt to become alienated by an interpersonal style characterized by irrational bursts of anger.

Psychological symptoms

When considering the domain of social behavior and interpersonal adjustment, empirical research suggests that guilt, on balance, is the more moral or adaptive emotion. Guilt appears to motivate reparative action, foster other-oriented empathy, and promote constructive strategies for coping with anger. But are there intrapersonal or intrapsychic costs for those individuals who are prone to experience guilt? Does guilt-proneness lead to anxiety, depression, and/or a loss of self-esteem? Conversely, is shame perhaps less problematic for intrapersonal as opposed to interpersonal adjustment?

The answer is clear in the case of shame. Research over the past two decades consistently indicates that proneness to shame is related to a wide variety of psychological symptoms. These run the gamut from low self-esteem, depression, and anxiety to eating disorder symptoms, posttraumatic stress disorder (PTSD), and suicidal ideation (Andrews et al. 2000, Ashby et al. 2006, Brewin et al. 2000, Crossley & Rockett 2005, Feiring & Taska 2005, Feiring et al. 2002, Ferguson et al. 2000, Ghatavi et al. 2002, Harper & Arias 2004, Henderson & Zimbardo 2001, Leskela et al. 2002, Mills 2003, Murray et al. 2000, Orsillo et al. 1996, Sanftner et al. 1995, Stuewig & McCloskey 2005; see also review in Tangney & Dearing 2002). The negative psychological implications of shame are evident across measurement methods, diverse age groups, and populations. Both the clinical literature and empirical research agree that people who frequently experience feelings of shame about the self are correspondingly more vulnerable to a range of psychological problems.

Although the traditional view is that guilt plays a significant role in psychological symptoms, the empirical findings have been more equivocal. Clinical theory and case studies make frequent reference to a maladaptive guilt characterized by chronic self-blame and obsessive rumination over one’s transgressions (Blatt 1974, Ellis 1962, Freud 1924/1961, Hartmann & Loewenstein 1962, Rodin et al. 1984, Weiss 1993). Recently, however, theorists and researchers have emphasized the adaptive functions of guilt, particularly for interpersonal behavior (Baumeister et al. 1994, 1995a; Hoffman 1982; Tangney 1991, 1994, 1995b; Tangney et al. 1992; Tangney & Dearing 2002).

In an effort to reconcile these perspectives, Tangney (1996) argued that earlier work failed to take into account the distinction between guilt and shame. Once one conceptualizes guilt as a negative emotion in response to a specific failure or transgression, there’s no compelling reason to expect guilt to be associated with poor psychological adjustment. Instead, guilt is most likely to be maladaptive when it becomes fused with shame. The advantages of guilt are lost when a person’s guilt experience (“Oh, look at what a horrible thing I have done”) is magnified and generalized to the self (“… and aren’t I a horrible person”). Ultimately, it’s the shame component of this sequence—not the guilt component— that poses the problem, as the person becomes saddled with feelings of contempt and disgust for a bad, defective self.

Moreover, such painful feelings of shame are difficult to resolve. Shame—and, shame-fused guilt—offers little opportunity for redemption. It is a daunting challenge to transform a self that is defective at its core. Thus, guilt with an overlay of shame is most likely the source of the painful self-castigation and rumination so often described in the clinical literature. In contrast, there are typically a multitude of paths to redemption in the case of uncomplicated guilt focused on a specific behavior. A person (a) often has the option of changing the objectionable behavior; (b) or even better yet, has an opportunity to repair the negative consequences; (c) or at the very least, can extend a heartfelt apology. And when it is not possible to make these external amends, one can resolve to do better in the future.

Consistent with this conceptual analysis, empirical studies that fail to take into account the distinction between shame and guilt, or that employ adjective checklist-type (and other globally worded) measures that are ill-suited to distinguish between shame and guilt, report that guilt-proneness is associated with psychological symptoms (Boye et al. 2002, Fontana & Rosenbeck 2004, Ghatavi et al. 2002, Harder 1995, Jones & Kugler 1993, Meehan et al. 1996). For example, using the Interpersonal Guilt Questionnaire (O’Connor et al. 1997), Berghold & Locke (2002) found that solely the “self-hate” guilt scale differentiated between a control group and adolescents diagnosed with anorexia nervosa. (The authors concluded that, in fact, shame—not guilt—is more important to a clinical understanding of this eating disorder.)

On the other hand, measures sensitive to Lewis’s (1971) distinction between shame about the self versus guilt about a specific behavior (e.g., scenario-based methods assessing shame and guilt with respect to specific situations) show that the propensity to experience “shame-free” guilt is essentially unrelated to psychological symptoms. Numerous independent studies converge: guilt-prone children, adolescents, and adults are not at increased risk for depression, anxiety, low self-esteem, etc. (Gramzow & Tangney 1992; Leskela et al. 2002; McLaughlin 2002; Quiles & Bybee 1997; Schaefer 2000; Stuewig & McCloskey 2005; Tangney 1994; Tangney & Dearing 2002; Tangney et al. 1991, 1992, 1995).

It is worth noting, however, that in most scenario-based measures of shame and guilt (including the Test of Self-Conscious Affect, or TOSCA), the majority of situations are relatively ambiguous regarding responsibility or culpability. For the negatively valenced (but not positively valenced) situations, respondents are asked to imagine events in which they clearly failed or transgressed in some way. Problems are likely to arise when people developed an exaggerated or distorted sense of responsibility for events beyond their control or for which they have no personal involvement (Ferguson et al. 2000, Tangney & Dearing 2002, Zahn-Waxler & Robinson 1995). Survivor guilt is a prime example of such a problematic guilt response that has been consistently linked to psychological maladjustment (Kubany et al. 1995, 2004; O’Connor et al. 2002). In an experimental study of elementary school–aged children, Ferguson et al. (2000) varied the degree to which situations in a scenario-based measure were ambiguous with respect to responsibility. They found a positive relationship between internalizing symptoms (e.g., depression) and proneness to guilt specifically in situations where responsibility was ambiguous.

In short, the benefits of guilt are evident when people acknowledge their failures and transgressions and take appropriate responsibility for their misdeeds. In such situations, the interpersonal benefits of guilt do not appear to come at a cost to the individual. The propensity to experience “shame-free” guilt in response to clear transgressions is generally unrelated to psychological problems, whereas shame is consistently associated with maladaptive processes and outcomes at multiple levels.

Linking moral emotions to risky, illegal, and otherwise inadvisable behavior

Because shame and guilt are painful emotions, it is often assumed that they motivate individuals to avoid doing wrong. From this perspective, anticipated shame and guilt should decrease the likelihood of transgression and impropriety. But what exactly do the data show?

Empirical studies of diverse samples, employing a range of measures, clearly indicate that guilt-proneness is inversely related to antisocial and risky behavior. In a study of college undergraduates (Tangney 1994), guilt-proneness was associated with endorsing such items as “I would not steal something I needed, even if I were sure I could get away with it.” Similarly, Tibbetts (2003) found that college students’ guilt-proneness was inversely related to self-reported criminal activity. Among adolescents, proneness to shame-free guilt has been negatively correlated with delinquency (Merisca & Bybee 1994, Stuewig & McCloskey 2005; although Ferguson et al.1999 found a negative relationship between guilt-proneness and externalizing symptoms among boys, the opposite was true for girls). The moral emotions appear to be well established by middle childhood and have implications for moral behavior for years to come (Tangney & Dearing 2002). Children prone to shame-free guilt in the fifth grade were, in adolescence, less likely to be arrested, convicted, and incarcerated. They were more likely to practice safe sex, and they were less likely to abuse drugs. Importantly, these findings held when controlling for family income and mothers’ education. Guilt-prone college students, too, are less likely to abuse drugs and alcohol (Dearing et al. 2005). Even among adults already at high risk, guilt-proneness appears to serve a protective function. In a longitudinal study of jail inmates, guilt-proneness assessed shortly after incarceration negatively predicted recidivism and substance abuse during the first year post-release (Tangney et al. 2006).

The pattern of results for shame is quite different, with virtually no evidence supporting the presumed adaptive nature of shame. In studies of children, adolescents, college students, and jail inmates, shame does not appear to serve the same inhibitory functions as guilt (Dearing et al. 2005, Stuewig & McCloskey 2005, Tangney et al. 1996b). To the contrary, research suggests that shame may even make things worse. In a study of children, Ferguson et al. (1999) found that shame-proneness was positively correlated with externalizing symptoms on the Child Behavior Checklist. In a sample of college students, Tibbetts (1997) found a positive relationship between shame-proneness and intentions toward illegal behavior. Shame-proneness assessed in the fifth grade predicted later risky driving behavior, earlier initiation of drug and alcohol use, and a lower likelihood of practicing safe sex (Tangney & Dearing 2002). Similarly, proneness to problematic feelings of shame has been positively linked to substance use and abuse in adulthood (Dearing et al. 2005, Meehan et al. 1996, O’Connor et al. 1994, Tangney et al. 2006).

The differential link of shame and guilt to moral behavior may not generalize across all populations with respect to all behaviors. Harris (2003) assessed event-specific experiences of shame and guilt among drunk-driving offenders following their appearance in court or at a restorative justice conference. In contrast to most extant studies, Harris found no evidence that shame and guilt form distinct factors. It’s important to note that this study focused on a unique,homogeneous sample (convicted drunk drivers, many of whom have substance abuse problems) and a single type of transgression. Harris’s findings raise the intriguing possibility that individuals with substance abuse problems may not have well-differentiated experiences of shame and guilt. Alternatively, guilt and its attendant empathic focus on the harmed other may be less relevant to transgressions, such as drunk driving, that typically do not result in objective physical harm to others. (That is, the magnitude of consequences of an automobile accident is potentially huge, whereas the probability of its occurrence on any given occasion is rather small. Most drunk-driving offenders are arrested for erratic driving, not at the scene of an accident involving actual harm to another person.)

In sum, empirical results converge, indicating that guilt but not shame is most effective in motivating people to choose the moral paths in life. The capacity for guilt is more apt to foster a lifelong pattern of moral behavior, motivating individuals to accept responsibility and take reparative action in the wake of the occasional failure or transgression. In contrast, research has linked shame with a range of illegal, risky, or otherwise problematic behaviors. Thus, when considering the welfare of the individual, his or her close relationships, or society, feelings of guilt represent the moral emotion of choice.

New directions in research on shame and guilt

Context- or domain-specific shame and guilt

Some clinicians have lamented the research literature’s heavy focus on dispositional shame (Leeming & Boyle 2002). Andrews (1998) notes that at least three different conceptualizations of the high-shame individual are implicit in the range of current dispositional measures of shame. Some researchers conceptualize shame-proneness as the propensity to experience shame across a range of situations (operationalized by scenario-based measures such as the TOSCA-3). Others conceptualize high-shame individuals as those who frequently or continuously experience global shame, an affect not necessarily connected to particular events (operationalized by global adjective checklists, such as the PFQ-2, and by the Internalized Shame Scale). A third and more recent conceptualization of “high shame” is explicitly domain specific— individuals who are chronically shamed about particular circumscribed behaviors or personal characteristics (e.g., physical appearance, level of education, race/ethnicity, and stuttering).

A number of researchers have developed measures to assess shame and guilt with respect to specific domains. For example, researchers concerned with the psychology of eating disorders and those exploring hypotheses drawn from the Objectification Theory of Frederickson & Roberts (1997) have assessed feelings of shame specifically in reference to one’s body. “Body shame” has been consistently associated with self-objectification and eating disorder symptoms (Hallsworth et al. 2005). Andrews (1995, 1998) has examined the link between childhood abuse and body shame (see below).

Regarding guilt, researchers have begun to examine the nature and implications of domain-specific feelings of guilt associated with trauma. Trauma-related guilt cognitions, such as false beliefs about responsibility or pre-outcome knowledge, are reliably associated with symptoms of depression among diverse samples of trauma survivors (Blacher 2000; Kubany et al. 1995, 2004; Lee et al. 2001). Moreover, cognitive processing therapy and prolonged exposure interventions appear to be effective at reducing trauma-related guilt cognitions (Nishith et al. 2005, Resick et al. 2002).

Styles of coping with the shame (and guilt) experience

Most theory and research on shame and guilt has focused on the events that lead up to these emotional experiences, the phenomenology of these emotions, or the consequences of these emotions for motivation and behavior. Less attention has been directed toward how people cope with aversive feelings of shame and guilt. Drawing on Nathanson’s (1992) Compass of Shame theory, Elison et al. (2006a) developed a measure of individual differences in coping with shame. The Compass of Shame Scale (COSS-4) consists of four 10-item scales representing the poles of Nathanson’s Compass of Shame plus a fifth assessing adaptive responses. More specifically:

• ○“Attack Self” assesses inward-directed anger and blame (e.g., self-disgust)
• ○“Withdrawal” assesses the tendency to hide or withdraw when shamed (e.g., avoid others)
• ○“Avoidance” assesses disavowal and emotional distancing or minimization (e.g., minimizing the importance of a failing grade)
• ○“Attack Other” assesses outward-directed anger and blame (e.g., blaming someone else for the failure or transgression)
• ○“Adaptive” assesses acknowledgment of shame and motivation to apologize and/or make amends

Some clear parallels exist between the scales of the COSS-4 and the scales of the TOSCA. Attack Self and Withdrawal bear a close resemblance to the two types of items that comprise the TOSCA Shame scale— negative self-appraisals and avoidance. The Adaptive Responses scale bears a close resemblance to the TOSCA Guilt scale. Attack Other bears considerable resemblance to the TOSCA Externalization of Blame scale. And Avoidance resembles the TOSCA Detachment scale (although the TOSCA Detachment scale appears less internally consistent than the COSS-4 Avoidance scale). The use of different terms to describe similar types of response most likely reflects differences in theoretical formulation. Nathanson (1992) draws on modern psychoanalytic theory, Tomkin’s (1963) emotion theory, and associated attachment theory. The TOSCA measures were informed by social-cognitive theories of emotion, with much influence from Lewis (1971). As one might expect, of the four scales of the COSS-4, the Withdrawal and Attack Self scales are most highly correlated with shame, as well as measures of more general psychological adjustment (Elison et al. 2006a,b).

In a sample of undergraduates, Campbell & Elison (2005) found that both subscales of the Self-Report Psychopathy Scale (SRPS) were negatively related to the guilt-like Adaptive Response to Shame scale and positively related to Attack Others and Avoidance scales. The SRPS subscale assessing antisocial lifestyle paralleling Hare’s (1991) Factor 2 on the Psychopathy Checklist-Revised was positively correlated with Attack Self and Withdrawal scales—scales that assess shame much as defined by Lewis (1971) and Tangney (1996). But the Primary Psychopathy subscale, assessing psychopathic personality features akin to Hare’s (1991) Factor 1, was negatively or negligibly related to shame per se—the Attack Self and Withdrawal scales. Future research examining how correlates of the COSS-4 parallel or differ from the TOSCA is needed.

Psychobiological correlates of shame

Researchers have recently begun to evaluate psychobiological markers of shame, examining biological responses to laboratory manipulations designed to threaten the social self (Dickerson et al. 2004b, Gruenewald et al. 2004; see Dickerson et al. 2004a for a review). Dickerson et al. found that participants who wrote about incidents wrought with self-blame, in contrast to participants who wrote about daily activities, evidenced increased levels of self-reported shame (and guilt) from pretest to post-test. More importantly, these same participants evidenced increased proinflammatory cytokine activity from pretest to post-test, and this response was significantly predicted by increases in self-reported shame. Consistent with theory differentiating shame and guilt, shame uniquely predicted this immune-related response; changes in neither guilt nor general negative affect significantly predicted changes in the cytokine response. Recent applied research is equally suggestive: Among people with HIV, persistent feelings of shame predicted t-cell decline, an indicator of compromised immune function (Weitzman et al. 2004).

Gruenewald et al. (2004) examined cortisol responses of individuals performing stressful speaking and arithmetic tasks with and without an audience. Individuals in the social evaluation condition reported more shame (and lower self-esteem) than did individuals in the nonevaluative condition. Moreover, those individuals in this shame-eliciting condition also evidenced significant increases in cortisol levels. Similarly, among children, nonverbal expressions of shame and embarrassment during laboratory tasks were associated with greater cortisol changes during the session, relative to other nonverbal behavioral styles (Lewis & Ramsay 2002).

Considering these patterns of immunoresponse in toto, Dickerson et al. (2004a) note, “… shame may be experienced particularly in conditions characterized by negative social evaluation and rejection. The cortisol and proinflammatory cytokine systems also appear to be responsive to social-evaluative threat. While tentative, there is support for the notion that the activation of these systems under the very specific condition of threat to the social self may hinge on the experience of shame and related emotions” (p. 1205).

Cardiovascular reactivity seems likewise associated with experiences of shame. For example, in addition to evaluating cortisol response, Gruenewald et al. (2004) also evaluated heart rate and blood pressure changes in response to the stressful speaking and arithmetic tasks. Although heart rate and systolic blood pressure increased in both the social evaluative and nonevaluative conditions, the response was somewhat more marked in the social evaluative condition. Extending this work with a clever laboratory manipulation of experienced emotions, Herrald & Tomaka (2002) evaluated cardiovascular reactivity in the wake of pride, shame, and anger. They found that the negatively valenced emotions of shame and anger resulted in higher levels of cardiovascular reactivity than pride; importantly, participants in the shame condition showed higher peripheral resistance (associated with hypertension) and participants in the anger condition showed higher cardiac contractility (associated with coronary disease).

In sum, there seems to be distinct physiological correlates corresponding to the experience of shame. Such physiological markers may prove to be useful as a measurement tool in future research on situation-specific states of shame.

Childhood abuse and the propensity to experience shame

Clinicians have long reported that victims of abuse or trauma are often haunted by feelings of shame. This may especially be true in cases of child maltreatment because of its secretive and hidden nature (Deblinger & Runyon 2005). The experience of abuse at a young age may instigate and reinforce shame-inducing thoughts (Andrews 1998). Also, severely punitive parenting practices may engender in children feelings of helplessness and self-blame, which may then lead to a globalized sense of shame. Although child maltreatment in its different forms (physical abuse, sexual abuse, neglect, harsh parenting) has long been theorized to engender a vulnerability to shame, systematic empirical research has been conducted only recently.

A number of studies have found a relationship between childhood physical and sexual abuse and specific forms of shame, including body shame (Andrews 1995, Andrews & Hunter 1997) and shame about a traumatic event (Andrews et al. 2000). In addition, Murray & Waller (2002) found a relationship between unwanted sexual experience of any sort and internalized shame. Although Hoglund & Nicholas (1995) reported no relationship between a history of physical abuse and shame-proneness, they did find a link between shame-proneness and history of emotional abuse. In this same vein, Gilbert et al. (1996) found that put-downs and shaming practices by parents were associated with adult children’s shame-proneness. Each of these studies, however, was based on retrospective reports of maltreatment and parenting practices, which have known weaknesses (Widom et al. 2004).

Nonetheless, when considering studies using prospective or observational designs, the results for nonsexual abuse and shame are similar. Bennett et al. (2005) report an association between physical abuse and nonverbal shame, although there was not a significant relationship for neglect. In addition, Alessandri & Lewis (1996) found girls coded as maltreated to have higher nonverbal shame. More generally, negative or harsh parenting has been associated with the propensity to experience shame (Alessandri & Lewis 1993, 1996; Ferguson & Stegge 1995; Mills 2003). Stuewig & McCloskey (2005) report a relationship between harsh parenting in childhood and shame-proneness in adolescence, a relationship that was mediated by rejecting parenting practices also measured in adolescence.

The relationship between sexual abuse and shame seems to be less straightforward. In research studies of individuals who have experienced sexual abuse, shame has been consistently implicated in poor outcomes such as depression and PTSD symptoms (Feiring & Taska 2005; Feiring et al. 1996, 2002; Talbot et al. 2004). Feiring & Taska (2005) have also found abuse-specific shame to be moderately stable across time.

However, neither Alessandri & Lewis (1996), using observational measures of shame, nor Stuewig & McCloskey (2005), using self-reports of shame-proneness, found a relationship between history of sexual abuse and shame, but both studied small samples of sexually abused individuals. Another reason for these null findings may be that complex emotions surround not only the abusive act but also how the individual copes with the experience. Using facial coding data for shame, Bonanno et al. (2002) found that individuals with a documented history of sexual abuse who did not disclose the abuse in an interview had higher levels of observed shame than those individuals who did disclose their sexual abuse history. There was no difference in shame between those who did disclose and a nonabused comparison group. In a follow-up (Negrao et al. 2005), individuals who did disclose their sexual abuse history were higher on shame coded from narratives compared with those who did not disclose and those in a nonabused comparison group.In other words, individuals who disclosed their abuse histories expressed more shame verbally, whereas those who did not disclose expressed more shame nonverbally, relative to control participants.

In sum, the findings regarding the relationship between childhood abuse and subsequent difficulties with shame are mixed, no doubt due in part to the fact that studies have employed different measures and conceptualizations of both maltreatment and shame (Berliner 2005). Nonetheless, taken together, the weight of evidence suggests that people who experience maltreatment in childhood are somewhat more vulnerable to shame issues later in life.

Vicarious or “collective” shame and guilt: group-based self-conscious emotion

Thus far, this review has focused almost exclusively on shame and guilt experienced in reaction to one’s own misdeeds. In recent years, a number of investigators have substantially expanded the literature on self-conscious emotions by considering “vicarious” or “group-based” shame and guilt—feelings experienced in response to the transgressions and failures of other individuals. This research represents an exciting integration of self-conscious emotions theory with the social psychological literature on social identity, group, and intergroup processes. To the extent that the self is, in part, defined by our interpersonal relations and group memberships, it is possible to construe the behavior of an in-group member as reflecting on the self. Thus, personal causality is not always a prerequisite for the experience of shame or guilt.

In many ways, the phenomena of vicarious shame and guilt parallel personal shame and guilt experiences. Lickel, Schmader, and colleagues (Lickel et al. 2004, 2005) have developed a process model linking specific types of appraisals with vicarious experiences of shame and guilt, respectively. They present compelling evidence that group-based shame is most likely elicited when a threatened shared identity is salient—that is, when concerns about maintaining a positive group identity arise. Vicarious guilt, on the other hand, is more likely when one’s interpersonal dependence with the perpetrator is salient, and when relational-based concerns are highlighted by a focus on harm to another group or individual. For example, Lickel et al. (2005) found that vicarious shame (but not guilt) experiences were positively related to their ratings of the relevance of an offending behavior to the identity shared by the respondent and the perpetrator. The link between identity concerns and vicarious or group-based shame are evident in both correlational and experimental studies (Iyer et al. 2006, Schmader & Lickel 2006).

Degree of interdependence with the perpetrator appears to be uniquely related to vicarious guilt (Lickel et al. 2005). However, identification with the perpetrating group can also have implications for vicarious, group-based guilt as well (Branscombe & Doosje 2004, Doosje et al. 1998), especially when individuals are prompted to focus on the harm done (Iyer et al. 2003)

Of particular applied relevance to current international conflicts, when people are provided with ambiguous information about group members’ transgressions, those who are highly identified with the group appear to capitalize on the ambiguity, reporting less vicarious shame (Johns et al. 2005) and group-based guilt (Doosje et al. 1998) relative to those who are less identified, and whose self is presumably less threatened.

As with personal guilt experiences, group-based guilt has been associated with empathy (Zebel et al. 2004) and a motivation to repair or make amends (Iyer et al. 2003, Lickel et al. 2005, Swim & Miller 1999, Zebel et al. 2004). And as with personal shame experiences, vicarious group-based shame (but not guilt) has been linked to a desire to distance oneself from the shame-eliciting event (Johns et al. 2005, Lickel et al. 2005). Furthermore, the link between anger and shame is evident when considering vicarious shame (Iyer et al. 2006, Johns et al. 2005, Schmader & Lickel 2006). Nonetheless, there are some indications that vicarious or group-based shame may have a kinder, gentler side than personal shame. For example, under some circumstances, group-based shame appears to motivate a desire to change the image of the group in a proactive fashion (Lickel et al. 2006).

Embarrassment

Embarrassment appears to be less centrally relevant to the domain of morality than are shame and guilt. For example, adults’ ratings of personal shame-, guilt-, and embarrassment-eliciting events indicate that when people feel embarrassed, they are less concerned with issues of morality than when they feel shame or guilt (Tangney et al. 1996a). Nonetheless, certain conditions exist under which embarrassment may support or undermine people’s efforts to live life in a manner consistent with their moral standards.

Miller (1995) defines embarrassment as “an aversive state of mortification, abashment, and chagrin that follows public social predicaments” (p. 322). Embarrassment accounts from hundreds of high school students and adults (Miller 1992) indicate that the most common causes of embarrassment are “normative public deficiencies”—situations in which a person behaves in a clumsy, absent-minded, or hapless way (tripping in front of a crowd, forgetting someone’s name, unintended bodily-induced noises). Other common embarrassment-inducing situations include awkward social interactions and being conspicuous (e.g., during the “birthday” song). Generally, events causing embarrassment seem to signal that something is amiss— some aspect of the self or one’s behavior needs to be carefully monitored, hidden, or changed.

The motivations prompted by embarrassment, however, may have implications for moral behavior. Research indicates that embarrassed people are inclined to behave in conciliatory ways in order to win approval and (re)inclusion from others (Cupach & Metts 1990, 1992; Leary et al. 1996; Miller 1996; Sharkey & Stafford 1990). In other words, upon feeling embarrassment (or to avoid embarrassment), people are inclined to conform and curry favor. Thus, depending on the local norms of the immediate social environment, embarrassment may prompt adherence to broadly accepted moral standards or to locally endorsed deviant acts.

As with shame and guilt, there are individual differences in the degree to which people are prone to experience embarrassment. Research has shown that embarrassability is associated with neuroticism, high levels of negative affect, self-consciousness, and a fear of negative evaluation from others (Edelmann & McCusker 1986, Leary & Meadows 1991, Miller 1995b). To the extent that embarrassment-prone individuals are highly aware of and concerned with social rules and standards, they may be especially vulnerable to the influence of peer pressure.

Moral Pride

Thus far, this chapter has focused on negatively valenced moral emotions. We turn now to one of the long-neglected positively valenced moral emotions—morally relevant experiences of pride. Of the self-conscious emotions, pride is the neglected sibling. Mascolo & Fischer (1995) define pride as an emotion “generated by appraisals that one is responsible for a socially valued outcome or for being a socially valued person” (p. 66). From their perspective, pride serves to enhance people’s self-worth and, perhaps more importantly, to encourage future behavior that conforms to social standards of worth or merit (see also Barrett 1995).

Most theoretical and empirical research on pride emphasizes achievement-oriented pride (Tracy & Robins 2004b). Although pride may most often arise in response to scholastic, occupational, or athletic achievement, self-conscious experiences of pride in moral contexts may be an important component of our moral emotional apparatus. Feelings of pride for meeting or exceeding morally relevant standards (and for inhibiting impulses to behave immorally) may serve important motivational functions, rewarding and reinforcing one’s commitment to ethics of autonomy, community, and divinity.

In parallel to the self-versus-behavior distinction of guilt and shame, it may be useful to distinguish between two types of pride. Along similar lines, Tangney (1990) distinguished between “alpha” pride (pride in self) and “beta” pride (pride in behavior), M.Lewis (1992) distinguished between hubris (pride-fulness) and pride (experienced in reference to a specific action or behavior), and Tracy & Robins (2004b) distinguished between hubris and more event-specific achievement-oriented pride. Tracy & Robins (2006), drawing on multiple methods, present compelling empirical evidence for these two types of pride.

Little empirical research has been conducted on individual differences in proneness to pride in self (or pride in behavior, for that matter). The Tests of Self-Conscious Affect (e.g., Tangney et al. 1989; see Tangney & Dearing 2002 for details) each contain measures of the propensity to experience alpha pride and beta pride, respectively. These sub-scales, however, have very modest reliabilities, largely because they draw on only a few items. Thus, we and other investigators have made little use of these ancillary scales. Lewis (1992) views hubris as largely maladaptive, noting that hubristic individuals are inclined to distort and invent situations to enhance the self, which can lead to interpersonal problems. It remains to be seen how individual differences in pride or hubris relate to the capacity to self-regulate or to choose the moral path in life. One possibility is that pride and hubris represent the flip side of guilt and shame—one the “modern,” adaptive moral emotion and the other, its evil twin.

OTHER-FOCUSED MORAL EMOTIONS

Thus far, our review of theory and research on moral emotion has focused on the self-conscious emotions of shame, guilt, embarrassment, and pride. These emotions vary in valence and in attributions regarding the particular source of offense (e.g., self versus self’s behavior). But these self-conscious emotions are similar in that in each case, the emotion is elicited when some aspect of the self is scrutinized and evaluated with respect to moral standards. Recently, Haidt (2000, 2003) added importantly to our thinking about the nature of “moral emotions.” In his work, Haidt focuses primarily on the emotions of elevation and gratitude—emotions that are experienced when observing the admirable deeds of others, and that then motivate observers to engage in admirable deeds themselves.

In fact, by crossing the two dimensions of focus (self versus other) and valence (positive versus negative), one can conceptualize four categories of moral emotion (see Haidt 2003, following Ortony et al. 1988). To date most theory and research on moral affect has emphasized the negatively valenced self-conscious quadrant. With the advent of the positive psychology movement and Haidt’s groundbreaking work, we anticipate that the next decade will see exciting new developments in our understanding of the moral functions of negatively and positively valenced other-directed emotions.

Righteous Anger, Contempt, and Disgust

Anger is a negatively valenced, other-focused emotion not typically considered in the morally relevant sphere. People may experience anger for a very broad range of situations—e.g., when insulted, frustrated, inconvenienced, or injured in any one of a number of ways. According to appraisal theorists (Lazarus 1991, Roseman 1991, Smith & Ellsworth 1985), people typically feel angry when they appraise an event as personally relevant, inconsistent with their goals, and when the event appears to be caused (often intentionally) by a responsible other. The emphasis is on perceptions of actual or potential self-harm (e.g., a personally relevant goal has been thwarted or frustrated, a valued possession has been threatened or harmed) in conjunction with attributions of intentionality and/or responsibility on the part of the offending other.

Righteous anger, however, arises in response to a special class of anger-eliciting events, those in which the perpetrator’s behavior represents a violation of moral standards. In such cases, the harm need not be personally experienced. One can feel anger upon witnessing morally repulsive behavior aimed at a third party. Rozin et al. (1999) presented evidence that righteous anger tends to occur more specifically in response to violations of the ethic of autonomy—the ethic most familiar in Western culture. Righteous anger can serve moral functions in that it can motivate “third-party” bystanders to take action in order to remedy observed injustices.

The emotions of contempt and disgust also stem from negative evaluations of others, but seem somewhat less apt than righteous anger to motivate morally corrective action. Among participants in both the United States and Japan, Rozin et al. (1999) found that feelings of contempt were differentially linked to violations of the ethic of community (e.g., violations of social hierarchy), whereas feelings of disgust were linked to violations of the ethic of divinity (e.g., actions that remind us of our animal nature, such as defecation, problems with hygiene, etc., as well as assaults on human dignity, such as racism and abuse).

Elevation

Just as disgust is the moral emotion people experience when observing violations of the ethic of divinity, elevation is the positive emotion elicited when observing others behaving in a particularly virtuous, commendable, or superhuman way (Haidt 2000). In a study of college students, Haidt et al. (2002) explored the phenomenology of elevation, asking participants to recall “a manifestation of humanity’s ‘higher’ or ‘better’ nature.” Participants reported warm, pleasant, “tingling” feelings in their chest, they felt open to other people as their attention turned outward, and they felt motivated to help others and to become better people themselves. In this respect, elevation appears to be the quintessential positive emotion, especially apt to foster a “broaden and build” (Frederickson 2000) orientation to the world.

Gratitude

Gratitude is another example of an other-oriented, positively valenced moral affect. People are inclined to feel gratitude specifically in response to another person’s benevolence—that is, when they are the recipient of benefits provided by another, especially when those benefits are unexpected and/or costly to the benefactor. Gratitude is a pleasant affective state, distinct from indebtedness, which implies an obligation and is often experienced as a negative state.

McCullough et al. (2001) classify gratitude as a moral affect, not because the experience and expression of gratitude is in and of itself “moral,” but because feelings of gratitude (a) result from moral (e.g., prosocial, helping) behavior of the benefactor, and (b) engender subsequent moral motivation on the part of recipients. They observe that grateful people are often motivated to respond prosocially— both to their benefactor and toward others not involved in the gratitude-eliciting act. Moreover, expressions of gratitude can serve as a moral reinforcer, encouraging benefactors’ helping behavior in the future (Bennett et al. 1996, Clark et al. 1988, Goldman et al. 1982).

Gratitude not only benefits benefactors and relationships. Those who benefit most from the experience and expression of gratitude are grateful people themselves. In a series of experimental studies, feelings of gratitude enhanced psychological resilience, physical health, and the quality of daily life (Emmons & McCullough 2003). In fact, both dispositional and situation-specific episodes of gratitude have been linked to psychological well-being and adaptive behavior in non-clinical samples (Emmons & Shelton 2002; Frederickson et al. 2003; Kendler et al. 2003; McCullough et al. 2001, 2002) and among combat veterans with PTSD (Kashdan et al. 2006).

EMPATHY: A MORAL EMOTIONAL PROCESS

Finally, we discuss briefly a morally relevant emotional process—other-oriented empathy. (For a more complete review, see Eisenberg et al. 2004, 2006.) In contrast to the other moral emotions discussed in this review, empathy is not a discrete emotion. Rather it is an emotional process with substantial implications for moral behavior. Current conceptualizations of empathy integrate both affective and cognitive components. Feshbach (1975), for example, defines empathy as a “shared emotional response between an observer and a stimulus person.” She suggests that empathic responsiveness requires three interrelated skills or capacities: (a) the cognitive ability to take another person’s perspective, (b) the cognitive ability to accurately recognize and discriminate another person’s affective experience, and (c) the affective ability to personally experience a range of emotions (since empathy involves sharing another person’s emotional experience). Similarly, Coke and colleagues (1978) proposed a two-stage model of empathic responding, whereby perspective-taking facilitates empathic concern, which in turn leads to a desire to help.

Some researchers have made a distinction between “true” empathy and sympathy. Eisenberg (1986) explains that sympathy involves feelings of concern for the emotional state of another, but does not necessarily involve the vicarious experience of the other person’s feelings or emotions (e.g., emotional matching). Thus, one may feel concern (sympathy) for an angered individual without being vicariously angered oneself (an empathic reaction).

Others have distinguished between other-oriented empathy and self-oriented personal distress (Batson 1990, Batson & Coke 1981, Davis 1983). Other-oriented empathy involves taking another person’s perspective and vicariously experiencing similar feelings. These responses often involve feelings of sympathy and concern for the other person, and often lead to helping behavior. Importantly, the empathic individual’s focus remains on the experiences and needs of the other person, not on his or her own empathic response.In contrast, self-oriented personal distress involves a primary focus on the feelings, needs, and experiences of the empathizer. Empirical research underscores the importance of this distinction. Empathic concern for others has been linked to altruistic helping behavior, whereas self-oriented personal distress is unrelated to altruism (Batson et al. 1988) and may in fact inter-fere with prosocial behavior (Davis&Oathout 1987; Eisenberg et al. 1990, 1993; Estrada 1995).

Empathy and its close cousin sympathy have been cited as central to the human moral affective system for at least three reasons (Eisenberg et al. 2004, 2006). First, empathic reactions to others’ distress often elicit feelings of concern for the distressed other (Feshbach 1975). Second, such empathic concern often prompts behavior aimed at helping the distressed other (Batson 1991, Eisenberg & Miller 1987, Feshbach 1987). Third, feelings of empathy are apt to inhibit aggression and other behaviors that are harmful to others (Feshbach & Feshbach 1969, Miller & Eisenberg 1988).

SUMMARY AND FUTURE DIRECTIONS FOR RESEARCH

This review has considered the implications of moral standards and moral emotion for moral decisions and moral behavior. In this sense, the structure of this review reflects the current state of the field. Little research has examined the relation between moral standards and moral emotional factors, much less their interactive influence in moderating the link between moral standards and people’s moral behavior. Our hope is that this framework will encourage integrated research along such exciting lines. Future directions for research include evaluating the relative importance of cognitive and emotional factors in various domains of morality, as well as the degree to which particular emotional factors are differentially more important in influencing behavior among particular subpopulations (e.g., corporate managers, criminal offenders) and at different points in development.

In addition, this review may help clarify several points of conceptual confusion evident in portions of the literature. For example, in the guilt literature, some theory and associated measures have confounded proneness to guilt with moral standards or other related attitudes and beliefs (e.g., Mosher 1966; see Tangney 1996 for discussion). Although feelings of guilt generally arise from some failure or violation of moral standards, proneness to guilt (an affective disposition) is conceptually distinct from moral standards (a set of beliefs guiding one’s evaluation of behavior). With the advantage of greater conceptual clarity, future researchers can address many questions about the functions and costs of various forms of moral emotion. Such research has potential to pay off substantially, informing educational, judicial, and social policies that foster adaptive moral processes and ultimately moral behavior that benefits all.