Abstract
Intergroup conflict is a costly and persistent aspect of social life, and one that often carries great moral significance for those who participate in it. Ostensibly moral behaviors can provide a path to social status in groups, as when self-sacrificing in-group members gain respect and prestige relative to their peers. This article bridges these two ideas to examine the perceived morality and status worthiness of intergroup revenge and forgiveness, and the moral accounts used to justify them. Using an original survey experiment conducted on a national probability sample in the United States, we examine everyday intergroup conflicts across national, sports, and political identities. We find forgiveness is perceived as more moral, and in turn more status worthy, than revenge. Justifications for moral judgments typically drew on accounts of harm/care, reciprocity, and avoiding chaos and disorder. This contributes to research on conflict, group dynamics, status, and morality.
Status is a fundamental dimension of inequality within and between groups. Unlike material forms of inequality, it can emerge only from interactions between others and self, operates through shared beliefs, and cannot be directly seized from others (Ridgeway and Markus 2022, this issue). Thus individuals are motivated to behave in ways that garner respect and deference from others (Bai 2017; Bai, Ho, and Yan 2020; Flynn et al. 2006; Hardy and Van Vugt 2006; Milinski, Semmann, and Krambeck 2002; Willer 2009). Existing status research focuses intragroup relations, including the traits, behaviors, and resources that cause some people to have more status than others (Correll and Ridgeway 2006).
Intergroup conflict is an important structural condition shaping these intragroup status processes (Benard et al. 2021; Benard and Doan 2020). Intergroup conflict provides opportunities for individuals to demonstrate valued qualities, such as group commitment, and in doing so gain status in groups. For example, Ukrainian President Volodymyr Zelensky was widely praised for refusing U.S. offers to evacuate prior to the Russian invasion, instead staying to participate in the defense of Kyiv (Braithwaite 2022; Harris, Francis, and Dixon 2022). In this article, we examine whether perceptions of morality shed light on how conflict behaviors and status processes are related. We aim to understand the relationship between perceived morality, status, and two contrasting responses to conflict: intergroup revenge and forgiveness. To do so, we draw on two findings from recent research.
First, individuals view intergroup conflicts in deeply moral terms, especially when valued in-groups are implicated in these conflicts (Böhm, Thielmann, and Hilbig 2018; Graham, Haidt, and Nosek 2009; Halevy et al. 2015; Rai and Fiske 2011). Conflict shapes the moral evaluations of behavior and which behaviors are considered moral varies across social contexts (Rai and Fiske 2011). In some cases, aggression is considered heroic when directed at an out-group member but despicable otherwise (Halevy and Cohen 2019). Indeed, in honor cultures, failing to seek revenge for an intergroup affront is shameful (Beckerman et al. 2009; Black-Michaud 1975; Gould 2003). The perceived morality of forgiveness also varies situationally. Forgiveness is viewed as virtuous in some contexts, such as nonviolent civil rights movements (McAdam and Tarrow 2000), but is morally unacceptable in others (Exline et al. 2003). Our work finds that in everyday contexts, intergroup revenge is viewed as less status worthy than forgiveness (Benard et al. 2021).
Second, individuals who demonstrate ostensibly moral behaviors such as altruism, generosity, and self-sacrifice gain social status (Bai 2017; Bai, Ho, and Yan 2020; Flynn et al. 2006; Hardy and Van Vugt 2006; Milinski, Semmann, and Krambeck 2002; Willer 2009). Status—the relationships of respect, prestige, and deference within groups—plays a fundamental role in structuring group life and motivating behavior (Anderson and Kilduff 2009; Bales et al. 1951; Berger, Cohen, and Zelditch 1972; Correll and Ridgeway 2006; Strodtbeck, James, and Hawkins 1957).
Taken together, the findings that conflict carries moral weight for its participants and that moral behaviors are considered status worthy suggest that conflict provides opportunities to gain status by demonstrating moral behavior. Building on this work and drawing on the status theory of collective action (Willer 2009), costly signaling theories (Spence 1973), and the moral virtue theory of status attainment (Bai 2017), we explore the status value of revenge, forgiveness, and morality in intergroup conflict. We evaluate the perceived morality of revenge and forgiveness, whether moral behaviors are perceived to be status worthy, and whether morality provides a pathway from intergroup revenge or forgiveness to intragroup status. Drawing on moral foundations theory (Graham, Haidt, and Nosek 2009), we examine the moral accounts used to justify evaluations of revenge and forgiveness as morally right or wrong, and how these accounts are related to perceived status. We evaluate these relationships using original, nationally representative experimental survey data with quantitative and qualitative measures.
This work makes several contributions. We use the concept of morality to bridge existing lines of research on intergroup conflict and the status worthiness of cooperative and competitive behaviors. Intergroup conflict is a persistent and costly aspect of social life as well as a key problem for understanding human groups (Cook 2000; Fiske 2002; Halevy and Cohen 2019). This article elucidates how intergroup conflict and intragroup relations shape one another, a question that has fascinated social scientists for more than a century (Barclay and Benard 2013; Benard and Doan 2011; Coser 1956; Gouldner 1954; Halevy and Cohen 2019; Pickett, Bonner, and Coleman 2002; Sherif 1966; Simmel [1908] 1955; Sumner and Keller 1906). Yet this work rarely focuses on the status dynamics of intergroup relations.
By focusing on status, this article has implications for understanding how conflict shapes status inequality. Within groups, conflict may spur the development of unequal status hierarchies, or high-ranking group members may exploit real or artificial threats to solidify their position (Barclay and Benard 2013; Pickering and Kisangani 2005). Between groups, cycles of revenge may increase intergroup oppression whereas norms of forgiveness may limit these cycles (Corey and Joireman 2004). Finally, this study contributes to understanding the role of status in cycles of conflict. Some groups become locked in recurring conflict; others establish peaceful, productive relations (Fearon and Laitin 1996). Understanding the moral accounts people use to justify forgiveness or vengeance has practical implications for encouraging peaceful intergroup relations. Once identified, these moral accounts could be used to frame conflict prevention messages in ways that are convincing and culturally appropriate.
THEORY
Status is a key factor shaping intragroup relationships. Definitions of the term vary across literatures (Faris 2012; Henrich and Gil-White 2001; Martin 2009). We draw on expectation states theory to use the term to refer to one’s place in a group hierarchy of respect, prestige, and deference (Berger, Cohen, and Zelditch 1972). Many groups, including work teams, juries, labor unions, and youth gangs, organize themselves along status lines; some group members are accorded greater respect, prestige, and influence than others (Bales et al. 1951; Gould 2003; Sherif 1966; Strodtbeck, James, and Hawkins 1957; Van Vugt, Hogan, and Kaiser 2008). Although symbolic, status is a widely desired social resource with material consequences. It incentivizes contributions to group goals and compliance with group norms while generating inequality and sometimes status competition within groups (Anderson et al. 2012; Anderson and Kilduff 2009; Correll and Ridgeway 2006; Willer 2009).
Paths to Status in Groups
Individuals can follow several paths to status in groups. Historically, most research focuses on a competence-based path (Bai 2017; Berger, Cohen, and Zelditch 1972; Cheng, Tracy, and Henrich 2010).1 This approach, elaborated by expectation states theory, argues that status hierarchies emerge when groups aim to collectively address a task or problem (Berger, Cohen, and Zelditch 1972; Correll and Ridgeway 2006). To do so, they evaluate which of their members are expected to make the greatest contribution to solving the problem at hand; these members receive more respect, attention, and deference. The expectation is that deferring to competent group members will improve group performance.
Acknowledging the importance of the competence-based pathway, we focus on prosociality as route to status. Diverse strands of research have identified an alternative set of pathways to status in groups that broadly hinge on gaining status through behaviors that demonstrate valued characteristics other than competence. These characteristics include altruism, sacrifice for other group members, generosity, and trustworthiness. This perspective is derived from expectation states (Ridgeway 1982; Willer 2009) and costly signaling theories (Hardy and Van Vugt 2006; Milinski, Semmann, and Krambeck 2002; Nelissen 2008). It argues that individuals who sacrifice for the group’s well-being, who are more generous with other group members, or who otherwise incur costs for the group’s benefit are presumed to possess greater intrinsic motivation to help the group, trustworthiness, or other group-beneficial qualities (Barclay 2004; Willer 2009).
Two interpretations of why these behaviors translate into greater status and leadership opportunities prevail. The status theory of collective action (Willer 2009) and other signaling-based approaches (Barclay 2004; Hardy and Van Vugt 2006; Milinski, Semmann, and Krambeck 2002; Nelissen 2008) suggest that conferring status to those with group-beneficial qualities pays off because they use status and influence to benefit the group. The moral virtue theory of status attainment argues that actors who embody the moral values of their group are perceived as virtuous, or excelling in moral domains, and in turn gain status (Bai 2017; Bai, Ho, and Yan 2020).
To the extent that virtuous behaviors yield group benefits, the moral virtue theory and signaling-based theories overlap. For example, Feng Bai (2017, 208) argues that “to be morally praiseworthy, virtue, inevitably, involves voluntary self-sacrifice for the good of others, beyond conformity to moral norms” (emphasis in the original). However, the moral virtue theory also argues that enacting moral values is not necessarily instrumentally beneficial to the group, such as when a moral emphasis on order or obedience impairs performance on group task requiring creativity (Bai 2017, 210), or when acts of altruism leave one too exhausted to contribute to group goals (Bai, Ho, and Yan. 2020, 504). Instead, the mechanism linking virtue to status is a “warm glow” of admiration that enhances one’s social standing.
This raises the question of whether behaviors such as altruism, generosity, and self-sacrifice function as costly signals of group motivation, signs of virtuous character, or both. Our data include separate measures of group motivation and perceived morality, allowing us to assess whether the evidence is consistent with one or both pathways.
Research on these pathways to status focuses on behavior that is unambiguously helpful to the group, such as contributing money or effort to group goals (Barclay 2006; Hardy and Van Vugt 2006; Willer 2009), or organizational citizenship behaviors (Bai, Ho, and Yan 2020). Assigning status for acts of intergroup revenge or forgiveness is more complicated, because each behavior may be interpreted as a positive or negative contribution to group welfare. In the moment, it can be ambiguous whether revenge or forgiveness will have beneficial consequences. Revenge may signal the group’s toughness and cohesion, deterring future threats from out-groups (Gould 2003; Schelling 1980), or it may expose the group to counter-revenge or other sanctions (Nikiforakis and Engelmann 2011). Forgiveness may reduce costly intergroup conflict and provide opportunities for mutually beneficial relationships (McCullough, Kurzban, and Tabak 2013), or may create the perception that the in-group is fainthearted, or that the forgiver sympathizes with the out-group (Sherif 1966). This ambiguity may account for the cultural variation in the value placed on revenge, as well as groups’ internal conflict about pursuing revenge or forgiveness (Exline et al. 2003).
When individuals make these assessments before the long-term consequences of vengeful or forgiving behavior are known, how do they assign status for these behaviors? Our recent work examines this question (Benard et al. 2021). We use a vignette study depicting an in-group member who is confronted by an out-group member. The out-group member insults the in-group, and the in-group member responds in either a forgiving (encouraging peaceful relations between the two groups) or vengeful manner (verbal retaliation and a shove). Across three identities (national identity, sports fandom, and political party affiliation) we find that the forgiving in-group member is viewed as more group motivated and more status worthy, compared to the vengeful in-group member. Consistent with signaling-based accounts, perceived group motivation partially mediates the effect of forgiveness on status, suggesting that forgiveness leads to status by signaling group motivation.
Extending these findings, we ask here whether forgiveness and revenge are perceived as moral, whether perceived morality increases perceived status, and whether it does so separately from group motivation. We investigate how people justify the moral judgments they make about revenge and forgiveness. We then examine whether these judgments are correlated with the status worthiness of morality. Because group solidarity is amplified through “enthusiastic participation in group rituals” that feel righteous and morally good (Collins 1990), a better understanding of the moral foundations used to justify revenge and forgiveness elucidates an underexplored dimension of group cohesion, hierarchy, and social status. These questions follow.
Q1: Is forgiveness or revenge more often viewed as morally right?
We investigate whether revenge or forgiveness is more often viewed as morally right. Drawing on data from a vignette study of revenge and forgiveness (Benard et al. 2021), we examine an open-ended item in which respondents were asked to make moral assessments of vengeful or forgiving behavior. According to the moral virtue theory of status attainment, if one behavior is viewed as more morally right, it should be viewed as more status worthy (Bai 2017). We made no predictions as to whether revenge or forgiveness will more often be viewed as morally right, given that each behavior could be interpreted as morally right or wrong.
Q2: What moral accounts do people draw on to justify moral judgments about revenge and forgiveness?
After assessing the perceived morality of revenge or forgiveness, we evaluate why people perceive revenge or forgiveness as morally right or wrong. To do so, we asked survey respondents to explain their moral judgments of the forgiving or vengeful behavior in the survey vignette. To analyze these accounts in a theoretically informed way, we drew on moral foundations theory (Graham, Haidt, and Nosek 2009). This theory argues that humans have a set of evolved but culturally and socially modifiable moral foundations. Although the exact number and nature of these foundations is debated, much work focuses on five: harm-care, fairness-reciprocity, in-group-loyalty, authority-respect, and purity-sanctity. As the names suggest, the harm-care foundation focuses on caring for others and avoiding causing harm, and the fairness-reciprocity foundation on issues of justice and cooperation. These are the “individualizing” foundations (Graham, Haidt, and Nosek 2009). The latter three are the “binding” foundations, for their presumed role in maintaining social solidarity through group loyalty, subservience to group hierarchies, and respecting cultural boundaries.
For our purposes, we are less concerned with whether the theory’s underlying assumptions are correct (for example, whether moral foundations are evolved cultural universals or primarily intuitive versus reasoned; see Smith et al. 2017). Instead, our interest lies in the fact that these foundations are widely valued, to differing degrees, in a broad range of cultural settings, and so make a logical starting point for understanding the perceived morality of revenge and forgiveness. In our coding, we also inductively capture respondents’ understanding of other moral values beyond these five foundations.
Q3: Does perceived morality mediate the relationship between forgiveness and status?
We next assess whether demonstrating morality through vengeful or forgiving behaviors provides a path to status. If one of these behaviors is viewed as more morally right, the moral virtue theory of status attainment predicts it should also be viewed as more status worthy. Our work finds that individuals gain status from forgiving behavior, in part because it signals group motivation (Benard et al. 2021). Given the broad theoretical similarity between signaling and morality-based approaches—in that each suggests that certain behaviors are taken as a sign of desirable underlying qualities, and therefore status worthy—we assess whether the evidence is more consistent with morality and group motivation as separate or overlapping paths.
Q4: Are specific moral accounts differentially related to status?
As an exploratory step, we examined whether some moral accounts elicit more status than others. For example, are behaviors viewed as morally right because they are caring perceived as more or less status worthy than those viewed as morally right because they are reciprocal? Existing theory suggests that whether a behavior is viewed as morally right shapes its perceived status worthiness (Bai 2017), but we know of no work on the status worthiness of specific moral foundations. We examine this question in two ways. We assess whether mean levels of perceived status differ by specific moral accounts. We also examine whether specific moral accounts play a role in mediating the effect of forgiveness-revenge on status.
METHODS
To answer our research questions, we conducted a nationally representative online survey experiment using NORC’s AmeriSpeak panel, as part of a larger study we conducted on the status value of group motivated revenge and forgiveness (Benard et al. 2021). The AmeriSpeak panel uses a two-stage probability sample design based on the NORC National Sample Frame (Dennis 2017). Panelists are recruited into the sample and agree to complete two to three short surveys per month. A sample of 2,116 respondents were recruited for the study. Of these, thirty (1.4 percent) skipped all survey items. Of the 2,086 remaining, 1,928 (92.4 percent) responded to the morality question. An additional fifteen skipped all of the items in the status or group motivation scales, leaving 1,913 respondents for analysis (90 percent of the total sample).2
Procedure
After providing informed consent, respondents were asked a series of questions designed to gauge important in-groups. All respondents were U.S. citizens, providing one (national) in-group identity. Respondents were also asked about political party identification, whether they had a favorite sports team and, if so, to name the team. Following these preliminary questions, we elicited perceived out-groups from respondents with political party and sports team identifications by asking which political party’s interests most conflict with their own party’s interests, and which sports team is the biggest rival of their favorite team. We treat these identities as three conceptual replications based on national identity, sports fandom, and political party affiliation, to evaluate the robustness of the findings and map variation across key social identities such as vocations and avocations (sports fandom), political affiliation, ethnicity and religion broadly conceived (national identity) (Deaux et al. 1995).3 These identities are widely held, making them practical for use in a study of the general population. Pilot testing indicated that individuals holding these identities can easily identify rival out-groups.
Using these responses, respondents were randomly assigned to an identity scenario based on their in-group memberships. Respondents who identified with all three groups were randomly assigned to one of the three. Respondents who identified with either a political party or a sports team were randomly assigned to scenarios corresponding to one of the two identities they had. The survey was programmed to keep the sizes of the scenarios relatively equal (between 625 and 658 respondents per scenario).
After making assignments to a scenario, we measured respondents’ identification with the in-group along with the perceived level of conflict with the out-group. In the broader study, respondents were then randomly assigned to one of the four between-subjects experimental conditions (vignette character’s behavior: forgiving-vengeful × vignette character’s stated motivation: group-individual benefit). Here we focus on the behavior factor (whether the vignette character behaves in a forgiving or vengeful manner). We find few differences by group motivation in the outcomes of interest (for details, see Benard et al. 2021); thus we present results collapsed across the motivation conditions to save space. Respondents read a vignette that corresponded to the condition and scenario they were assigned, specific political party or sports in-groups and out-groups derived from their responses earlier in the survey.4
The vignette asked respondents to imagine witnessing a chance encounter between two strangers in a public setting. This setting was a local tourist attraction in the national identity scenario, a protest in the political affiliation scenario, and the parking lot of a game between the respondent’s favorite team and a rival team in the sports fandom scenario.5 In each scenario, the vignette describes an out-group member insult an in-group member, using a group-based insult. For example, in the sports scenario, the out-group members says, “Watch where you put your stuff, jerk! All you [In-group] fans are such idiots!” Depending on the experimental condition, the in-group member responds in either a forgiving or vengeful manner, and then turns to a friend and provides either an individually motivated or group-motivated explanation for their behavior. Following the vignette, respondents evaluated the perceived social status of the in-group vignette character, completed manipulation check questions, and answered additional items, including the morality measure and items used in the moderation analysis.
Key Variables
Forgiveness. Our primary independent variable is a condition indicator for whether the in-group vignette character behaves in a forgiving or vengeful manner (1 = forgiving, 0 = vengeful). In the forgiveness condition, the in-group member tells the out-group member that they do not need to argue, and recommends that each person return to their original activities (such as enjoying the game or expressing themselves at the protest). This is consistent with scholarship viewing forgiveness as a proactive effort to improve intergroup relations, not solely the absence of revenge (McCullough 2001; McCullough, Kurzban, and Tabak 2013). In the revenge condition, the in-group member shoves the out-group member and issues a verbal warning. A manipulation check using two 9-point, Likert-type items finds that respondents viewed the focal character as significantly more forgiving in the forgiving vignette relative to the vengeful vignette (M = 6.54 vs. M = 4.11, t = –27.65, p < .0005, two-tailed).6
Perceived Group Motivation. We used two 9-point, Likert-type items to measure perceived group motivation. We asked respondents to assess the extent to which the vignette character was mostly concerned with helping himself or [other in-group members] and acting on his behalf or on behalf of [other in-group members], averaged to form a scale (α = .84).
Moral Evaluation. Morality is measured qualitatively with a two-part question and coded into two separate measures.7 The first question asks respondents, “Thinking about how the [in-group member], handled this situation, would you say that their response was morally right or wrong?” We coded responses into four mutually exclusive categories: “yes,” if the respondent believed the behavior was morally right (“Yes, I think it was morally right”), “no,” if the respondent believed the behavior was morally wrong (“I do not believe that morally it was appropriate”), “maybe,” which included statements that were neutral, ambivalent, or mixed (“It was neither right or wrong; it had a little of both”), and “other,” which included statements that did not assess the action’s morality (“I don’t think morality came into play in this situation”).8
Moral Accounts. Our second measure of morality captures moral values respondents drew on to explain their moral judgment. After their moral evaluation, we ask respondents, “What made it right or wrong?” As noted, our initial set of coding categories drew on five moral foundations specified by moral foundations theory—harm-care, fairness-reciprocity, in-group–loyalty, authority-respect, and purity-sanctity (Graham, Haidt, and Nosek 2009)—but we also used open coding to identify moral accounts that did not fit into these categories.
Following the open coding, the final codebook included a total of nine codes for moral justifications: harm-care, fairness, reciprocity, chaos-disorder, in-group–loyalty, respect-authority, purity-sanctity, self-control, and out-group–related justifications. The harm-care code was used for evaluations of morality that drew on people being harmed or cared for, either physically or emotionally (“I think it was morally wrong; I don’t think favorably of being physically aggressive with others or harming them emotionally”). We divided fairness and reciprocity into two codes, due to a tendency for respondents in our pilot data to distinguish between these moral justifications. Respondents tended to use fairness to denote unjust treatment, such as being the object of stereotyping (for example, one respondent in the sports identity scenario felt it was wrong for a vignette character to assume all fans of a team have similar characteristics: “Also he stereotyped the other fan as all of them being the same which is not right either”), and used reciprocity tended to denote a disproportionate response (“It was not morally right in that they made a big fuss over nothing”).
We included a code for chaos-disorder, which is sometimes grouped with authority in moral foundations theory. Our respondents used this type of reasoning to denote morality based on whether the in-group member’s actions caused or could have caused chaos or disorder: “I think this person is probably morally right, because they . . . avoided a larger conflict that might have unintended consequences (stabbing, shooting, expulsion from parking lot”).
Codes for in-group loyalty were used for justifications around the idea of betraying or not betraying one’s in-group (“He did the right thing because he spoke out in defense of someone else . . . [and] because he was part of a group trying to persuade others and he made his side look more reasonable”). The respect-authority codes were used for moral accounts drawing on the respect for authorities, other people, or social traditions. We coded for purity-sanctity for justifications drawing on notions of disgust, purity, and standards of decency (“They did not fire back with the same type of vulgar language”).
We developed two novel codes through open coding. One, which we called self-control, captured evaluations of morality that take into account whether or not someone demonstrated self-control or restraint (“I believe it was morally right. He took the high road. He could’ve insulted the foreign tourist in retaliation but chose not to”). The second, which we called moral-out, captured references to the out-group’s morality (or lack of morality).
Respondents gave different moral accounts for their assessments of the vignette character’s behavior, depending on whether the behavior was vengeful or forgiving and whether it was perceived as morally right or wrong. Categories do not sum to 100 percent because responses could be coded into multiple categories, or none of the categories.
Status. Respondents rated how the in-group member “is probably viewed in groups that he belongs to” on five 9-point, bipolar scales: respected, honorable, influential, a leader, and prominent (Ridgeway and Erickson 2000; Willer 2009). These items were average to create a single status measure (α = .88).
Analytic Strategy
To succinctly present the results, we pooled data across the three identity scenarios for the analyses presented in the main text. Perceived morality was largely similar when examining the three identity scenarios separately.9 Our analyses proceed in four steps. First, we evaluate whether forgiveness is viewed as more moral and status worthy than revenge using tests of proportions across the conditions. Second, we present our qualitative results outlining accounts respondents give in justifying their moral evaluations and test for differences in the prevalence of these moral accounts across conditions. Third, we conduct a mediation analysis to see whether moral evaluation and moral accounts explain status differences between forgiving and vengeful group members. Best practices are debated for both control variables and survey weights when it comes to survey experiments (Miratrix et al. 2018). In keeping with traditional presentations of experimental results, we present unweighted models without controls.10 In supplementary analyses, we evaluated whether the perceived morality of revenge or forgiveness varied by individual-level factors such as strength of group identity or demographics.11 While there is some individual-level variation, for example, those strongly identified with the group and men are more likely to view revenge as morally right, the overall patterns were similar to those in the main text.
RESULTS
In this section, we examine four research questions: whether revenge and forgiveness are perceived as morally right or wrong, what justifications are offered for these judgments, whether morality provides a path to status in groups, and whether specific moral accounts vary in their perceived status worthiness.
Q1: Is forgiveness perceived as more moral?
Table 1 includes mean status ratings and comparisons across conditions of the proportion of respondents who viewed the focal vignette character’s behavior as morally right, morally wrong, morally ambiguous, or something else altogether. Our prior work using these data shows that forgiveness is more status worthy than revenge (Benard et al. 2021). This result is replicated here (M = 6.46 in the forgiveness condition vs. M = 5.05 in the revenge condition, p < .001). We also ask whether forgiveness is viewed as more moral than revenge. As the table indicates, 76 percent of respondents viewed the forgiving vignette character as morally right compared to 20 percent who viewed the vengeful vignette character as morally right (Δ = 0.56, p < .001, two-tailed). Inversely, respondents are also less likely to view forgiveness as morally wrong (7 percent) versus revenge (55 percent; Δ = 0.48, p < .001). Also shown in table 1, a small proportion of respondents gave ambivalent evaluations, about 2.5 percent in each condition, indicating that Americans tend to not equivocate or have trouble determining the morality of forgiveness compared to revenge. More respondents gave nonevaluative responses than ambivalent ones; respondents were also significantly more likely to give nonevaluative responses in the revenge condition than the forgiveness condition (Pr(“Other”|Revenge) = 0.22 vs. Pr(“Other”| Forgiveness) = 0.14, p < .001).
Means by Condition
Collectively, the results indicate that Americans are more likely to view forgiveness in this setting as status worthy, more likely to view it as morally right, less likely to view it as morally wrong, and more likely to give it an evaluation than they are revenge. The proportion of respondents who viewed the focal character’s behavior as morally wrong is not merely a reflection of those who did not view it as morally right, suggesting that moral evaluation is not black and white. Indeed, a sizable minority of respondents do not view the given scenario as at all relating to morality. To better understand the social construction of morality, we examine different moral accounts respondents give to justify their moral evaluation.
Q2: What moral accounts do people draw on to justify moral judgments about revenge and forgiveness?
Although all nine values were given to justify respondents’ moral evaluations, one received a clear plurality of support for three of the four revenge-forgiveness x morally right-wrong categories (see figure 1). Among respondents who viewed revenge as morally right, a plurality (37 percent, n = 73/198) justified this evaluation based on reciprocity. Those who viewed revenge as morally wrong focused on the harm caused by revenge (37 percent, n = 201/537). Among respondents who viewed forgiveness as morally right, a plurality (27 percent, n = 192/714) focused on forgiveness as a means to combat chaos and disorder. It is less clear why some respondents viewed forgiveness as morally wrong, in part because so few did. Parsing these responses suggests that respondents in this category felt that even the forgiving response was too assertive, and that the focal character should have either ignored the antagonist or been more apologetic. This included statements such as “The American could have just said “excuse me” or quietly moved away from the foreign tourist”; and “His response did not even take into consideration the feelings of the visitors. He was completely self centered. Thus leaned very much towards wrong.”
Proportions of Moral Accounts by Condition and Moral Evaluation
Source: Authors’ tabulation.
As figure 1 shows, reciprocity, harm, and chaos are three primary moral accounts used across moral evaluations. Interestingly, reciprocity—the account most used to argue for the moral rightness of revenge—was also the second most commonly invoked account to argue that revenge is morally wrong (18 percent, n = 96). Respondents who invoked reciprocity viewed the shove as a disproportionate response, and therefore in violation of the principle of reciprocity. Thus the difference appears to hinge on whether a verbal insult is viewed as provocative enough to justify physical retaliation. Respondents in the camp viewing retaliation as a violation of the reciprocity principle made arguments such as “He was morally wrong. A person does not have to act that way over such a small issue”; “He was not morally right, because he had not been touched. He took it upon himself to make the issue a physical one”; and “Morally wrong for sure. There’s not reason to fight over something so silly like that.”
In contrast, many of reciprocity-based explanations for why revenge is morally right cited a right—and in some cases an obligation—to stand up for oneself and one’s in-groups, as illustrated by the following examples:
“We still have an honor culture particularly in some parts of the country and it is morally right to stand up against some insulting you, or your country. It was right because it was in response to an unreasonable insult. An insult against ALL Americans. IF it had just been personal, it would be easy to ignore, or end with a go F*** yourself and move on. The foreigner insulted an entire nation and its population. It should have been answered.”
“Morally right up to a point. He shouldn’t have to accept being talked to like that (generalized as an ‘ugly American’), but the shoving and escalation didn’t really help the perception of Americans in this situation. If I had to choose him acting this way, or doing nothing, I would go with this response honestly. That’s probably a bit of my overt patriotism coming out there—‘no one should talk bad about the USA.’”
“Morally right—yes. He had a right to protect his honor and that of his associates. If someone was to call his wife a whore would he not have the moral right and near-duty to defend her honor? This situation is not as drastic, but it is the same principle.”
These points were also echoed in shorter comments, such as “because as a human you don’t just stand there and take another persons abuse because they believe in another system”; “He was just sticking up for his party”; and “stood up for what he believes in.”
Those who viewed revenge as morally wrong most commonly gave harm or care-related justifications. Many viewed the escalation from a verbal dispute to physical aggression as intrinsically wrong: “It was wrong. You shouldn’t react that way to someone’s words. You definitely shouldn’t get physical.” “Morally wrong. It is never ok to put your hands on another person.” “Morally wrong, because he sought to do the other person physical harm.”
Along these lines, other responses argued for a more forgiving response: “Wrong. Give people a chance. Intolerance and violence shouldn’t be the gut reaction. The American tourist should have moved on considering he knew nothing of the other person’s situation.” “Wrong, because if you love your neighbor as yourself, you wouldn’t respond that way. A gentle answer turns away wrath, love your enemies. Hurt people hurt people. . . . Love is the only right response.”
Respondents who viewed forgiveness as morally right most often justified this assessment with accounts of chaos-disorder. These accounts focused on efforts to deescalate the situation: “Morally right; he took what could be an explosive situation and tamped it down.” “He was right not to say anything at the time things could have spiraled out of control.” “I think he handled the situation well, deescalating rather than going to the easier route and arguing back; there’s no moral high ground to starting a fight.”
A number of responses concern the potential spread of disorder: “He was right to diffuse the situation and prevent it from becoming a physical altercation or one that would have ruined either group’s night.” “I would say it was right. The reason it was right, the Eagles fan had the foresight to see a major brawl could have erupted if he had chose to be a combatant with the Cowboys fan.” “Right, What made it right was he didn’t contribute to an unnecessary uproar and spoil everybody’s good time. A true warrior wins the fight by avoiding it altogether.”
Q3: Is morality a pathway to status?
Our previous work finds that forgiving individuals are perceived to be greater in status than vengeful individuals in part because they are perceived as more group motivated (Benard et al. 2021). We assess whether respondents also perceive forgiving individuals as more status worthy because they perceive them as more moral. We evaluate whether the results are consistent with morality and group motivation serving as separate paths to status.
We evaluate this question using the Karlson-Holm-Breen (KHB) method, which allows for multiple mediators and comparisons across models with continuous and binary responses (Kohler, Karlson, and Holm 2011). We use the khb command in Stata to fit a reduced model (no mediators) which regresses status on the forgiving-vengeful condition (forgiving = 1), using the group-individual motivation condition as a control (group motivation = 1), and a full model that includes a binary indicator for whether the respondent viewed the behavior as morally right (1 = morally right, 0 = all other assessments), and the measure of perceived group motivation as mediators.
Table 2 includes the coefficients from the reduced and full models as well as the breakdown of the indirect effect of forgiveness through moral evaluation.12 The total effect of the forgiving manipulation is b = 1.4 (table 2), indicating that controlling for the group motivation manipulation, respondents view forgiving actors as 1.4 scale points more status worthy than vengeful actors. The total indirect effect via the mediators is 0.92 (p < .001). The model thus estimates that, of the 1.4-point increase in perceived status in the forgiving condition, 0.92 points (66 percent) are transmitted indirectly, through increases in perceived group motivation and morality. Of this total indirect effect, 0.52 points (57 percent) are estimated to occur through perceived group motivation and 0.40 points (43 percent) through perceived morality.13
Coefficients from KHB Mediation Analysis of Status on Moral Evaluation
Q4: Are specific moral accounts differentially related to status?
We explored whether specific moral accounts are differentially related to status. Extending the mediation model in table 2, we fit an alternative model that adds three common moral accounts (reciprocity = 1; harm = 1; chaos = 1) as potential mediators.14 These models find little evidence that moral accounts independently explain the total effect of forgiveness on status net of moral evaluation and group motivation, so we do not include them in our final mediation analysis.
Indeed, when we examined mean status by the three most commonly offered moral accounts: reciprocity, chaos, and harm, we found that status varied little by moral account. Figure 2 presents status by moral account, graphed separately by revenge-forgiveness condition and moral evaluation (right-wrong) to avoid confounding moral accounts with the experimental conditions or perceived morality. We limited our assessment to these three moral accounts due to sample size issues. Because only four to five respondents both viewed forgiveness as morally wrong and cited one of these three moral accounts, we exclude this category from the graph. Even with these steps, only seventeen respondents who viewed revenge as morally right cited chaos or harm in their justifications. The other categories ranged from fifty-two to 201 observations. Status did not appear to be driven by the moral account offered. Instead, whether the behavior was viewed as morally right or wrong was the main factor determining status worthiness.
Mean Status by Moral Account
Source: Authors’ tabulation.
DISCUSSION
Intergroup conflict is an influential, recurrent, and costly part of social life. It presents dilemmas without clear solutions: will retaliating for this grievance deter future aggression from rivals or perpetuate an endless cycle of retaliation? Will forgiving my enemy lead to productive, peaceful relationships, or is it a sucker’s bet? This ambiguity likely underlies the variation in the kinds of behaviors that groups value in conflict. Why do we valorize those who make peace with rival out-groups at certain times but elevate those who destroy our rivals at other times? The answers to these questions implicate key human motives, relationships, and social structures, as we seek to manage our often-competing interests, loyalty and attachment to in-groups, and striving for prestige, honor, and respect.
Increasingly, scholars have recognized that morality is closely intertwined with these factors. Indeed, morality and a sense of moral righteousness is interwoven and essential to furthering a group’s political goals, status claims, and their intragroup cohesion (Koenig 2022, this issue). Further, participating in conflicts is not only a way toward a policy goal, but also a way to establish oneself as a moral, status-worthy group member (Leicht 2022, this issue).
Building on this recognition of the importance of morality in motivating group behaviors, we ask whether revenge or forgiveness are seen as morally right, why respondents view them this way, and whether moral behavior in conflict provides a path to status in groups. We examine whether the status worthiness of moral behavior varies based on the specific moral account attached to that behavior.
We answer these questions by extending prior work on group-motivated revenge and forgiveness. We find that, for the conflict we examined, forgiveness was more often viewed as morally right than revenge. We find evidence consistent with the idea that morality serves as a pathway to intragroup status, alongside perceived group motivation. Respondents who viewed forgiveness as morally right most often justified these evaluations in terms of the chaos and disorder avoided, whereas those who viewed revenge as morally right most often drew on the reciprocity principle. Respondents who viewed revenge as morally wrong most often drew on the harm-care principle, but a substantial number also viewed revenge as a disproportionate violation of the reciprocity principle. Our results suggest that, despite the different accounts that Americans draw on to make sense of their moral evaluations, these accounts do not themselves create status differentials. Instead, moral evaluations coexist with established mediators such as group motivation in creating status differentials.
It is likely that the extent to which conflict produces opportunities to gain or lose status is moderated by other group characteristics, such as the degrees of compression, clarity, and rigidity in status hierarchies (Accominotti, Lynn, and Sauder 2022, this issue). For example, occupants of rigid status hierarchies may be less inclined to participate in intergroup conflict, given the relatively low possibility of gaining status by doing so. Collectively, these results point to moral evaluations as a key mechanism to create status differences and inequality within groups and highlight the contours of social groups for whom it is a stronger mechanism.
Our findings also have implications for social inequalities. Within groups, conflict can create or exacerbate unequal status hierarchies, as Georg Simmel ([1908] 1955) suggests. Leaders can exploit conflict to fend off competitors for their position (Barclay and Benard 2013; Pickering and Kisangani 2005). These processes might also lead some group members to have higher propensities to bear the individual-level burdens of these group-motivated behaviors. For example, if participating in conflict is primarily a way for low-status individuals to gain status, then low-status individuals may be disproportionately exposed to violence. Between groups, norms of revenge versus forgiveness can encourage intergroup oppression and create vicious cycles of revenge (Papachristos 2014).
Limitations
Our design has a number of strengths. By using a national probability sample, subdivided into three conceptual replications based on different group identities, as well as experimentally manipulating vengeful and forgiving behavior, we seek a balance of both external and internal validity. Our use of both closed and open-ended questions allows us to test our research questions yet remain open to new insights. That said, our findings could be strengthened or extended in several ways.
First, we examine a specific scenario: witnessing an out-group member insult an in-group member in a public setting, and the in-group member responding. The high cost of fielding a large, nationally representative sample placed some limitations on the number of conditions and scenarios we could examine. We adopted this particular scenario in part because it provides a look at more common, everyday forms of conflict that many people experience, rather than the lethal violence examined in a number of other studies of intergroup revenge (Gould 2003; Papachristos 2009). We also use this design because it captures a crucial moment in conflict—the instance when a verbal dispute might become a physical one. Recent work suggests that conflict is costly in part because it has a tendency to reverberate through social networks via cycles of retaliation and score settling (Papachristos 2014). Understanding when people do and do not approve of group members crossing this line helps us understand the conditions under which conflicts are more likely to spiral out of control.
The relative value groups place on revenge and forgiveness likely varies across structural and cultural factors. Our findings provide a starting point for mapping the morality of revenge and forgiveness cross-culturally. For example, the finding that forgiveness is most often viewed as morally right because of its role in preventing chaos and disorder suggests that forgiveness should be more valued as the likelihood or costs of disorder increase or in cultures in which avoiding disorder is emphasized.
Another limitation of our study is that our mediators and outcome variables—status, group motivation, and morality—are measured cross-sectionally. This means that, although we have causal evidence of the effect of revenge and forgiveness on these variables, we do not have causal evidence of the mediating effect. We also cannot be sure that perceptions of morality and group motivation are fully separate pathways to status. Future work could manipulate these pathways to tease out their causal link to status.
Finally, our study uses a vignette design and thus measures attitudes rather than behavior. We view this design as appropriate for our research because we sought to understand how people view acts of revenge and forgiveness and not whether respondents could predict their own vengeful or forgiving behavior. We ask respondents about relatively common, everyday conflicts of the type that they have likely witnessed in person, read about, or seen on social media. We do not ask them to form opinions about rare or extremely emotional events, or to report their opinions about highly personal or controversial issues. This study focuses on questions we believe our respondents will be both willing and able to answer.
CONCLUSION
People view intergroup forgiveness as moral and correspondingly status worthy while praising its value in preventing chaos and disorder in intergroup conflict. Yet this finding is not universal: some respondents judged revenge to be morally right. These findings highlight the value of morality as a link between two productive, but largely separate literatures: those on intergroup conflict and social status. Concepts of morality are bound up in group identities, and thus play a key role in conflict. At the same time, ostensibly moral behaviors, such as altruism, generosity, and self-sacrifice, are judged to be both moral and status worthy. Intergroup forgiveness illustrates both of these tendencies: forgivers are perceived as moral and group motivated, and both motives appear to increase perceived status worthiness. More broadly, this work speaks to the fundamental idea that conflict between groups shapes relationships between groups. Adding to a large body of work on the cohesion-shaping effects of conflict, we contribute to a growing set of findings indicating that conflict shapes intragroup status structures as well. In contributing to our conceptual understanding of intergroup conflict, we aim to lay the groundwork for practical steps toward reducing the costs of conflict.
FOOTNOTES
↵1. Individuals can also gain influence in groups through the use of threats and coercion. Although some scholars view this as a path to status (Bai 2017), we see coerced deference as analytically distinct from the voluntarily conferred deference characteristic of status hierarchies (Ridgeway and Markus 2022, this issue; Henrich and Gil-White 2001).
↵2. Sample demographics are included in online supplement 1. All supplemental materials can be viewed at https://www.rsfjournal.org/content/8/6/122/tab-supplemental.
↵3. Kay Deaux and her colleagues use these labels broadly: the ethnicity cluster includes regional identities such as Southerner, racial identities such as African American, and national identities such as American.
↵4. In developing the vignettes, we read news articles and watched social media videos depicting similar altercations. We elicited feedback from two introductory sociology classes (n = 25 and n = 40) and pilot tested the vignettes on Amazon’s mTurk platform. NORC also conducted and provided recordings for cognitive interviews with survey respondents (n = 9).
↵5. Full text of all scenarios is available in online supplement 7.
↵6. An additional two participants did not fill out the forgiving-vengeful manipulation check questions, leaving an analytic sample of 1,911 for the manipulation check only. The manipulation check consisted of two items: the extent which the in-group member was viewed as not forgiving or forgiving and vengeful or not vengeful, on 9-point bipolar scales. These were averaged to form a single measure (α = .83).
↵7. Teams of coders coded each item until they reached acceptable levels of interrater reliability; the minimum threshold was a Krippendorf’s alpha of 0.9. Open codes were developed using pilot test data from Amazon’s mTurk platform. For more details of the coding procedure, see online supplement 8.
↵8. Spelling errors in survey responses are in the original. Quotations marks have been added or adjusted for clarity as needed.
↵9. See online supplement 2.
↵10. For robustness checks, models with controls and those with survey weights are presented in online supplements 3 and 4. Results are substantively similar to those presented here.
↵11. See online supplement 6.
↵12. Detailed results are included in online supplement 5.
↵13. We conducted a sensitivity analysis of the indirect effect, using the multimed command in R (Tingley et al. 2014). This indicated that indirect effect of morality is expected to be 0 when the amount of residual variance explained by interaction heterogeneity is 16 percent, or when the total amount of variance explained by interaction heterogeneity is 10.5 percent (available from the authors).
↵14. See online supplement 5.
- © 2022 Russell Sage Foundation. Benard, Stephen, Long Doan, D. Adam Nicholson, Emily Meanwell, Eric L. Wright, and Peter Lista. 2022. “To Forgive Is Divine? Morality and the Status Value of Intergroup Revenge and Forgiveness.” RSF: The Russell Sage Foundation Journal of the Social Sciences 8(6): 122–39. DOI: 10.7758/RSF.2022.8.6.07. We thank Sydney Alcaraz, Callie Cleckner, Benjamin Gallati, Melissa Garcia, Patrick Kaminski, Eunhye Lee, Jennifer Lee, Hannah Lloyd, Christian Negron Rolon, Alyssa Place, Kara Snawder, Jenny Yang, and Iris Zhao for their work coding the open-ended responses, and Benjamin Hartmann for writing the R code for the sensitivity analyses. We also thank the editors and conference participants for their helpful feedback, and the National Science Foundation (Award ID #1728889) and Indiana University for supporting this research. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation or other sponsors. Direct correspondence to: Stephen Benard, at sbenard{at}indiana.edu, Department of Sociology, Indiana University, Bloomington, IN, 47405, United States.
Open Access Policy: RSF: The Russell Sage Foundation Journal of the Social Sciences is an open access journal. This article is published under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.
REFERENCES
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.