Own and other person’s transgressions: Guilt as a function of self-construal

Previous research has shown that self-construal predicts pride responses to success. Independent self-construal increases pride responses to own achievements whereas interdependent self-construal increases pride responses to other person’s achievements. In the current research, I examine if other self-conscious emotions such as guilt are similarly influenced by self-construal. In this study, German students (N = 88, 75 female, 13 male) were primed with an either independent or interdependent self-construal. Guilt was elicited either by one’s own transgression or by the transgression of a close other. Results show that an independent self-construal leads to stronger guilt responses to own transgressions than an interdependent self-construal whereas no such difference was obtained for transgressions of close other’s. In sum, these findings suggest that guilt is influenced by self-construal.

Emotions, such as guilt and shame, are regarded as selfconscious emotions because they hinge on the evaluation of the self in relation to an ideal or standard. Although the emotions of guilt and shame are often used interchangeably (Tangney et al. 1996), guilt is focused on undoing behavior and can, therefore, be seen as approach motivation (Niedenthal et al. 1994;Wicker et al. 1983). Thus, guilt instigates behavior such as repairing the originated damage or harm and signaling that norms are accepted. Shame, on the other hand, represents a form of avoidance behavior and is rather focused on undoing aspects of the self (Niedenthal et al. 1994;Wicker et al. 1983). Thus, shame is related to the desire to disappear or hide. According to Tangney et al. (1995), shame but not guilt is associated with depression.
Since guilt arises in interpersonal interactions, it can be conceived of as an inherently social emotion. According to Baumeister et al. (1994, p. 245), guilt can be conceived of as an unpleasant emotional state that is "associated with possible objections to his or her actions, inaction, circumstances, or intentions." Thus, the discrepancy between own behavior and social norms plays a central role in the elicitation of guilt. However, the social nature of guilt is even more articulated if we consider that guilt can also arise from behavior that is wrong in the eyes of other's. From this point of view, it seems obvious that guilt hinges at least partly on the ability to take the perspective of other's (Leith and Baumeister 1998).
However, there is a second aspect of why guilt can be conceived of as an inherently social emotion. Be it in the eyes of someone else or in one's own eyes, guilt can be conceived of as emotion that is elicited by misdeeds of the self. Due to the social nature of guilt, however, guilt can also arise from the misdeeds of someone you feel close to. Lickel et al. (2005) observed that vicarious guilt depends on the perceived interdependence with the wrongdoer. Interdependence refers to the relationship of those with whom you have the "opportunity for shared communication and influence over one another" (Lickel et al. 2005, p. 148). Thus, that someone can feel guilty for the behavior of someone else hinges, among other things, on the perceived relationship to the wrongdoer. If a close other does something wrong, she or he might perceive oneself responsible for repairing the damage. Such vicarious emotions are not restricted to guilt but are well known for other selfconscious emotions such as pride. Research has shown that one can feel pride because of the association with successful other's (Cialdini et al. 1976). Thus, in this research, it is not the real mutual exchange but someone else is subjectively construed as being part of the self that is sufficient to trigger pride. This research shows that such a construal of associations to other's is somewhat malleable. According to Markus and Kitayama (1991), culture shapes how people think of themselves and how they are related to other's.

Cultural Influences
Cultural differences such as those between individualist or collectivistic cultures are related to differences in selfconstrual (Markus and Kitayama 1991). According to Markus and Kitayama (1991), participants from western (i.e. individualist) cultures judge themselves to be more dissimilar to other's than other's to themselves. Indian participants, however, judge themselves to be more similar to other's than the other to the self. This indicates that the representation of the self as separated from other's is more elaborated in western cultures. Markus and Kitayama (1991) point out that such different self-descriptions have their origin in culturally shaped self-construals. An independent self is prevailing in individualistic cultures, which is characterized by the self as being separate from the social context. In collectivistic cultures, however, an interdependent self is prevailing, which is characterized by being connected with social context. Although self-construals can be conceived of as dispositions that are shaped by culture, they can nevertheless vary in accessibility depending on the social role. For example, in many cultures, the female role is more related to an interdependent self-construal whereas the male role is more related to the independent self-construal (Cross and Madson 1997). Accordingly, self-construal can be made temporally accessible and influence information processing despite being nondominant within a given culture (Gardner et al. 1999;Trafimow et al. 1991). In a similar vein, it is a central assumption of the "culture as situated cognition model" (Oyserman 2011) that situational cues can change the accessibility of culturally relevant cognitive concepts.
Self-construal influences not just cognition and behavior, but also emotional experience. According to Markus and Kitayama (1991), the same emotion is elicited by different conditions depending on the self-construal. Based on this assumption, individual achievements should elicit stronger pride responses in members of an individualistic culture whereas achievements of close other's should elicit stronger pride responses in members of collectivistic cultures. Support for this prediction comes from a study by Stipek et al. (1989): In their study, Chinese and American students were asked to describe situations that elicit pride. They observed that American students attributed pride more often to their own effort than Chinese students. On the other hand, Chinese participants responded with more pride to high achievements of close other's. However, it is not clear if these differences in emotional responses are causally due to differences in selfconstrual. To show such a causal influence, it is necessary to temporarily manipulate the accessibility of the relative selfconstrual and assess the effects of this manipulation on emotion.
To determine if the cultural differences in pride responses indeed hinge on differences in the accessible self-construal, Neumann et al. (2009) manipulated the self-construal of German college students temporarily by a priming procedure (Trafimow et al. 1991). After independent self-construal was induced, more pride was experienced in response to individual achievements whereas in response to achievements of close other's, more pride was experienced after interdependent selfconstrual was induced. Taken together, these findings suggest that different situations, such as one's own or other person's achievements, interact with the prevailing self-construal to elicit the self-conscious emotion of pride. In the current research I try to apply these assumptions to guilt as another selfconscious emotion. Derived from Markus and Kitayama's (1991) proposal that the same emotion is elicited by different conditions depending on self-construal, I suggest that a fit between self-construal and situation influences guilt responses. More specifically, I predict that independent selfconstrual triggers more guilt in response to own transgressions whereas interdependent self-construal leads to more guilt in response to transgressions of close other's. For this purpose, vignettes were used that report either own transgressions or transgressions of close other's. Two versions of the vignettes were implemented so that those vignettes that are used as own transgressions in the first version are used as transgressions of close other's in the second version. Before responding to these vignettes, either the independent or the interdependent selfconstrual was made accessible by a priming procedure.

Method
Participants Eighty-eight students (75 female, 13 male) from the University of Trier participated (36 female, 8 male in the interdependent priming condition and 39 female, 5 male in the independent priming condition). Participants' mean age was M = 23.58 years (SD = 6.81). Psychology students received course credit for participation. Specification of the sampling goal was based on a study by Neumann et al. (2009), Experiment 2). The required sample to detect an interaction effect of η 2 p = .091 with an alpha value of .05 and a power of 1 ß = .80 is N = 16 (G*Power; Faul et al. 2009). However, since it is unclear if guilt shows a similar effect as pride and we expected experimental dropout, we oversampled and collected data from N = 88. The study was carried out in accordance with the provisions of the World Medical Association's Declaration of Helsinki.

Procedure
The experiment was conducted in our lab. Participants were escorted to cubicles where they found further instruction. They first signed an informal consent sheet. Next, they responded to questions referring to their age, sex, field of study, and nationality. The priming task was then administered in paper-pencil format. Participants were informed that they were participating in two different studies. Consistent with Trafimow et al. (1991), participants in the independent self-construal condition were instructed to "please think about several differences you have with your family and friends. Please write these differences down on the sheet of paper in front of you." Participants in the interdependent self-construal condition were instructed to "please think about several similarities you have with your family and friends. Please write these similarities down on the sheet of paper in front of you." They were then informed that the next task was part of the second study. In this task, they were presented 14 different scenarios that were constructed to elicit guilt (see Appendix). The leading principle in constructing the scenarios was that guilt is elicited by actions that violate norms (Niedenthal et al. 1994;Wicker et al. 1983). The presentation language was German. Two versions were developed: In the first version that was presented to one half of the participants, the first seven scenarios focus on own transgressions and the second half on transgressions of relevant other's. In the second version presented to the other half of the participants, the first seven scenarios focus on transgressions of close other's and the other half focus on own transgressions. Thus, since the sequence of the scenarios was the same for all participants, the identical scenario was presented from the own transgression perspective for one half of the participants and from the perspective of a relevant other person for the other half of the participants. This counterbalancing strategy was used to exclude that the differences between own and other person's transgressions is due to the selection of the items. Participants were asked to imagine the scenarios as vividly as possible. The subjective intensity of guilt was assessed to each scenario with a Likert scale ranging from 0 (not present) to 7 (very strongly present). Participants were fully debriefed following completion of the scenario task. We expected stronger guilt responses for own transgressions if the independent rather than the interdependent self-construal was primed and stronger guilt responses for transgressions of relevant other's if the interdependent rather than the independent selfconstrual was primed. The whole experiment lasted approximately 20 min.

Results
The data of all 88 participants were included in the analysis. Participants' guilt responses to the scenarios were averaged across own transgressions (Cronbach's α = .92) and transgressions of relevant other's (Cronbach's α = .83). The intensity of the guilt responses were then subjected to a Priming (independent vs. interdependent self-construal) x Transgression (own vs. other) analysis of variance (ANOVA), with Priming as a between-subjects factor and Transgression as a within-subjects factor. Reported confidence intervals (95% CI) refer to the differences between two means.
A  Table 1.
To find out if the used scenarios exert an impact on the results, a Scenario x Priming analysis was conducted. Results revealed no main effect for Scenario, F(1, 87) = 2.13, p = .148, no interaction of Scenario x Priming, F(1, 87) = 1.20, p = .277, and no other main effect or interaction (F < 1).

Discussion
The current research was guided by the view that a fit between self-construal and situation shapes self-conscious emotions. More specifically, stronger guilt responses for own The intensity of the guilt response ranged from 0 (not present) to 7 (very strong) transgressions were expected if the independent rather than the interdependent self-construal was made accessible, and stronger guilt responses for transgressions of close other's were expected if the interdependent rather than the independent self-construal was primed. In line with this prediction, evidence for the influence of the activated self-construal on guilt responses was obtained. Stronger guilt responses were observed to own transgressions if the independent rather than the interdependent self was accessible whereas no such difference was obtained for guilt responses to transgressions of close other's. Apparently, and similar to other self-conscious emotions such as pride, self-construal exerts different influences on actions attributed to the self and actions attributed to close other's. Thus, our research supports the proposal that the same emotion is elicited by different conditions depending on self-construal (Markus and Kitayama 1991). Own transgressions and transgressions of close other's can have different consequences depending on the currently accessible selfconstrual. Nevertheless, there are several limitations of the current research. An important limitation is that the study was not replicated and not preregistered. Thus, it would be interesting to replicate the study and also include a sample from a collectivistic culture. A second limitation is that the contrast between independent and interdependent priming was reliable within own transgressions but not within transgressions of close other's. One reason for this discrepancy might refer to differences in the scenarios used to assess either own or close other's transgressions. However, since the transgressions named in the items were counterbalanced between own and close other's transgressions and an inclusion of this counterbalancing factor in the analysis does not yield an effect, such an explanation cannot be reconciled with the findings. Nevertheless, beyond the specific transgressions employed in the scenarios, one apparent difference refers to the varying social closeness of the persons named in the scenarios of close other's transgressions. Accordingly, compared to the scenarios of own transgressions, items are more heterogeneous and therefore the reliability is lower. It is thus conceivable that the priming procedure was not successful in this condition because of the varying social closeness and the other dimensions that introduce variability. 1 Another reason for the lack of priming effect within close other's transgressions might be that the study was conducted in an individualistic culture with a prevailing independent self-construal. The prevailing independent self-construal leads to stronger and more reliable guilt responses to own transgressions rather than to close other's transgressions. It is therefore conceivable that strong guilt responses to transgressions of close other's are required to demonstrate an effect of the temporal change in accessibility of self-construal. In other words, responses to close other's' transgressions were too weak to be influenced by the priming procedure.
Another limitation of the study is due to the lack of a control group without a manipulation of the self-construal: Therefore, in the current design, it is not possible to disentangle if the independent self increases the guilt response or if the interdependent self decreases the guilt response.
In our study, 85% of the participants are female. It is well known that there are important gender differences in self-construal: Men in individualistic countries are usually thought to develop a more independent self-construal than woman (Cross and Madson 1997). It is thus conceivable that a larger proportion of men in the study might have increased the difference between independent and interdependent priming in the own transgression condition whereas a larger proportion of woman might have increased the difference in the close other condition.
One might object that guilt due to own transgressions and vicarious guilt to transgressions of close other's is not the same emotion. For example, the experience and the action readiness (e.g., readiness to repair) might be different. Interestingly, a recent study has shown that own transgressions and transgressions of close other's recruit the same patterns in the anterior middle cingulate cortex (Li et al. 2020). Of course, one cannot conclude that both types of guilt response are experienced similarly. Thus, it remains open as to whether both forms of guilt are the same.
Importantly, however, self-conscious emotions are not only influenced by culture but the expression of these emotions fulfills important functions within the culture. For example, the expression of guilt and pride over one's own action amplifies the construal of the self as independent and unique (Mesquita 2001). On the other hand, the expression of guilt and pride over actions of close other's emphasizes the construal of the self as embedded in a social structure. In this respect, self-construal and self-conscious emotions are closely interrelated.
Funding Open Access funding enabled and organized by Projekt DEAL.
Data Availability The datasets generated and analyzed during the current study are available from the corresponding author on request.

Compliance with Ethical Standards
Conflict of Interest The author declares that he has no conflict of interest. The author has received no financial support for the research, authorship, and publication of this article.

Ethical Approval
The study was carried out in accordance with the provisions of the World Medical Association's Declaration of Helsinki. According to the guidelines of the German Science Foundation (DFG), approval by an ethical review board is required only if the study imposes high physical or mental risk (https://www.dfg.de/foerderung/faq/geistes_ sozialwissenschaften/ index.html). Since it is unlikely that responding to the items employed in the current study imposes high mental risk, no ethical approval was obtained.
Informed Consent was obtained from all individual participants included in the study.