Dynamics of Moral Repair: Forgiveness,Self-Forgiveness,and the Restoration of Value Consensus as Interdependent Processes

Abstract

Most psychological research has investigated victims’ forgiveness and offenders’ self-forgiveness separately, ignoring interactive and dynamic processes between them. We suggest that both parties are interdependent in their attempts to revalidate the values violated by the wrongdoing. In the present study, both partners of close relationships dyads (including 164 complete couples) were surveyed over three time-points following the report of a wrongdoing by one of the partners. Latent growth modeling showed that victims’ forgiveness was associated with growth in their perception of a value consensus with the offender. Victims’ value consensus perception was associated with growth in offenders’ perception of value consensus and engagement in genuine self-forgiveness (working through). However, directly, forgiveness was associated with decline in offenders’ genuine self-forgiveness, while offenders’ self-punitiveness was associated with decline in victims’ forgiveness. The findings highlight the regulatory function of victim forgiveness and the pivotal role of restoring value consensus in interactive moral repair.

Keywords

forgiveness self-forgiveness value consensus value affirmation reconciliation

Humans almost inevitably hurt or wrong others with whom they are in relationship. Hurt can arise from failing to meet expectations or honor promises, ignoring others, showing disrespect, breaking trust, or being abusive. Relationship transgressions, even simple ones, can escalate, and the long-term fallout from mismanaged interpersonal transgressions can be costly and may include relationship breakups and productivity loss (see Whitton et al., 2018). Wrongdoing has at least two sides: one person (“victim”) has been harmed, hurt, or insulted by another person (“offender”). Current social psychological research has helped understand each individual party’s behavior, including when victims seek punishment of offenders (e.g., Yoshimura & Boon, 2018) and offenders self-punish (e.g., De Vel-Palumbo et al., 2018), when victims seek and offenders offer apologies (e.g., Fehr & Gelfand, 2010; Schumann, 2018), and—important here—when victims offer forgiveness and offenders self-forgive (e.g., Fehr et al., 2010; Woodyatt et al., 2017).

However, most research has focused separately on victim or offender perspectives. It has focused on intrapersonal processes that lead victims to forgive or offenders to self-forgive, but has largely ignored the interactive, dynamic processes between victim and offender. Yet, interactions between the parties are critical and often cause a conflict to escalate or to resolve. For example, researchers have argued that for reconciliation to occur, victims and offenders must meet each other’s psychological needs arising from a transgression (Shnabel & Nadler, 2008). Specifically, a transgression constitutes a violation of values presumed to be shared between victim and offender, which could potentially cause a rift in their common identity (Okimoto & Wenzel, 2008). We reason that both parties are interdependent in their attempts to regain social consensus that revalidates violated values.

In the present research, we explore whether acts of forgiveness and self-forgiveness can communicate one’s reaffirmation of values and the shared identity that those values represent. Conversely, the communicated affirmation of values by one party may also motivate the other party to engage in forgiveness or self-forgiveness. We examine forgiveness and self-forgiveness as interdependent processes in a dynamic of moral repair, that is, the restoration of moral selves and relationships that are constituted by shared values and standards (see Walker, 2006).

Forgiveness and Self-Forgiveness as Mechanisms of Moral Repair

Forgiveness has been defined as a victim’s prosocial change toward an offender, involving less vengeful and avoidant, and more benevolent motives (McCullough et al., 1997). Forgiveness can be an intrapsychic process of developing more positive motives and sentiments toward the offender, and can involve interpersonal communications to the offender (Baumeister et al., 1998). Similarly, it can be an emotional transformation, or a conscious decision to forgive that regulates negative behavioral intentions and modifies motives and sentiments (Worthington, 2006). Crucially, however, forgiveness does not condone, downplay, or excuse the offender’s behavior nor let the offender off the hook (see Enright & The Human Development Study Group, 1991).

Self-forgiveness has been defined, analogously, as transformation of emotions and motives of offenders, away from self-punishment and self-condemnation toward greater benevolence and compassion for oneself (Hall & Fincham, 2005). As with forgiveness, self-forgiveness implies that offenders do not downplay the wrongfulness of their behavior, the harm caused, or their responsibility. Rather, offenders need to take responsibility for their actions, experience remorse, make amends (if possible), and separate their wrongdoing from their self-worth to regain moral integrity (Fisher & Exline, 2010; Wenzel et al., 2012). While interpersonal forgiveness is a unitary construct (Forster et al., 2019), the empirical evidence shows that different dimensions of self-forgiveness need to be distinguished (Woodyatt & Wenzel, 2013a). For genuine self-forgiveness, offenders need to work through their responsibility and guilt, seek to learn from their wrongdoing, commit to doing better, and thus reaccept themselves as a good, if imperfect, person. However, acceptance of guilt can also imply self-punitiveness; feelings of self-condemnation and the need to be unforgiving to self. These processes align with dual-process accounts wherein self-forgiveness includes forms of both emotional release from self-condemnation and moral reform through facing one’s wrongdoing (Griffin et al., 2015).

Both forgiveness and self-forgiveness can have positive consequences for victim and offender, respectively, by reducing negative affect, stress, and rumination, and increasing mental and physical health (Davis et al., 2015; Riek & Mania, 2012). They may also promote moral and relationship repair by increasing trust and self-trust, empathy, and willingness to reconcile (Wenzel et al., 2010; Woodyatt & Wenzel, 2013a). It is therefore crucial to understand how forgiveness and self-forgiveness develop.

Previous research has focused on intra-individual factors of victim (for a review, see Fehr et al., 2010) or offender (e.g., Hall & Fincham, 2008), as they determine (self-)forgiveness as an outcome. Some research also suggests reciprocal causality for many of these intrapersonal determinants. For example, forgiveness is not only caused by victims’ empathy and identification with the offender, but it also increases victims’ empathy and identification with the offender (Karremans & Van Lange, 2004). Forgiveness is not only impeded by rumination, but also reduces rumination (Ysseldyk et al., 2007). Self-forgiveness can result from conciliatory behavior from an offender, but it can also lead to acts of reconciliation (Carpenter et al., 2014; Woodyatt & Wenzel, 2013a). Thus, forgiveness and self-forgiveness may involve cyclical, dynamic processes that defy a simple cause–effect paradigm. Evidence also shows that an offender’s apology increases a victim’s forgiveness (McCullough et al., 1997), and perceived victim forgiveness increases an offender’s self-forgiveness (Hall & Fincham, 2008). This preliminary evidence suggests that forgiveness and self-forgiveness are bilateral and negotiated, and therefore best understood as dyadic, interdependent processes.

However, there are few, if any, dyadic studies on the interplay between forgiveness and self-forgiveness. Dyadic studies in this field tend to use symmetric designs where both relationship partners report on their experiences in the same roles (not complementary roles). For example, in a study by Pelucchi et al. (2013), both relationship partners recalled a wrongdoing they committed against the other, rated their self-forgiveness for their own wrongdoing, as well as their relationship satisfaction. It was found that the more unforgiving either partner was toward themselves, the less satisfied the other partner was with their relationship. While interesting, this cross-partner effect does not address the forgiveness/self-forgiveness interplay.

The Violation and Revalidation of Shared Values

To understand the interdependencies between forgiveness and self-forgiveness, we may consider the key symbolic meaning of a transgression and the shared damage it has caused for both parties. Transgressions can lead to physical, emotional, or material harm, as well as damage to either party’s personal integrity in terms of threats to the status/power of the victim and the moral image of the offender (Shnabel & Nadler, 2008). Fundamentally, however, a transgression represents the violation of a shared understanding of appropriate behavior, based on values (e.g., honesty, trust, respect) that were assumed to be shared between the parties and within their relevant community (Wenzel et al., 2008). In communities and relationships (communities of two), shared values help to define the members’ or partners’ common identity. Transgressing these values threatens that shared identity; commitment to their identity motivates the parties to revalidate the values.

In principle, the involved parties can independently affirm violated values without engaging with the other, through (self-)punitive or (self-)forgiving acts (e.g., De Vel-Palumbo et al., 2018; Wenzel & Okimoto, 2010). In fact, they may even do so through differentiation from the other, such as when victims (or third parties) impose a punishment to censure the offense or symbolically exclude the offender from their moral community, thus reinforcing the values within the rest of the community (Okimoto & Wenzel, 2009). However, where reconciliation is the goal, hence the repair of the shared identity, value consensus between the parties is paramount. It is perceived consensus that validates values as true and defining of one’s shared identity (Turner, 1987). This would apply in particular (but not only) to close relationships where the identity and community of two cannot continue to exist without both parties remaining a part of it. We propose that restoration of value consensus is a key mechanism of moral repair and reconciliation following a transgression.

Dynamics of Forgiveness, Self-Forgiveness, and Perceived Value Consensus

Based on this perspective, several dynamics are possible—within each party and, crucially, between both parties (see Figure 1). First, there are theoretical arguments for reciprocal relationships between forgiveness and value affirmation within victims. When a victim decides to forgive the offender, that act implies an expectation, trust, or self-persuasion that the offender is still committed to the violated values. This can alleviate the victim’s concern about shared values and repair their perceptions of a value consensus (Wenzel & Okimoto, 2010, 2012); it can restore a subjective sense of “we” identity or closeness (Karremans & Van Lange, 2008). Conversely, the victim’s perception of value consensus with the offender promotes the victim’s emotional forgiveness. It increases forgiving sentiments toward the offender and willingness to reconcile (Wenzel & Okimoto, 2010). Within victims, there may thus be a reciprocal dynamic between forgiveness and perceptions of value consensus (Figure 1, paths a₁ and a₂).

Figure 1.

A model of forgiveness/self-forgiveness dynamics.

Second, there are theoretical arguments for reciprocal relationships between genuine self-forgiveness and value affirmation within offenders. Offenders may accept responsibility and work hard at thinking through what they have done—constitutive elements of genuine self-forgiveness. In doing so, they realize that what they have done does not define them as a wrongdoer or offender, and that they are indeed committed to shared values. Genuine self-forgiveness could thus bolster offenders’ value consensus perceptions. Conversely, offenders’ affirmation of violated values has been found to reduce their defensiveness or reluctance to accept responsibility, and to promote genuine self-forgiveness (Wenzel et al., 2012; Woodyatt et al., 2017). Thus, also within offenders, there may be a reciprocal dynamic between self-forgiveness and perceptions of value consensus (Figure 1, paths b₁ and b₂).

Third, central to the present study, there are arguments for plausible reciprocal relationships between the victim’s and the offender’s value consensus. One party’s communication of value consensus may validate and boost the other’s affirmation of shared values. Both parties thus engage in mutual social influence toward perceptions of shared value consensus (Figure 1, paths c₁ and c₂). Similarly, both parties would likely respond with conciliation to the partner’s affirmation of a shared value consensus. On one hand, victims are likely to see the offender’s affirmation of shared values as reason to be more forgiving to the offender: The offender is seen as reaffirming the violated values, understanding their wrongs, and committing to better future behavior (e.g., through an explicit apology; see Fehr et al., 2010). On the other hand, offenders are likely to feel assured of their acceptance by the victim’s affirmation of a value consensus, hence to be less defensive (Woodyatt & Wenzel, 2013b) and more inclined to genuinely forgive themselves (paths d₁ and d₂).

Finally, the victim’s forgiveness and the offender’s self-forgiveness may be reciprocally related to each other. Specifically, if the victim sees the offender taking responsibility for wrongdoing and working toward genuine self-forgiveness, the victim may respond empathically and conciliatorily with forgiveness. Likewise, if the offender sees the victim’s forgiveness, they may respond with reciprocal conciliation (Kelln & Ellard, 1999) and engage in a genuine working through of their wrongdoing as a pathway to self-forgiveness (Hall & Fincham, 2008; Figure 1, paths e₁ and e₂).

These theoretical propositions (Figure 1) relate to genuine self-forgiveness as working through one’s responsibility, which implicates value reaffirmation and consensus seeking. For self-punitiveness (or, overcoming it as a path to self-forgiveness), we would not necessarily expect the same relationships. Self-punitiveness shares with genuine self-forgiveness an acknowledgment of responsibility and blame; but, in contrast to genuine self-forgiveness, it is antithetical to self-forgiveness as an outcome. While we do not advance an alternative model for self-punitiveness, we explore it in the present research for comparison.

The Present Study

This research investigates the dyadic dynamics between forgiveness, self-forgiveness, and shared value affirmation, which are hypothesized to unfold after an interpersonal wrongdoing as victim and offender engage with each other. To explore our conceptualization empirically, innovative methodological approaches are required that focus on dyads and the effects of their mutual engagement.

In this study, we investigated real-life transgressions between actual relationship partners. We assessed both partners’ cognitions and feelings repeatedly over several days following the transgression, thus providing longitudinal data for analysis. Treating the dyad as the unit of analysis, we used multivariate (or parallel) latent growth modeling (Bollen & Curran, 2006) to simultaneously model the initial level and trajectory of victim and offender concepts (forgiveness, self-forgiveness, value consensus perceptions), and to assess reciprocal associations between these concepts. Specifically, we investigated how the initial level of one concept was associated with the change over time in another concept, both within victim and offender roles as well as—crucially—between parties.

Method

Procedure

We employed a prospective-longitudinal design, where the dyad was the unit of analysis. Interested couples were invited to participate in the study before the critical event:

Over the next days or possibly weeks, we would like to ask you to be aware of any wrong that you feel your relationship partner commits against you—be it minor or major. It could be anything of which you think your partner should not have behaved in that way to you. It could be an act of disrespect, a trust violation, an indiscretion, psychological hurt, physical harm, and so on.

When one relationship partner felt wronged by the other, the wronged person (victim) completed an initial questionnaire within 24 hr, briefly describing the event. This brief survey triggered an email to both parties, asking them to complete an online survey within 24 hr. The completion of that first survey then triggered another email, with 24 hr delay, which instructed participants to access the next online survey and complete it within the next 24 hr (i.e., up to 48 hr after completion of the first survey). This process was repeated for a third survey. Thus, in total the study could be completed within 48 to 120 hr (for a similar design, with equivalent timeframe but focused on victims only, see Wenzel et al., 2010). We chose the narrow timeframe because, given the prospective approach, we expected rather nonserious incidents that would commonly not cause a long-lasting upset, often being resolved within a few days (see McCullough et al., 2010). Both parties provided survey responses from their respective perspectives at three measurement points.¹

Participants

We recruited University students and their relationship partners for this study. Given the novelty of this research and uncertainty as to what effect sizes to expect, we conducted this research in two stages. In Stage 1, we tested the overall model on an initial sample of relationship dyads, guided by recommendations that latent growth models require a minimum of 50 (Hamilton et al., 2003) to 100 cases (Curran et al., 2010). However, these rough recommendations do not relate to dyadic analysis with parallel growth processes (i.e., for two or more variables in parallel). Hence, we used the findings from the initial sample for a more precise power analysis using Monte Carlo simulation. We targeted 75 couples for our initial sample (Sample A). In total, 82 couples initiated the study. For 6 of those, one partner did not complete the surveys, and there was therefore only data for one role (five victims, one offender). However, all the data were used in our analyses (using full information maximum likelihood estimation in AMOS).

Based on the model results from Sample A, we conducted a Monte Carlo analysis in MPlus, which established that we needed to double the sample size for sufficient statistical power (see Online Supplementary Materials). For Sample B, we recruited 90 relationship dyads who initiated the study. Two offenders failed to respond, resulting in incomplete dyads. A further three participants in the victim role and three participants in the offender role dropped out at Times 2 or 3, leaving incomplete data for these dyads. Nevertheless, all data were included in the analyses. Thus, both subsamples combined comprised 164 complete dyads, plus eight individuals where the partner data was missing. Of the 164 complete dyads, 151 were different-sex and 11 same-sex partners (nine female, two male couples), and another two couples indicated their gender as nonbinary. The individual average age was 24.0 years. Of the 174 female participants, 122 (70.9%) were in the victim role; of the 157 male participants, 46 (26.7%) were in the victim role.

The following item statistics are based on the complete sample. Likewise, we report below the findings for the total sample only, but the results for the two subsamples (for the final model) are detailed in the Online Supplementary Materials. The findings were highly similar between the two subsamples, suggesting replicability of the results.

Measures

The surveys that each of the participants completed contained a larger battery of questions, for exploratory purposes and outside the scope of this article (see Online Supplementary Materials). Here, we only detail the measures relevant for the present investigation. Except for the description/classification of the incident and the measure of severity, which were only assessed once within both parties’ initial questionnaires, the main variables were repeated at each measurement point. The measures varied depending on the victim or offender role that participants assumed in the situation. All items were rated using 7-point response options from 1 = strongly disagree to 7 = strongly agree. For multi-item scales, scores were computed by averaging item responses.

Seriousness

Both parties rated the seriousness of the wrongdoing via an equivalent three-item scale: “What do you currently think about the incident? My partner’s [for offenders: My] behavior was . . . wrong; serious; inexcusable” (victims: α = .69. offenders: α = .75).

Shared value consensus

Each individual’s perception of shared value consensus with their partner was measured with same three items for both victim and offender roles (adapted from Wenzel & Okimoto, 2010): “I have similar values and beliefs as my partner,” “My partner and I share an understanding of what is important,” and “My partner and I agree about what is right and wrong” (victims: αs = .82, .86, .85, for Times 1 to 3, respectively; offenders: αs = .78, .87, .91, for Times 1 to 3, respectively).

Forgiveness

As a measure of forgiveness, victims completed the 18-item Transgression-Related Interpersonal Motivation scale (TRIM-18 scale; McCullough et al., 2006). The scale contains five items assessing revenge (e.g., “I’ll make him/her pay”), seven avoidance-related items (e.g., “I’m trying to keep as much distance between us as possible”), and six benevolence-related items (e.g., “Although his/her actions hurt me, I have goodwill for him or her”). However, this three-factor structure does not always reveal itself empirically (McCullough et al., 2006). In the present research, there was a strong first factor explaining 41% of the variance at Time 1, on which all items loaded >.5 except for four items that fell below this threshold (but still with loadings from .41 to .50). This is consistent with findings by McCullough et al. (2010) and Forster et al. (2019) who, using item response theory, concluded that forgiveness is adequately represented as a single dimension from less vengeful to more benevolent and non-avoidant responding. In the present study, after reverse-scoring revenge and avoidance items, Cronbach’s alpha was high (α = .90, .93, and .93, at Times 1 to 3, respectively), and all items were averaged into a composite scale score.

Self-forgiveness

The offenders’ self-forgiveness was measured in terms of two processes (see Griffin et al., 2015): lack of self-punitiveness (or self-condemnation), and genuine self-forgiveness via working through one’s wrongdoing and guilt. The measures were taken from the Differentiated Process Scale of Self-Forgiveness (Woodyatt & Wenzel, 2013a), which measures self-punitiveness with seven items such as “I deserve to suffer for what I have done,” and genuine self-forgiveness also with seven items, including “I have spent time working through my guilt.” Consistent with the conceptual distinction, principal component analysis yielded two components that explained (at Time 1) 30% and 26% of the variance, respectively, following Varimax rotation. Coefficient reliability was high for both self-punitiveness (α = .86, .89, and .90, at Times 1, 2, and 3, respectively) and genuine self-forgiveness (α = .85, .86, and .90, at Times 1, 2, and 3, respectively). Self-punitiveness and genuine self-forgiveness were positively correlated at all three time-points (rs = .59, .53, and 50), reflecting that both responses involve taking responsibility and accepting guilt. However, conceptually self-punitiveness is a negative and genuine self-forgiveness a positive indicator of self-forgiveness (Woodyatt & Wenzel, 2013a).

To check construct validity, we also included a one-item, direct measure of self-forgiveness: “I have been able to forgive myself for what happened.” Regressing the single-item self-forgiveness at Time 2, controlling for its Time 1 measure, onto self-punitiveness and genuine self-forgiveness measured at Time 1, yielded a significant negative lagged effect of self-punitiveness (β = –.29, p =. 006) and a positive lagged effect of genuine self-forgiveness (β = .19, p =. 049). Similarly so for Times 2 and 3: regressing the single-item self-forgiveness at Time 3, controlling for its Time 2 measure, onto self-punitiveness and genuine self-forgiveness measured at Time 2, yielded a significant negative lagged effect of self-punitiveness (β = –.20, p =. 048) and a positive lagged effect of genuine self-forgiveness (β = .20, p = .030; see Online Supplementary Materials for details). Hence, lack of self-punitiveness and engaging in genuine self-forgiveness are independent processes predicting a state of subjective self-forgiveness. For the present analyses, we therefore used both self-punitiveness and genuine self-forgiveness as different processes of self-forgiveness, although genuine self-forgiveness was our main focus due to its predicted relationship with shared value consensus.

Results

Diverse relationship transgressions were reported. According to victims’ own classifications (multiple responses possible), 60.5% were acts of disrespect, 17.4% neglect or rejection, 15.7% wrongful accusations, 14% betrayal of trust, 12.8% insults or verbal abuse, 9.9% deceit, secrecy or lying, 4.1% involvement with another person, and 16.9% were other types of wrongdoing. As shown in Table 1, victims rated the transgression as significantly more serious than offenders did, paired samples t(162) = 7.34, p < .001, even though both parties’ seriousness ratings were significantly correlated (see Table 2). The means and correlations for all other variables are also shown in Tables 1 and 2, respectively.

Table 1.

Means (and Standard Deviations) for the Complete Sample.

Variables	Time 1	Time 2	Time 3
Seriousness-V	4.33 (1.30)	–	–
Seriousness-V	n = 169
Seriousness-O	3.39 (1.45)	–	–
Seriousness-O	n = 165
Value consensus-V	5.18 (1.25)	5.23 (1.27)	5.45 (1.15)
Value consensus-V	n = 170	n = 167	n = 163
Values consensus-O	5.46 (1.12)	5.62 (1.03)	5.84 (1.01)
Values consensus-O	n = 166	n = 162	n = 159
Forgiveness-V	5.67 (0.93)	5.89 (0.96)	6.04 (0.96)
Forgiveness-V	n = 170	n = 167	n = 163
Genuine self-forgiveness-O	4.29 (1.21)	4.29 (1.26)	4.29 (1.43)
Genuine self-forgiveness-O	n = 165	n = 162	n = 159
Self-punitiveness-O	2.52 (1.22)	2.47 (1.16)	2.33 (1.17)
Self-punitiveness-O	n = 165	n = 162	n = 159

Note. All variables were measured on scales from 1 to 7. V = victim; O = offender.

Table 2.

Intercorrelations.

Variables	Gender-V	Gender-O	Seriousness-V	Seriousness-O	Value consensus-V			Value consensus-O			Forgiveness-V			Genuine SF-O			Self-punitiveness-O
Variables	Gender-V	Gender-O	Seriousness-V	Seriousness-O	T1	T2	T3	T1	T2	T3	T1	T2	T3	T1	T2	T3	T1	T2
Gender-V	–
Gender-O	−.63***	–
Seriousness-V	−.10	.03	–
Seriousness-O	−.04	−.004	.33***	–
Value C.-V T1	−.05	.08	−.27***	−.01	–
Value C.-V T2	−.09	.09	−.10	.08	.65***	–
Value C.-V T3	−.08	.10	−.05	.03	.59***	.72***	–
Value C.-O T1	.05	.02	−.12	.13	.32***	.34***	.28***	–
Value C.-O T2	.02	−.03	−.19*	−.03	.33***	.34***	.28***	.70***	–
Value C.-O T3	−.06	.09	−.06	.03	.42***	.38***	.28***	.67***	.73***	–
Forgiveness-V T1	−.03	.09	−.38***	−.13	.50***	.45***	.40***	.22**	.29***	.27***	–
Forgiveness-V T2	−.17*	.20*	−.26***	−.06	.34**	.50***	.43***	.25**	.28***	.26**	.66***	–
Forgiveness-V T3	−.17*	.19*	−.09	−.05	.25*	.43***	.52***	.15	.15	.25**	.53***	.77***	–
Genuine SF-O T1	.04	−.08	.11	.50***	.13	.15	.18*	.25**	.16*	.05	−.03	−.06	−.04	–
Genuine SF-O T2	.15	−.10	.12	.48***	.14	.17*	.16*	.26**	.20*	.16*	.02	−.06	.03	.70***	–
Genuine SF-O T3	.15	−.09	.22**	.37***	.13	.10	.10	.14	.05	.12	−.12	−.22**	−.12	.64***	.81***	–
Self-Punit.-O-T1	−.01	−.08	.28**	.54***	−.15	−.06	−.04	.01	−.14	−.15	−.27***	−.26**	−.23***	.59***	.45***	.46***	–
Self-Punit.-O-T2	−.04	−.04	.33**	.49***	−.14	−.10	−.12	.01	−.13	−.14	−.29***	−.33**	−.28***	.42***	.53***	.53***	.71***	−
Self-Punit.-O-T3	−.06	−.003	.29**	.42***	−.18*	−.15	−.15	−.05	−.23**	−.17*	−.32***	−.31***	−.27***	.38***	.37***	.50***	.72***	.75***

Note. Value C. = value consensus; Genuine SF = genuine self-forgiveness; Self-Punit. = self-punitiveness; V = victim; O = offender; T1, T2, and T3 = time-points 1, 2, and 3, respectively.

p < .05. **p < .01. ***p < .001.

Latent Growth Models

We applied latent growth modeling (Bollen & Curran, 2006) to investigate the trajectories of the main variables, and ultimately to test the predicted dynamic relationships within victim and offender roles as well as between them, treating the dyad as unit of analysis. Latent growth modeling, using structural equation modeling (AMOS 25), allowed us to estimate the intercept (initial status) and slope (growth rate) for variables repeatedly measured over the three time-points (see Figure 2). The intercepts were modeled as a latent factor with loadings to all three repeated measures fixed to one. The slopes were modeled as a latent factor with loadings to the relevant Time 1 measure fixed to zero, the loading to the Time 2 measure set to 0.5, and the loading to the Time 3 measure set to one.

Figure 2.

Latent growth model in Study 1.

We proceeded in three steps to build toward the overall models that would test the theoretical predictions in their entire complexity. We first investigated simple growth models of each variable of interest and tested their development over time: victim’s forgiveness and victim-perceived values consensus, and offender genuine self-forgiveness (and self-punitiveness) and offender-perceived value consensus. Given our primary interest in partner dynamics between victim and offender, we then tested two-variable parallel growth models of corresponding victim and offender variables, investigating effects of initial levels of each variable on change in the other variable. Finally, we combined these models into overall, four-variable parallel growth models, testing again for cross-variable intercept-to-slope effects. In our two- and four-variable parallel growth models, we also included the victims’ and the offenders’ gender, as well as both parties’ seriousness-of-offense ratings as covariates; these were treated as time-invariant covariates measured at the beginning of the study.

Single-Variable Growth Models

The results of the four single-variable latent growth models are summarized in Table 3. A comparison with Table 1 shows that the estimated means of the latent intercepts closely corresponded to the means of the respective variables at Time 1 (initial levels). The latent slopes reflect, given the chosen scaling of the slope indicator weights (0, 0.5, 1), the accumulated average change over the three timepoints; the estimated means of the latent slopes closely correspond to the differences in means between Times 1 and 3. There were significant positive slopes (i.e., increase over time) for victim and offender-perceived value consensus as well as victims’ forgiveness, whereas offenders’ self-punitiveness had a significant negative slope and, thus, decreased over time. For genuine self-forgiveness, the analysis found a nonsignificant slope on average; but it also indicated significant interindividual variance in the slope.

Table 3.

Unconditional Single-Variable Growth Models.

Variable		Estimate	SE	p
Value consensus—victim
Intercept	Mean	5.16	0.09	<.001
Intercept	Variance	1.22	0.21	<.001
Slope	Mean	0.30	0.08	<.001
Slope	Variance	0.80	0.30	.008
Model fit: χ²(1) = 1.54, p = .215; CFI = .998; RMSEA = .056 (.000, .221)
Value consensus—offender
Intercept	Mean	5.44	0.08	<.001
Intercept	Variance	0.85	0.15	<.001
Slope	Mean	0.40	0.07	<.001
Slope	Variance	0.10	0.23	.655
Model fit: χ²(1) = .077, p = .781; CFI = 1.00; RMSEA = .000 (.000, .133)
Forgiveness—victim
Intercept	Mean	5.68	0.07	<.001
Intercept	Variance	0.70	0.11	<.001
Slope	Mean	0.35	0.07	<.001
Slope	Variance	0.69	0.17	<.001
Model fit: χ²(1) = 1.25, p = .263; CFI = .999; RMSEA = .038 (.000, .211)
Genuine self-forgiveness—offender
Intercept	Mean	4.29	0.09	<.001
Intercept	Variance	1.02	0.17	<.001
Slope	Mean	−0.01	0.09	.947
Slope	Variance	0.67	0.31	.030
Model fit: χ²(1) = .001, p = .972; CFI = 1.00; RMSEA = .000 (.000, .000)
Self-punitiveness—offender
Intercept	Mean	2.55	0.09	<.001
Intercept	Variance	0.98	0.17	<.001
Slope	Mean	−0.20	0.07	.005
Slope	Variance	0.06	0.28	.835
Model fit: χ²(1) = .316, p = .574; CFI = 1.00; RMSEA = .000 (.000, .167)

Note. CFI = confirmatory fit index; RMSEA = root mean square error of approximation.

Two-Variable Parallel Growth Models

We next tested a series of two-variable parallel growth models. The first of these included victims’ and offenders’ value consensus perceptions, both represented by their latent intercepts and latent slopes. We included paths from the intercepts to slopes, both within victim and offender roles and, of particular interest, between victim and offender. The latter represent effects of victims’ initial perceptions of value consensus on change in offenders’ value consensus perceptions and, vice versa, effects of offenders’ initial perceptions of value consensus on change in victims’ value consensus perceptions. As covariates, victim gender and victim-perceived seriousness of wrongdoing were modeled as being correlated with the intercept of victim-perceived value consensus, and offender gender and offender-perceived seriousness were modeled as being correlated with the intercept of offender-perceived value consensus. In addition, because it was plausible that value consensus perceptions would be more slowly restored following perceived serious wrongdoing, paths from seriousness perceptions on value consensus slopes were included within victim and offender roles. A second model paired victim forgiveness and offender genuine self-forgiveness in the same way. A third model paired victim forgiveness and offender self-punitiveness. The results are summarized in Table 4.

Table 4.

Structural Paths and Covariances of Two-Variable Latent Growth Models.

Structural relationships			Coefficient	SE	p
Model 1: victim and offender-perceived value consensus
Structural paths:
Value C.-V (S)	←	Value C.-V (I)	−0.35	0.10	<.001
	←	Value C.-O (I)	0.11	0.08	.211
	←	Seriousness-V	0.13	0.06	.038
Value C.-O (S)	←	Value C.-V (I)	0.20	0.06	<.001
	←	Value C.-O (I)	−0.27	0.12	.024
	←	Seriousness-O	−0.06	0.04	.142
Covariances:
Value C.-V (I)	↔	Seriousness-V	−0.39	0.12	<.001
Value C.-V (I)	↔	Gender-V	−0.01	0.02	.528
Seriousness-V	↔	Seriousness-O	0.62	0.15	<.001
Gender-V	↔	Gender-O	−0.17	0.02	<.001
Value C.-O (I)	↔	Seriousness-O	0.16	0.12	.158
	↔	Gender-O	0.02	0.02	.335
Model fit: χ²(31) = 55.43, p = .004, χ²/df = 1.79; CFI = .966; RMSEA = .068 (.037, .096)
Model 2: forgiveness/genuine self-forgiveness
Structural paths:
Forgiveness-V (S)	←	Forgiveness-V (I)	−0.11	0.14	.455
	←	Genuine SF.-O (I)	−0.03	0.07	.612
	←	Seriousness-V	0.19	0.06	.003
Genuine SF.-O (S)	←	Forgiveness-V (I)	−0.34	0.12	.005
	←	Genuine SF.-O (I)	0.11	0.20	.583
	←	Seriousness-O	−0.13	.10	.210
Covariances:
Forgiveness-V (I)	↔	Seriousness-V	−0.43	0.09	<.001
Forgiveness-V (I)	↔	Gender-V	−0.01	0.02	.357
Seriousness-V	↔	Seriousness-O	0.41	0.12	<.001
Gender-V	↔	Gender-O	−0.17	0.02	<.001
Genuine SF.-O (I)	↔	Seriousness-O	0.83	0.14	<.001
Genuine SF.-O (I)	↔	Gender-O	−0.01	0.02	.444
Model fit: χ²(31) = 49.78, p = .018, χ²/df = 1.60; CFI = .977; RMSEA = .060 (.025, .089)
Model 3: forgiveness/self-punitiveness
Structural paths:
Forgiveness-V (S)	←	Forgiveness-V (I)	−0.17	0.14	.219
	←	Self-Punit.-O (I)	−0.13	0.06	.037
	←	Seriousness-V	0.20	0.06	.001
Self-Punit.-O (S)	←	Forgiveness-V (I)	−0.26	0.09	.004
	←	Self-Punit.-O (I)	−0.03	0.17	.869
	←	Seriousness-O	−0.11	0.09	.227
Covariances:
Forgiveness-V (I)	↔	Seriousness-V	−0.48	0.10	<.001
Forgiveness-V (I)	↔	Gender-V	−0.02	0.02	.290
Seriousness-V	↔	Seriousness-O	0.34	0.11	.003
Gender-V	↔	Gender-O	−0.17	0.02	<.001
Self-Punit.-O (I)	↔	Seriousness-O	0.88	0.14	<.001
Self-Punit.-O (I)	↔	Gender-O	v0.02	0.02	.220
Model fit: χ²(31) = 46.86, p = .034, χ²/df = 1.51; CFI = .981; RMSEA = .055 (.016, .085)
Model 4: forgiveness/genuine self-forgiveness residualized
Structural paths:
Forgiveness-V (S)	←	Forgiveness-V (I)	−0.18	0.13	.172
	←	Genuine SF.-O (I)	0.06	0.09	.492
	←	Seriousness-V	0.16	0.06	.007
Genuine SF.-O (S)	←	Forgiveness-V (I)	−0.14	0.11	.213
	←	Genuine SF.-O (I)	0.06	0.22	.794
	←	Seriousness-O	−0.02	0.07	.820
Covariances:
Forgiveness-V (I)	↔	Seriousness-V	−0.41	0.09	<.001
Forgiveness-V (I)	↔	Gender-V	−0.01	0.02	.416
Seriousness-V	↔	Seriousness-O	0.52	0.14	<.001
Gender-V	↔	Gender-O	−0.17	0.02	<.001
Genuine SF.-O (I)	↔	Seriousness-O	0.33	0.11	.002
	↔	Gender-O	−0.002	0.02	.906
Model fit: χ²(31) = 63.55, p = .001, χ²/df = 2.05; CFI = .955; RMSEA = .078 (.051, .106)
Model 5: forgiveness/self-punitiveness residualized
Structural paths:
Forgiveness-V (S)	←	Forgiveness-V (I)	−0.25	0.13	.051
	←	Self-Punit.-O (I)	−0.21	0.09	.018
	←	Seriousness-V	0.18	0.06	.003
Self-Punit.-O (S)	←	Forgiveness-V (I)	−0.15	0.08	.063
	←	Self-Punit.-O (I)	−0.03	0.19	.862
	←	Seriousness-O	−0.01	0.06	.897
Covariances:
Forgiveness-V (I)	↔	Seriousness-V	−0.46	0.10	<.001
Forgiveness-V (I)	↔	Gender-V	−0.01	0.02	.360
Seriousness-V	↔	Seriousness-O	0.45	0.13	<.001
Gender-V	↔	Gender-O	−0.17	0.02	<.001
Genuine SF.-O (I)	↔	Seriousness-O	0.37	0.11	<.001
Genuine SF.-O (I)	↔	Gender-O	−0.01	0.02	.396
Model fit: χ2(31) = 57.46, p = .003, χ²/df = 1.85; CFI = .964; RMSEA = .071 (.041, .099)

Note. Value C. = value consensus; Genuine SF = genuine self-forgiveness; Self-Punit. = self-punitiveness; V = victim; O = offender; (I) = intercept; (S) = slope; CFI = confirmatory fit index; RMSEA = root mean square error of approximation.

We assessed the fit of our models with three different fit indices: A χ²/df ratio of two or less represents a good fit, and between two and three an acceptable fit; a confirmatory fit index (CFI) of .95 or higher represents a good fit and between .90 to .95 a marginal fit; and a root mean square error of approximation (RMSEA) of .06 or lower represents a good fit, between .06 to .08 an acceptable fit, and between .08 and .10 a marginal fit (Hair et al., 2010; Tabachnick & Fidell, 2019). By these criteria, the three two-variable parallel growth models fit the data well (see Table 4). In terms of the structural paths, the effects of intercept on the slope of the same variable (i.e., within victim or offender roles) are of lesser interest. These were significant only in the perceived value consensus model; for both victim-perceived and offender-perceived value consensus the intercept showed a negative effect on the slope. It is possible that a low perceived value consensus at Time 1 provides greater scope for increase over time.

Of main interest are the intercept-to-slope effects between partners. In Model 1, the intercept (initial status) of victim-perceived value consensus had a significant positive effect on the slope (growth) in offender-perceived value consensus, whereas the effect of the intercept of offender-perceived value consensus on the slope of victim-perceived value consensus was not significant. Victims’ belief in shared value consensus was related to the development of such perceptions in offenders, but not vice versa. In Model 2, the intercept of victim forgiveness had a significant negative effect on the slope of offender genuine self-forgiveness, whereas the effect of the intercept of genuine self-forgiveness on the slope of forgiveness was not significant. A victim’s early forgiveness was related to an offender’s diminishing engagement in genuine self-forgiveness. In Model 3, the intercept of victim forgiveness had a significant negative effect on the slope of offender self-punitiveness; and the intercept of offenders’ self-punitiveness had a negative effect on the slope of forgiveness. Victims’ forgiveness was related to a greater decrease in offenders’ self-punitiveness over time; conversely, offenders’ self-punitiveness was related to a decrease in victims’ forgiveness over time.

As discussed in the introduction, genuine self-forgiveness and self-punitiveness, despite their opposite relations to an overarching concept of self-forgiveness, have in common the offender’s acceptance of responsibility and guilt. Indeed, as can be seen in Table 2, the two variables were substantially positively correlated at each time-point (rs = .50–.59). To investigate their distinctive qualities, we therefore conducted two further models, in which we partialled out the overlapping variance. So as not to add further complexity to the models (in particular when moving to the four-variable models in the next section), we refrained from modeling both concepts within the same growth model but instead residualized their indicator variables. That is, at each time-point we regressed the measure of genuine self-forgiveness on self-punitiveness and saved the residual as a distinct measure of genuine self-forgiveness; and vice versa for the measures of self-punitiveness. We then conducted our latent growth models with these residualized indicators.

The model fit deteriorated somewhat compared with the previous models but remained acceptable (see Table 4). For the residualized measure of genuine self-forgiveness, the previously significant negative effect of the forgiveness intercept on the slope of genuine self-forgiveness was no longer significant. In contrast, for the residualized measure of self-punitiveness, both previously significant intercept-to-slope partner effects were again significant or marginally significant: The self-punitiveness intercept was related to a decline in forgiveness, and the forgiveness intercept was (marginally) related to a decline in self-punitiveness.

Overall Four-Variable Parallel Growth Models

In the final step of our analyses, we combined the two-variable models to four-variable parallel growth models (see Figure 2): victims’ perceived value consensus and forgiveness and offenders’ perceived value consensus and either genuine self-forgiveness or self-punitiveness. To highlight the distinctive aspects of genuine self-forgiveness and self-punitiveness, we again used a regression approach to partial out the variance shared with the respective other variable and treated the residualized measures as indicators within these models. (We report the equivalent analyses with the raw, non-residualized measures in the Online Supplementary Materials, and merely note here that the results were highly similar.)

The latent intercepts were modeled to be correlated within each role. That is, the two intercepts for victim-perceived value consensus and forgiveness were allowed to correlate, as were the intercepts of the offender-perceived value consensus and self-forgiveness variables. The same applied also to the covariates: Victim gender and seriousness ratings were allowed to correlate with the intercepts for victim-perceived value consensus and forgiveness, and offender gender and seriousness rating were allowed to correlate with intercepts for offender-perceived value consensus and self-forgiveness. Again, given that more serious wrongdoing might be more difficult (and thus slower) to repair, we included effects of victim-perceived seriousness on the slopes of victim-perceived value consensus and forgiveness, and effects of offender-perceived seriousness on the slopes of offender-perceived value consensus and self-forgiveness. Victim- and offender-perceived seriousness were free to correlate, and so were victim and offender gender. Most importantly, the tested models included all structural effects from each of the four latent intercepts (victim-perceived value consensus, forgiveness, offender-perceived value consensus, and self-forgiveness) to all four latent slopes, that is, within and between partners, thus permitting a test of the intra- and inter-role dynamics in their entirety as per our theoretical model (see Figure 1).

Model 6: Including genuine self-forgiveness

We first focus on the model that included genuine self-forgiveness of the offender as an aspect of self-forgiveness (with self-punitiveness partialled out—Model 6). The model fit statistics and estimated structural parameters are shown in Table 5. The fit was more marginal than for the previous simpler models. We inspected modification indices for possible reasons and found that two error covariances were the two largest contributors to the suboptimal fit: When freeing the error covariance between forgiveness at Time 3 and offender value consensus at Time 3, and the error covariance between the forgiveness indicators at Times 1 and 2, the fit would significantly improve. However, we could not see any theoretical argument for including these covariances in the model. Comparing the structural effects for the models with and without these error covariances furthermore showed very similar results (for details, see Online Supplementary Materials). Given that its fit statistics were still tolerable, we report here the theoretically specified model, without the error covariances.

Table 5.

Model 6: Structural Paths and Covariances of Four-Variable Latent Growth Models With Genuine Self-Forgiveness (Self-Punitiveness Partialled Out).

Structural relationships			Coefficient	SE	p
Structural paths:
Forgiveness-V (S)	←	Forgiveness-V (I)	0.82	0.33	.011
	←	Value C.-V (I)	−0.47	0.17	.005
	←	Seriousness-V	0.31	0.08	<.001
	←	Value C.-O (I)	−0.001	0.09	.994
	←	Genuine SF.-O (I)	−0.10	0.10	.322
Value C.-V (S)	←	Forgiveness-V (I)	0.83	0.16	<.001
	←	Value C.-V (I)	−0.69	0.10	<.001
	←	Seriousness-V	0.27	0.06	<.001
	←	Value C.-O (I)	−0.01	0.10	.907
	←	Genuine SF.-O (I)	0.20	0.12	.079
Value C.-O (S)	←	Forgiveness-V (I)	0.03	0.12	.839
	←	Value C.-V (I)	0.20	0.08	.012
	←	Value C.-O (I)	−0.11	0.18	.543
	←	Genuine SF.-O (I)	−0.22	0.14	.125
	←	Seriousness-O	−0.03	0.05	.480
Genuine SF.-O (S)	←	Forgiveness-V (I)	−0.50	0.17	.004
	←	Value C.-V (I)	0.33	0.11	.004
	←	Value C.-O (I)	−0.02	0.13	.860
	←	Genuine SF.-O (I)	−0.07	0.22	.761
	←	Seriousness-O	−0.02	0.06	.779
Covariances:
Forgiveness-V (I)	↔	Value C.-V (I)	0.57	0.10	<.001
	↔	Seriousness-V	−0.41	0.09	<.001
	↔	Gender-V	−0.02	0.01	.080
Value C.-V (I)	↔	Seriousness-V	−0.38	0.12	.001
Value C.-V (I)	↔	Gender-V	−0.01	0.02	.743
Seriousness-V	↔	Seriousness-O	0.47	0.13	<.001
Gender-V	↔	Gender-O	−0.17	0.02	<.001
Value C.-O (I)	↔	Seriousness-O	0.13	0.12	.262
Value C.-O (I)	↔	Gender-O	0.02	0.02	.346
Genuine SF.-O (I)	↔	Value C.-O (I)	0.36	0.08	<.001
	↔	Seriousness-O	0.32	0.11	.003
	↔	Gender-O	−0.002	0.02	.927
Model fit: χ²(84) = 195.07, p < .001, χ²/df = 2.32; CFI = .919; RMSEA = .088 (.072, .104)

Note. Value C. = value consensus; Genuine SF = genuine self-forgiveness; V = victim; O = offender; (I) = intercept; (S) = slope; CFI = confirmatory fit index; RMSEA = root mean square error of approximation.

Inspecting the structural covariances, we note that the intercepts of forgiveness and victim-perceived value consensus were positively correlated, as well as the intercepts of genuine self-forgiveness and offender-perceived value consensus (Table 5, lower panel). Hence, victims’ initial levels of forgiveness and offenders’ initial level of genuine self-forgiveness were positively related to the value consensus each perceived. This is consistent with theoretical assumptions of within-role relationships of value consensus perceptions with forgiveness and genuine self-forgiveness, respectively.

Intrapersonal dynamics

For victims, the model showed a positive effect of the forgiveness intercept on the slope of victim-perceived value consensus. Initial levels of forgiveness were positively related to the development of victims’ value consensus perceptions (in line with predicted path a₁; see Figure 3A). Conversely, victim-perceived value consensus (intercept) had a negative effect on forgiveness (slope); initial perceived value consensus was negatively related to growth in forgiveness (contrary to predicted path a₂). For offenders, although we expected that engagement in genuine self-forgiveness would similarly promote their value consensus perceptions, the intercept (initial status) of the offenders’ self-forgiveness had no significant effect on the slope of offender-perceived value consensus, and neither did initial levels of offender-perceived value consensus affect the growth rate of genuine self-forgiveness.

Figure 3.

Conceptual summary of results.

Interpersonal dynamics

For effects between partners, victims’ value consensus perceptions (intercept) had a positive effect on offenders’ value consensus perceptions (slope) as well as their genuine self-forgiveness (slope). The converse effects of offender-perceived value consensus (intercept) on victim-perceived value consensus (slope) and their forgiveness (slope) were not significant. These results are consistent with the earlier two-variable models. They indicate that victims’ initial belief in a shared value consensus is positively related with the development of such a perception in offenders (in line with path c₁) and also with their growing engagement in genuine self-forgiveness (in line with path d₁). In contrast, offenders’ value consensus does not seem to be related to victims’ growth in value consensus.

Furthermore, offender genuine self-forgiveness (intercept) had no significant effect on victims’ forgiveness (slope); offenders’ early engagement in genuine self-forgiveness does not seem to be related to change in victim’s forgiveness. Unexpectedly, victim’s forgiveness (intercept) had a negative effect on offenders’ genuine self-forgiveness (slope); initial levels of forgiveness seem to be related to a decrease in offenders’ engagement in genuine self-forgiveness (contrary to predicted path e₁). Again, these results are completely consistent with the two-variable models.

Covariates

Victim and offender gender did not show any significant relationships with the latent intercepts. Victim-perceived and offender-perceived seriousness were positively correlated. Victim-perceived seriousness was negatively related to the initial status of victim-perceived value consensus and forgiveness. A serious wrongdoing implies greater doubts about the offender sharing the same values and is less forgivable. Offender-perceived seriousness was not significantly related to the initial status of offender-perceived value consensus but was significantly positively related to that of genuine self-forgiveness, which may reflect the offenders’ acknowledgment of (and working through) their wrongdoing in genuine self-forgiveness. Moreover, victim-perceived seriousness was surprisingly a positive predictor of growth in victim-perceived value consensus and forgiveness, whereas offender-perceived seriousness was not predictive of growth in offender-perceived value consensus or genuine self-forgiveness.

Model 7: Including self-punitiveness

A second full model was tested with self-punitiveness in place of genuine self-forgiveness (with genuine self-forgiveness partialled out—Model 7). The model fit statistics and estimated structural parameters are shown in Table 6. The fit was again somewhat weaker than for the two-variable models, albeit acceptable. The same two error covariances as in Model 6 contributed mostly to the weaker fit. However, as with the previous model, we did not see good reasons to include these covariances and instead report here the theoretically specified model. In any case, the results for the models with or without these error covariances were highly similar (see Online Supplementary Materials).

Table 6.

Model 7: Structural Paths and Covariances of Four-Variable Latent Growth Models With Self-Punitiveness (Genuine Self-Forgiveness Partialled Out).

Structural relationships			Coeff.	SE	p
Structural paths:
Forgiveness-V (S)	←	Forgiveness-V (I)	1.38	0.58	.018
	←	Value C.-V (I)	−0.75	0.30	.013
	←	Seriousness-V	0.41	0.12	<.001
	←	Value C.-O (I)	−0.01	0.08	.853
	←	Self-Punit.-O (I)	−0.24	0.10	.014
Value C.-V (S)	←	Forgiveness-V (I)	0.96	0.19	<.001
	←	Value C.-V (I)	−0.72	0.12	<.001
	←	Seriousness-V	0.30	0.07	<.001
	←	Value C.-O (I)	0.06	0.09	.450
	←	Self-Punit.-O (I)	−0.05	0.10	.624
Value C.-O (S)	←	Forgiveness-V (I)	0.01	0.13	.951
	←	Value C.-V (I)	0.17	0.08	.036
	←	Value C.-O (I)	−0.28	0.13	.031
	←	Self-Punit.-O (I)	−0.10	0.11	.339
	←	Seriousness-O	−0.04	0.05	.426
Self-Punit.-O (S)	←	Forgiveness-V (I)	0.04	0.15	.785
	←	Value C.-V (I)	−0.14	0.09	.124
	←	Value C.-O (I)	0.08	0.10	.464
	←	Self-Punit.-O (I)	0.05	0.22	.831
	←	Seriousness-O	−0.02	0.07	.719
Covariances:
Forgiveness-V (I)	↔	Value C.-V (I)	0.58	0.10	<.001
	↔	Seriousness-V	−0.44	0.09	<.001
	↔	Gender-V	−0.02	0.01	.048
Value C.-V (I)	↔	Seriousness-V	−0.43	0.12	<.001
Value C.-V (I)	↔	Gender-V	−0.01	0.02	.548
Seriousness-V	↔	Seriousness-O	0.39	0.12	.001
Gender-V	↔	Gender-O	−0.17	0.02	<.001
Value C.-O (I)	↔	Seriousness-O	0.14	0.12	.238
Value C.-O (I)	↔	Gender-O	0.02	0.02	.274
Self-Punit.-O (I)	↔	Value C.-O (I)	−0.21	0.08	.010
	↔	Seriousness-O	0.38	0.11	<.001
	↔	Gender-O	−0.02	0.02	.327

Note. Value C. = value consensus; Self-Punit. = self-punitiveness; V = victim; O = offender; (I) = intercept; (S) = slope.

The structural covariances of the main variables (see Table 6, bottom panel) showed that the intercepts of forgiveness and victim-perceived value consensus were positively correlated, as in Model 6. The intercepts of self-punitiveness and offender-perceived value consensus were significantly negatively correlated, in contrast to genuine self-forgiveness in Model 6, which showed a positive correlation. The realization of sharing a value consensus appears to be implicated in genuine self-forgiveness, but the realization of not sharing a consensus, or violating it, is related to self-punitiveness.

Intrapersonal dynamics

The forgiveness intercept had a significant positive effect on victim-perceived value consensus (slope). Conversely, victim-perceived value consensus (intercept) had a negative effect on forgiveness (slope). These results are consistent with Model 6 (paths a₁ and a₂, Figure 3B). The intercept of self-punitiveness showed no significant effect on the slope of offender-perceived value consensus; neither did the intercept of offender-perceived value consensus show an effect on the slope of offenders’ self-punitiveness.

Interpersonal dynamics

Regarding cross-partner effects, victim’s perceived value consensus (intercept) had a positive effect on offenders’ value consensus perceptions (slope), consistent with Model 6 (path c₁). However, in contrast to its positive effect on offenders’ genuine self-forgiveness in Model 6, victims-perceived value consensus (intercept) did not have a significant effect on offenders’ self-punitiveness (slope). These results indicate again that victims’ early belief in a shared value consensus is positively related to the development of such a perception in offenders (but not the other way around); however, it does not seem to be related to offenders overcoming self-punitiveness over time.

Contrary to Model 6 where victims’ forgiveness (intercept) negatively affected the slope of offenders’ genuine self-forgiveness, victims’ forgiveness (intercept) had no significant effect on self-punitiveness (slope). Instead, conversely, self-punitiveness (intercept) had a significant negative effect on the slope of victims’ forgiveness. This finding indicates that offenders’ initial levels of self-punitiveness are negatively related to the development of forgiving sentiments in victims (path e₂).

Covariates

Victim gender was negatively correlated with the forgiveness intercept; men were less forgiving than women—this effect was marginally significant in Model 6. As in Model 6, victims’ perceived seriousness was negatively related to the initial status of victims’ perceived value consensus and forgiveness. Offenders’ perceived seriousness was not significantly related to the initial status of offenders’ perceived value consensus, but it was significantly positively related to that of self-punitiveness—as it was to genuine self-forgiveness in Model 6. As in Model 6, victims’ perceived seriousness was a positive predictor of growth in victims’ perceived value consensus and forgiveness.

Discussion

The present study provides support for our notion that forgiveness and self-forgiveness are dynamic and interdependent processes. Previous research has largely studied forgiveness and self-forgiveness as intrapersonal processes or, at best, as individual responses to partner behaviors (e.g., an offender’s apology, or the victim’s rejection of an apology). While it can be methodologically fruitful to isolate casual effects, it is critical to appreciate and understand the processes of moral repair as an interactional dynamic whereby two parties rebuild their value consensus interpersonally, in interaction with each other. Using a prospective-longitudinal-dyadic design, we explored the dynamic relationships between forgiveness, self-forgiveness, and value consensus perceptions in dyads. The longitudinal data permitted a latent growth approach to model the effects of initial status of variables on the growth rates of other variables. While there are limitations to this approach that we discuss below, the findings underscore that perceived value consensus is a key pivot in moral repair dynamics.

Intrapersonal Dynamics

In terms of intrapersonal processes, the findings suggest that victims’ forgiveness increases their perception of sharing a value consensus with the offender. This is in line with research by Wenzel and Okimoto (2010, 2012) where an act of forgiveness was shown to feed back into the forgiver’s own value consensus perceptions and sentiments toward the offender. The present study, using a different paradigm that taps into naturalistic processes as they unfold, adds further evidence that forgiveness is not a mere outcome of moral repair but rather an active contributor to the process. Forgiveness may function as a self-persuasion that reassures the victim of the offender’s commitment to shared values. It implies, and thus reinforces, trust that the offender is committed to the violated values.

Although we expected that perceived value consensus would relate to increases in victims’ forgiveness (Wenzel & Okimoto, 2010) by seeing the offender as trustworthy to be given another chance, this effect was not supported in the present study. In fact, the results indicated the opposite effect: Value consensus perceptions seemed to be related to a decrease in forgiveness. It is not clear why this would be the case. It is possible that the perception of sharing a value consensus with the offender makes further self-persuasion in the form of forgiveness unnecessary. Alternatively, victims may feel inclined to judge offenders more harshly (and unforgivingly) when they believe offenders knowingly violated principles despite understanding that these are important to the relationship or shared identity.

On the offenders’ side, the findings did not evidence any corresponding intrapersonal dynamics. Specifically, the present study did not indicate an (intercept-on-slope) effect for offender-perceived value consensus on genuine self-forgiveness. These findings appear at odds with previous evidence where experimentally manipulated value affirmation causally affected and promoted genuine self-forgiveness (Woodyatt et al., 2017; Woodyatt & Wenzel, 2014). However, those experimental manipulations involved instructions that led participants through an exercise of value affirmation. Hence, as an intervention, it may have had a more active impact on the offender’s processing of their wrongdoing, whereas the present study’s measure of value consensus represented a more passive gauge of the offender’s processing.

Still, value consensus and genuine self-forgiveness were positively correlated at the intercept-level, whereas self-punitiveness was negatively related to offender-perceived value consensus. Interestingly, offender-perceived seriousness of the offense was positively related to the intercept of both genuine self-forgiveness and self-punitiveness. Hence, both genuine self-forgiveness and self-punitiveness seem to involve an acceptance of the wrongfulness of one’s actions; however, genuine self-forgiveness further seems to imply a trust in one’s moral integrity, whereas such self-trust is lacking for self-punitiveness (see also Woodyatt & Wenzel, 2013a).

On the other hand, the finding that victim-perceived seriousness was positively related to the development of value consensus and forgiveness in victims is curious, considering that seriousness has generally been found to be negative related to forgiveness (Fehr et al., 2010). It is possible that a more serious wrongdoing initially implies lower levels of perceived value consensus and forgiveness, permitting more scope for growth in those variables. However, this should have been statistically accounted for by the effects of the intercepts of value consensus and forgiveness, yet the unexpected effects emerged regardless. Alternatively, it is possible that victims’ appraisals of seriousness partly reflect how important they find their relationship with the offender and how committed they are to it, which at the same time may make them more willing to engage in forgiveness (Finkel et al., 2002).

In sum, the intrapersonal dynamics revealed in our research suggest that victims are more clearly able to regulate their own value consensus perceptions through acts and sentiments of forgiveness. In contrast, offenders seem less able to do so and, as the following discussion of interpersonal dynamics shows, may be more reliant on victims to assure them of their value consensus.

Interpersonal Dynamics

The most exciting and distinctive aspect of the present research is that it facilitates recognition of the dynamics that occur between partners in their negotiation of moral repair. First, victims’ perceptions of value consensus showed a reliable effect (across two- and four-variable models) on the growth of value consensus perceptions in offenders. The victim’s perception that both parties share the same values, whether explicitly or subtly communicated to the partner, might reassure the offender of a continued value consensus, in spite of what the offender has done. Intrapersonally, victims’ forgiveness seems to nurture their belief in a value consensus with the offender (as discussed above); it can therefore be concluded that forgiveness also ultimately benefits offenders’ value consensus perceptions, mediated via the victims’ perceptions. Thus, forgiveness may not only lead offenders to reciprocate with conciliation out of a sense of indebtedness (Kelln & Ellard, 1999) or gratitude (Mooney et al., 2016). Rather, forgiveness also fosters offenders’ belief in shared values and, as sharing values increases feelings of acceptance, closeness, and love (Lomore et al., 2007), should therefore motivate offender efforts to repair the relationship.

Related, this study showed that victims’ perceptions of value consensus also reliably led to an increase in offenders’ engagement in genuine self-forgiveness. Hence, victim beliefs that both parties share a consensus about relevant values seems to encourage offenders to genuinely work through their wrongdoing: making an effort to learn from what they have done to change and rebuild their moral self. Notably, victims’ value consensus perception did not decrease offenders’ self-punitiveness. Thus, while previous research showed that offenders, when led through exercises of value reaffirmation, become more willing to engage in genuine self-forgiveness (Woodyatt & Wenzel, 2014), the present study showed that victims’ belief in value consensus, and the assurance this may communicate to offenders, can similarly encourage genuine self-forgiveness. As before, this also implies an indirect effect of forgiveness, mediated by victim-perceived value consensus: To the extent that forgiveness bolsters victims’ confidence that both parties share relevant values, it leads offenders to engage in genuine self-forgiveness. In other words, to the extent that the forgiveness is genuine and implies a true belief in the offenders’ fundamental goodness, it leads offenders to be genuine in working through their wrongdoing.

Two of the interpersonal dynamics were unexpected. First, victims’ forgiveness decreased offenders’ genuine self-forgiveness. We expected the opposite: simple reciprocation where victim conciliation is answered by offenders working through their wrongdoing. It is possible that the observed effect is an outcome of suppression due to controlling for the mediated path of the effect of forgiveness via victim-perceived value consensus, thus creating an opposite direct effect. In other words, genuine forgiveness implies the victim’s true belief that the offender shares relevant values with them, which leads the offender to respond with genuine self-forgiveness; however, forgiveness without such belief (which may be seen as an empty gesture or even a manipulative move) may prompt offender reactance and unwillingness to engage with their wrongdoing. Alternatively, it is possible that victims’ early forgiveness is seen by the offender as relieving them from efforts of working through. This would be consistent with McNulty’s (2011) finding that forgiveness can have a “dark side” in that it can free offenders from critical self-reflection and allow them to continue on with their hurtful behavior. This also raises questions about the role of the timing of forgiveness in moral repair—an issue that has received little research attention so far.

The second surprising interpersonal effect in the present study was that self-punitiveness negatively affected the development of forgiveness over time (observed in both the two and four-variable models). We did not have specific a priori predictions for the model that included self-punitiveness. However, past research suggests that offender self-punitiveness would be responded to with victim forgiveness, be it for feeling sympathy for the offender, seeing the self-punitiveness as a sign of remorse, believing it effectively restored justice, or for communicating the sharing of a value consensus with the victims (De Vel-Palumbo et al., 2019). Contrary to these arguments, the observed effect suggests that offenders’ self-punitiveness reduces victims’ forgiveness. It may be that self-punitiveness is considered an egocentric response where offenders feel bad about themselves rather than feeling for or with the victim. There is, indeed, evidence that self-punitiveness is related to offenders’ avoidance of the victim/issue and, over time, with a reduction of empathy for the victim (Woodyatt & Wenzel, 2013a). It is therefore possible that victims react with decreasing benevolence and forgiveness to an offenders’ self-punitiveness.

The main insight from these analyses, however, relates to the pivotal role that victims’ value consensus perceptions have in the process of moral repair. Victims’ belief in such a consensus seems to positively influence offenders to regain their own confidence in sharing a value consensus with victims and encourages them to engage in genuine self-forgiveness, working through their wrongdoing, and learning their lessons from it to rebuild their moral self and social relationships. Moreover, victims seem able to actively rebuild value consensus perceptions by engaging in forgiveness. Hence, forgiveness seems capable of instigating or promoting value consensus processes deemed beneficial for moral repair. At the same time, however, there remains a risk that forgiveness prematurely relieves the offender from engaging in efforts of working through their wrongdoing and thus undermines genuine repair.

Limitations and Research Directions

The latent growth approach we have used can be criticized for not completely disaggregating between- and within-person sources of variation (e.g., Curran & Bauer, 2011; Curran et al., 2014). The approach focuses on between-person differences in level and change of psychological variables, and associations between those. For example, individuals who differ in the level of one variable (e.g., forgiveness) may show different trajectories in another variable (e.g., more or less increase in value consensus perceptions over time). However, from this we need to distinguish, and cannot infer, time-specific within-person change (e.g., whenever an individual forgives their perception of value consensus increases). Ultimately, it depends on the theory whether between- or within-person effects, or both, are of interest and theoretically relevant (Curran & Bauer, 2011). Our theoretical arguments are not specific enough, and arguably both between and within-person associations are theoretically relevant. However, we limited our analysis to the between-person level because of the already considerable complexity posed by the parallel growth modeling of four variables across the two partners of a relationship dyad. Future research may adopt alternative growth curve modeling approaches that allow the investigation of within-person associations between variables over time (e.g., Curran et al., 2014). Certainly, this is an evolving area of research and we need to continue to evolve our means of more accurately testing the complexity of these relationship dynamics.

In addition, as the research is correlational in nature, future research will need to use experimental designs to investigate causality, even if this means again carving causal effects out of what is a more complex dynamic of reciprocal relationships. The present research demonstrates the complexity and reciprocity inherent in this dynamic, and thus provides a heuristic perspective for future research to attempt to isolate specific causal effects.

Furthermore, due to the prospective design with “ordinary” relationship couples, the reported transgressions were generally only rather minor relationship infractions. Hence, it is not clear whether the present findings generalize to more serious incidents. To study more serious transgressions with this type of design would require expanding the timeframe of the study considerably, likely over years. Alternatively, one would need to recruit high-risk groups (e.g., relationship partners with a history of abuse). However, such a study would raise major ethical questions (e.g., letting a serious transgression take its course without intervening) and would suffer restricted generalization of findings to specific at-risk groups or transgression types. Another limitation of the present method is that interpretation of the findings presumes that both parties somehow engaged with each other (e.g., signaling one’s trust that the offender shares the same values). However, the extent and nature of the engagement were uncontrolled and unobserved, and we therefore cannot be sure what role engagement, and certain qualities thereof, played in these processes.

Finally, the present study focused on relationship couples as dyads, although our theoretical analysis is not limited to relationship partners. All transgressions, whether between close others or strangers, can be understood as a violation of values presumed to be shared within some form of moral community, however defined. The restoration of consensus about these values would be critical to restorative justice and moral repair (Walker, 2006; Wenzel et al., 2008), although this would be particularly the case for parties highly identified with that shared community (Wenzel et al., 2010; Wenzel & Okimoto, 2012) whose identity is particularly tied to the shared values, as would be the case for couples. However, future research will nonetheless need to test the forgiveness/self-forgiveness dynamics in other interpersonal (and intergroup) contexts.

Conclusion

Moving beyond the one-party perspective and the intrapersonal investigations of prior (self-)forgiveness research, the present findings provide support for the moral repair processes as being intertwined between victim and offender. They suggest that forgiveness and self-forgiveness are interactional and negotiated. Finally, they highlight the centrality of the reaffirmation of value consensus as a vital element in these processes.

Supplemental Material

Wenzel_Online_appendix – Supplemental material for Dynamics of Moral Repair: Forgiveness, Self-Forgiveness, and the Restoration of Value Consensus as Interdependent Processes

Supplemental material, Wenzel_Online_appendix for Dynamics of Moral Repair: Forgiveness, Self-Forgiveness, and the Restoration of Value Consensus as Interdependent Processes by Michael Wenzel, Lydia Woodyatt, Tyler G. Okimoto and Everett L. Worthington in Personality and Social Psychology Bulletin

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by a Grant from the Australian Research Council, DP190102283.

ORCID iD

Michael Wenzel

Supplemental Material

Supplemental material is available online with this article.

Notes

References

Baumeister

R. F.

Exline

J. J.

Sommer

K. L.

(1998). The victim role, grudge theory, and two dimensions of forgiveness. In Worthington, Jr.

E. L.

(Ed.), Dimensions of forgiveness (pp. 79–104). Templeton Press.

Bollen

K. A.

Curran

P. J.

(2006). Latent curve models: A structural equation approach. Wiley.

Carpenter

T. P.

Carlisle

R. D.

Tsang

J.-A.

(2014). Tipping the scales: Conciliatory behavior and the morality of self-forgiveness. The Journal of Positive Psychology, 9, 389–401.

Curran

P. J.

Bauer

D. J.

(2011). The disaggregation of within-person and between-person effects in longitudinal models of change. Annual Review of Psychology, 62, 583–619.

Curran

P. J.

Howard

Bainter

Lane

McGinley

(2014). The separation of between-person and within-person components of individual change over time: A latent curve model with structured residuals. Journal of Consulting and Clinical Psychology, 82, 879–894.

Curran

P. J.

Obeidat

Losardo

(2010). Twelve frequently asked questions about growth curve modeling. Journal of Cognition and Development, 11, 121–136.

Davis

D. E.

M. Y.

Griffin

B. J.

Bell

Hook

J. N.

Van Tongeren

D. R.

Westbrook

C. J.

(2015). Forgiving the self and physical and mental health correlates: A meta-analytic review. Journal of Counseling Psychology, 62, 329–335.

De Vel-Palumbo

Wenzel

Woodyatt

. (2019). Self-punishment promotes reconciliation with third parties by addressing the symbolic implications of wrongdoing. European Journal of Social Psychology, 49, 1070–1086.

De Vel-Palumbo

Woodyatt

Wenzel

. (2018). Why do we self-punish? Perceptions of the motives and impact of self-punishment outside the laboratory. European Journal of Social Psychology, 48, 756–768.

10.

Enright

R. D.

, & The Human Development Study Group. (1991). The moral development of forgiveness. In Kurtines

Gewirtz

(Eds.), Handbook of moral behavior and development (Vol. 1, pp. 123–152). Lawrence Erlbaum.

11.

Fehr

Gelfand

M. J.

(2010). When apologies work: How matching apology components to victims’ self-construals facilitates forgiveness. Organizational Behaviorand Human Decision Processes, 113, 37–50.

12.

Fehr

Gelfand

M. J.

Nag

(2010). The road to forgiveness: A meta-analytic synthesis of its situational and dispositional correlates. Psychological Bulletin, 136, 894–914.

13.

Finkel

E. J.

Rusbult

C. E.

Kumashiro

Hannon

P. A.

(2002). Dealing with betrayal in close relationships: Does commitment promote forgiveness? Journal of Personality and Social Psychology, 82, 956–974.

14.

Fisher

M. L.

Exline

J. J.

(2010). Moving toward self-forgiveness: Removing barriers related to shame, guilt, and regret. Social and Personality Psychology Compass, 4(8), 548–558.

15.

Forster

D. E.

Billingsley

Russell

V. M.

McCauley

T. G.

Smith

Burnette

J. L.

McCullough

M. E.

(2019). Forgiveness takes place on an attitudinal continuum from hostility to friendliness: Toward a closer union of forgiveness theory and measurement. Journal of Personality and Social Psychology. Advance online publication. https://doi.org/10.1037/pspi0000227

16.

Griffin

B. J.

Worthington

E. L.

Jr. Lavelock

C. R.

Greer

C. L.

Lin

Davis

D. E.

Hook

J. N.

(2015). Efficacy of a self-forgiveness workbook: A randomized controlled trial with interpersonal offenders. Journal of Counseling Psychology, 62, 124–136.

17.

Hair

J. F.

Anderson

R. E.

Babin

B. J.

Black

W. C.

(2010). Multivariate data analysis: A global perspective (Vol. 7). Pearson.

18.

Hall

J. H.

Fincham

F. D.

(2005). Self-forgiveness: The stepchild of forgiveness research. Journal of Social & Clinical Psychology, 24, 621–637.

19.

Hall

J. H.

Fincham

F. D.

(2008). The temporal course of self-forgiveness. Journal of Social & Clinical Psychology, 27, 174–202.

20.

Hamilton

Gagne

P. E.

Hancock

G. R.

(2003, April). The effect of sample size on latent growth models [Conference session]. Annual Meeting of the American Educational Research Association, Chicago, IL, United States. https://eric.ed.gov/?id=ED476862.

21.

Karremans

J. C.

Van Lange

P. A.

(2004). Back to caring after being hurt: The role of forgiveness. European Journal of Social Psychology, 34, 207–227.

22.

Karremans

J. C.

Van Lange

P. A.

(2008). The role of forgiveness in shifting from “me” to “we.” Self and Identity, 7, 75–88.

23.

Kelln

B. R.

Ellard

J. H.

(1999). An equity theory analysis of the impact of forgiveness and retribution on transgressor compliance. Personality and Social Psychology Bulletin, 25, 864–872.

24.

Lomore

C. D.

Spencer

S. J.

Holmes

J. G.

(2007). The role of shared-values affirmation in enhancing the feelings of low self-esteem women about their relationships. Self and Identity, 6, 340–360.

25.

McCullough

M. E.

Luna

L. R.

Berry

J. W.

Tabak

B. A.

Bono

(2010). On the form and function of forgiving: Modeling the time-forgiveness relationship and testing the valuable relationships hypothesis. Emotion, 10, 358–376.

26.

McCullough

M. E.

Root

L. M.

Cohen

A. D.

(2006). Writing about the benefits of an interpersonal transgression facilitates forgiveness. Journal of Consulting and Clinical Psychology, 74, 887–897.

27.

McCullough

M. E.

Worthington

E. L.

Jr. Rachal

K. C.

(1997). Interpersonal forgiving in close relationships. Journal of Personality and Social Psychology, 73, 321–336.

28.

McNulty

J. K.

(2011). The dark side of forgiveness: The tendency to forgive predicts continued psychological and physical aggression in marriage. Personality and Social Psychology Bulletin, 37, 770–783.

29.

Mooney

Strelan

McKee

(2016). How forgiveness promotes offender pro-relational intentions: The mediating role of offender gratitude. British Journal of Social Psychology, 55, 44–64.

30.

Okimoto

T. G.

Wenzel

(2008). The symbolic meaning of transgressions: Towards a unifying framework of justice restoration. In Hegtvedt

Clay-Warner

(Eds.), Advances in group processes (pp. 291–326). Emerald.

31.

Okimoto

T. G.

Wenzel

(2009). Punishment as restoration of group and offender values following a transgression: Value consensus through symbolic labelling and offender reform. European Journal of Social Psychology, 39, 346–367.

32.

Pelucchi

Paleari

Regalia

Fincham

F. D.

(2013). Self-forgiveness in romantic relationships: It matters to both of us. Journal of Family Psychology, 27, 541–549.

33.

Riek

B. M.

Mania

E. W.

(2012). The antecedents and consequences of interpersonal forgiveness: A meta-analytic review. Personal Relationships, 19, 304–325.

34.

Schumann

(2018). The psychology of offering an apology: Understanding the barriers to apologizing and how to overcome them. Current Directions in Psychological Science, 27, 74–78.

35.

Shnabel

Nadler

(2008). A needs-based model of reconciliation: Satisfying the differential emotional needs of victim and perpetrator as a key to promoting reconciliation. Journal of Personality and Social Psychology, 94, 116–132.

36.

Tabachnick

Fidell

(2019). Using multivariate statistics (7th ed.). Pearson.

37.

Turner

J. C.

(1987). The analysis of social influence. In Turner

J. C.

Hogg

M. A.

Oakes

P. J.

Reicher

S. D.

Wetherell

M. S.

(Eds.), Rediscovering the social group: A self-categorization theory (pp. 68–88). Basil Blackwell.

38.

Walker

M. U.

(2006). Moral repair: Reconstructing moral relations after wrongdoing. Cambridge University Press.

39.

Wenzel

Okimoto

T. G.

(2010). How acts of forgiveness restore a sense of justice: Addressing status/power and value concerns raised by transgressions. European Journal of Social Psychology, 40, 401–417.

40.

Wenzel

Okimoto

T. G.

(2012). The varying meaning of forgiveness: Relationship closeness as a moderator of forgiveness effects on feelings of justice. European Journal of Social Psychology, 42, 420–431.

41.

Wenzel

Okimoto

T. G.

Feather

N. T.

Platow

M. J.

(2008). Retributive and restorative justice. Law and Human Behavior, 32, 375–389.

42.

Wenzel

Okimoto

T. G.

Feather

N. T.

Platow

M. J.

(2010). Justice through consensus: Shared identity and the preference for a restorative notion of justice. European Journal of Social Psychology, 40, 909–930.

43.

Wenzel

Turner

J. K.

Okimoto

T. G.

(2010). Is forgiveness an outcome or initiator of sociocognitive processes? Rumination, empathy, and cognitive appraisals following a transgression. Social Psychological and Personality Science, 1(4), 369–377.

44.

Wenzel

Woodyatt

Hedrick

(2012). No genuine self-forgiveness without accepting responsibility: Value reaffirmation as a key to maintaining positive self-regard. European Journal of Social Psychology, 42, 617–627.

45.

Whitton

S. W.

James-Kangal

Rhoades

G. K.

Markman

H. J.

(2018). Understanding couple conflict. In Vangelisti

A. L.

Perlman

(Eds.), The Cambridge handbook of personal relationships (2nd ed., pp. 297–310). Cambridge University Press.

46.

Woodyatt

Wenzel

(2013a). Self-forgiveness and restoration of an offender following an interpersonal transgression. Journal of Social and Clinical Psychology, 32, 225–259.

47.

Woodyatt

Wenzel

(2013b). The psychological immune response in the face of transgressions: Pseudo self-forgiveness and threat to belonging. Journal of Experimental Social Psychology, 49, 951–958.

48.

Woodyatt

Wenzel

(2014). A needs-based perspective on self-forgiveness: Addressing threat to moral identity as a means of encouraging interpersonal and intrapersonal restoration. Journal of Experimental Social Psychology, 50, 125–135.

49.

Woodyatt

Wenzel

De Vel-Palumbo

(2017). Working through psychological needs following transgressions to arrive at self-forgiveness. In Woodyatt

Worthington

E. L.

Wenzel

Jr., M.

Griffin

B. J.

(Eds.), Handbook of the psychology of self-forgiveness (pp. 43–58). Springer.

50.

Woodyatt

Wenzel

Ferber

(2017). Two pathways to self-forgiveness: A hedonic path via self-compassion and a eudaimonic path via the reaffirmation of violated values. British Journal of Social Psychology, 56, 515–536.

51.

Worthington

E. L.

Jr. (2006). Forgiveness and reconciliation: Theory and application. Brunner-Routledge.

52.

Yoshimura

S. M.

Boon

S. D.

(2018). Communicating revenge in interpersonal relationships. Lexington Books.

53.

Ysseldyk

Matheson

Anisman

(2007). Rumination: Bridging a gap between forgivingness, vengefulness, and psychological health. Personality and Individual Differences, 42, 1573–1584.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.09 MB