Apology–forgiveness cycle in restorative justice, but how?

The apology–forgiveness cycle is a simple but powerful process for conflict resolution. Given the prevalence of apology and forgiveness in restorative justice (RJ), the apology–forgiveness cycle may take place. However, there is a lack of theoretical understanding of the relationship between apology and forgiveness in the RJ processes. After identifying key elements and impediments of the apology–forgiveness cycle during RJ meetings based on the existing literature, we develop a theoretical model of the apology–forgiveness cycle during RJ encounters. This typology explains how the apology–forgiveness cycle is intertwined with the RJ process, subsequently facilitating, blocking, and changing its sequence. There are four cycles: (1) apology facilitating forgiveness, (2) apology without forgiveness, (3) forgiveness promoting apology, and (4) forgiveness without apology. We conclude by offering future directions for research on the apology–forgiveness cycle in RJ.


Introduction
To cause and receive offences is a very human and inevitable experience in our everyday life. Therefore, it needs to be dealt with properly to avoid further conflict. What will be done in its aftermath is critical to both effectively addressing the harm and, if appropriate and meaningful to the parties involved, maintaining the relationship between the transgressor and the victim. In this regard, what is called an 'apology-forgiveness cycle' (Tavuchis, 1991) may be an effective conflict-solving mechanism. It is a simple but powerful process. The transgressor apologies and the victim may then forgive the transgressor or vice versa (Fehr et al., 2010;Strang et al., 2014;Strelan et al., 2016). This sequence may, in turn, pave a path to reconciliation (Leunissen et al., 2013;Shnabel and Nadler, 2008).
Restorative justice (RJ) is a new mode of responding to crime. While what qualifies as RJ remains controversial (Doolin, 2007;Wood and Suzuki, 2016), RJ commonly involves a face-toface dialogue between a victim and an offender (Daly, 2016), who under the guidance of a trained facilitator and in the presence of their supporters come together in one place and discuss the offence, why it happened, and what needs to be done to repair the resultant harm (UNODC, 2020). Restorative justice emerged in the 1970s in response to critiques towards the conventional justice system as retributive (Zehr, 1990) and has since been utilised worldwide (Van Ness, 2001). Although there is a variation in terms of what form of RJ is utilised and at which stage of the conventional justice system RJ is situated (cf. Shapland, 2012), RJ is now an essential part of the conventional justice system in many countries, particularly in the area of youth justice (Zinsstag and Vangraechem, 2012).
As RJ is widely used to deal with the aftermath of a crime, questions arise in relation to the apology-forgiveness cycle. Is the apology-forgiveness cycle something that should be aimed for in RJ? While defining apology (cf. Slocum et al., 2011) and forgiveness (cf. Jenkins, 2019) have both been contested, an apology can be defined as 'an acknowledgement of an offense together with an expression of remorse ... [and] an ongoing commitment by the offending party to change his or her behavior' (Lazare, 2004: 263), and forgiveness as 'a willingness to abandon one's right to resentment, negative judgement, and indifferent behaviour toward one who unjustly injured us, while fostering the undeserved qualities of compassion, generosity, and even love toward him or her ' (Enright et al., 1998: 46-47). There is a consensus in the literature that both apology (Hayes, 2006;Stubbs, 2007) and forgiveness (Van Stokkom, 2008;Walgrave, 2011) should not be considered a primary goal of RJ. Even without apology and forgiveness, RJ can still be beneficial for victims and offenders in terms of fairness, restoration, and legitimacy (e.g. Strang et al., 2013;Wilson et al., 2017). However, if apology and forgiveness are offered and take place in a sequence during RJ meetings, it has the potential to lead to what Daly (2002: 70) referred to as 'a "nirvana" story of repair and good will' because the apology-forgiveness cycle may be 'the key to reconciliation, victim satisfaction, and decreasing recidivism' (Retzinger and Scheff, 1996: 316).
To our knowledge, however, there is no research that has deeply explored the apology-forgiveness cycle in RJ. Some scholars only touched lightly upon it when they referred to either apology or forgiveness in RJ. For instance, concerning the ritual of apology in RJ, Bolívar et al. (2013: 131) mentioned that 'an offenders' apology could have ... a meaning independently of a victim's willingness to forgive'. Similarly, focusing on the role of forgiveness in RJ, Shapland (2020) argued that forgiveness does not necessarily require an apology. One exception may be Retzinger and Scheff (1996) who modelled the apology-forgiveness cycle in RJ. However, they focused solely on one cycle where apology precedes forgiveness. As will be discussed below, there are other possible patterns of a sequential relationship between apology and forgiveness in the RJ process.
This lack of research may be, in part, because there is limited empirical and theoretical RJ literature on both apology and forgiveness. This scarcity may be related to concerns over the perceived roles of apology and forgiveness in RJ (cf. Suzuki and Jenkins, 2020). As a result of victimisation, victims have various needs including participation, vindication, validation, restoration, accountability, and so on (Bolitho, 2015;Daly, 2017;Pemberton and Vanfraechem, 2015). Apology is also one of them (Choi et al., 2010b;Pali and Karin, 2011;Strang, 2002) because apology has the potential to contribute to meeting other victim needs such as vindication (Zehr, 2002), restoration (Bolívar, 2010), and accountability (Keenan, 2014). Despite such an important role and function of apology, RJ scholars argue that an apology should not be regarded as a primary goal in RJ; rather, it must emerge spontaneously from the offender and not be coerced (e.g. Hayes, 2006). Forgiveness in RJ is similarly viewed as an emergent standard as RJ scholars emphasise that forgiveness is a 'by-product' that should ultimately be left up to victims' discretion (e.g. Braithwaite, 2002). In sum, both apology and forgiveness are considered a secondary goal in RJ.
Nevertheless, the apology-forgiveness cycle may appear in RJ because both apology and forgiveness may simultaneously be present in the RJ meetings. This assumption about the prevalence of the apology-forgiveness cycle may be right given the fair prevalence of apology and forgiveness in RJ encounters. Recent research suggested that a simple verbal apology (e.g. 'I'm sorry') was offered in one-third of RJ cases, and 'full' apologies that involve (1) admitting responsibility for the behaviour and outcomes, (2) acknowledging the harm done and that it was wrong, (3) expressing regret or remorse for the harm done, (4) offering to repair the harm or make amends, and (5) promising not to repeat the behaviour in the future and to work toward good relations were provided in approximately one-fifth of cases (Dhami, 2012: 47). Similarly, Suzuki and Jenkins (2020) found in their recent review of the literature that forgiveness might be present in 30-80% of RJ cases. Thus, there is a strong likelihood that the apology-forgiveness cycle may take place in the RJ processes.
As apology and forgiveness hold substantive benefits for victims and offenders, it is imperative to understand the mechanism of how and why the apology-forgiveness cycle may or may not take place and how it produces such positive outcomes (cf. Bazemore and Green, 2007). Since the benefits of RJ, such as satisfaction, fairness, restoration, and legitimacy, have already been confirmed (e.g. Strang et al., 2013;Wilson et al., 2017), it is now the time to embark on a new research journey to advance our knowledge about how RJ 'works' (e.g. see recent attempts by Bolitho, 2017;Suzuki and Yuan, 2021). Exploring the micro-mechanism of the apology-forgiveness cycle during RJ encounters can be part of this attempt. Addressing the under-investigated topic of the apologyforgiveness cycle will not only help to improve the quality of RJ practices, leading to further benefits, but such focus is also in line with the recent call for a research shift in RJ from 'what works' to 'how it works' (e.g. Lanterman, 2021; Suzuki, 2020) because we do not know how this cycle may or may not occur in the RJ meetings. Taken together, examining the apology-forgiveness cycle greatly adds not only to the literature on apology and forgiveness in RJ but also to the overall RJ literature.
The aim of this research is to advance a theoretical understanding of the relationship between apology and forgiveness in the RJ processes. More specifically, as Suzuki and Jenkins (2020) did in terms of the relationship between self-forgiveness and desistance in RJ, we intend to situate the apology-forgiveness cycle within the RJ process. To this end, we take two steps. Drawing on the existing but limited literature on the apology-forgiveness cycle, we first seek to identify key elements and impediments in the apology-forgiveness cycle in RJ by analysing what elements of RJ may facilitate the apology-forgiveness cycle and what impediments may hinder it. For the latter, Blecher (2011) has listed some impediments in the apology-forgiveness cycle, but only from an apology-related perspective. We aim to extend the list by drawing on the literature on the apology-forgiveness cycle. Based on the key elements and impediments that are identified as relevant to the apology-forgiveness cycle during RJ meetings, we then seek to develop a theoretical typology of the apology-forgiveness cycle in RJ. Forgiveness can be given without an apology (e.g. Shapland, 2020). Furthermore, forgiveness from victims may elicit an apology from offenders in the RJ processes (Exline et al., 2007). Hence, considering a variety of sequences in the apologyforgiveness cycle is important because there may also be a case where forgiveness may precede an apology. We intend to identify patterns of sequential relationships between apology and forgiveness.
This article starts by describing the benefits of the apology-forgiveness cycle. This is followed by a discussion of the key elements and impediments of the apology-forgiveness cycle in RJ. From this debate, we offer a theoretical typology of the apology-forgiveness cycle within the RJ process. We conclude by providing implications for future research on the apology-forgiveness cycle in RJ.
Before moving forward, we wish to define the scope of this paper by making four points. First, because it is possible that forgiveness may precede an apology or that only one or the other is offered, the sequences of the apology-forgiveness cycle must remain fluid. Put differently, we take into account the case where apology does not facilitate forgiveness, the case where forgiveness does not elicit an apology, and the case where only either an apology or forgiveness is provided. Since we intend to unravel the complex theoretical relationship between apology and forgiveness, it is necessary to consider every pattern of a sequential relationship between apology and forgiveness. In this sense, the concept of the apology-forgiveness cycle in this paper is used as a broad notion that captures the complex sequential relationship between apology and forgiveness in the RJ processes. Second, we wholeheartedly agree that both apology and forgiveness should not be coerced in RJ (Armour and Umbreit, 2018). This paper is only concerned about the micro-mechanism of how the apology-forgiveness cycle may or may not occur in RJ. Hence, we do not enter a debate over the appropriateness of apology and forgiveness in RJ. To examine the relationship between apology and forgiveness in the RJ meetings, the existence of both or either apology or forgiveness during RJ encounters is purposely taken for granted. Third, since our interest lies in the apology-forgiveness cycle, we intend to not discuss and analyse apology and forgiveness separately unless necessary. Although we may touch upon some literature on only either apology or forgiveness due to the limited literature, particularly in relation to the benefits of the apology-forgiveness cycle, we focus on research pertinent to the apology-forgiveness cycle. Finally, while both apology (e.g. Szablowinski, 2012) and forgiveness (e.g. Green et al., 2008) can be provided by a 'third party' (e.g. support persons, community representatives), we focus on the apologyforgiveness cycle between victims and offenders. This is because, in our view, victims and offenders are the 'primary' stakeholders in a strict sense that have wronged and have been wronged directly by crime and require restoration and reconciliation in their relationship.

Benefits of the apology-forgiveness cycle
While there are few studies on the benefits of the apology-forgiveness cycle, there is abundant literature on the benefits of apology and forgiveness, respectively. Given this limitation, we use the latter to describe the overall benefits of the apology-forgiveness cycle. We argue that this approach is appropriate not only because of the limited literature but also because it may be possible to estimate the overall benefits of the apology-forgiveness cycle by examining those of apology and forgiveness. respectively. We start by discussing the benefits of an apology.
An apology may have a healing effect on victims as it indicates that offenders take the responsibility for what they have done to their victims (Bennett, 2006), assuring victims that victimisation is not their fault (Pranis, 2004). This moral action can satisfy the justice needs of the victims (Herman, 2005) and bring about the restoration of self-respect and dignity (Lazare, 2004), thereby helping the victims move on from the impact of crime (Vines, 2017). Subsequently, an apology may reduce the desire for revenge towards the offenders in victims (Walgrave, 2011). Indeed, empirical research suggested that an apology can reduce anger among victims (Anderson et al., 2006).
Interestingly, the benefits of an apology extend beyond victims because offenders may also benefit from apologising. By committing a crime, offenders breach the norm accepted in the community, resulting in diminished trust in the wrongdoers. An apology can be the first step in restoring trust as it signals that they have 'received, understood, and the moral message directed at him [or her] and that he [or she] accepts the values implicit in it' (Bennett, 2006: 129). An apology, as a moral act, indicates that the offender now subscribes to the norm accepted in society (Vines, 2017). As revealed in the research, this can, in turn, increase the trustworthiness of offenders (Ma et al., 2019). Therefore, offenders who have apologised may no longer be considered an 'outsider' who does not comply with the rules in the community; rather, they may be welcomed as lawabiding citizens who should be reintegrated into the community (Johnstone, 1998).
Forgiveness also has benefits for both victims and offenders. Yet, unlike research on the benefits of apology, there are only a few studies that empirically examined the benefits of forgiveness for offenders (e.g. Jenkins, 2018;Wenzel et al., 2021). In contrast, there is ample evidence for the benefits of forgiveness for victims. According to the existing literature, forgiveness can promote positive health outcomes (Rasmussen et al., 2019) such as lower blood pressure levels (Lawler et al., 2003) and emotional restoration such as anger reduction (Lawler-Row et al., 2008) and protection against depression (Toussaint et al., 2008). Moreover, research suggests that forgiveness is associated with restoration of a sense of justice (Wenzel and Okimoto, 2010) and restoration of value consensus with the offenders (Wenzel et al., 2021). Forgiveness may be particularly beneficial for victims because it has the potential to heal the impact of victimisation (Armour and Umbreit, 2005). Forgiveness cannot undo the harm caused by crime as it is not 'condoning, forgetting or reconciling' (Enright andKittle, 2000: 1623), but it can help the victim remember 'the hurts in a different way', thereby 'changing the emotional attachment to the offence' (Szablowinski, 2010: 473).
If apology and forgiveness have these benefits, respectively, the possible benefits of the apology-forgiveness cycle may be substantive. In addition, these possible benefits of the apologyforgiveness cycle align with the positive outcomes that are expected from RJ meetings, including responsibility (Calhoun and Pelech, 2013), emotional restoration and recovery (Angel et al., 2014), and reintegration (Braithwaite and Mugford, 1994). Due to limited research on the apology-forgiveness cycle in general, and because of the potential benefits of the apology-forgiveness cycle, it is imperative to unpack the micro-mechanism of how the apology-forgiveness cycle may or may not occur in RJ. Since the role and functions of apology and forgiveness in the RJ process are not probed deeply, this examination may provide an insight into why RJ encounters benefit victims and offenders.

Key elements and impediments of apology-forgiveness cycle in restorative justice
In the RJ context, Retzinger and Scheff (1996) described what is now known as the apology-forgiveness cycle as a 'core sequence'. Furthermore, while they used the term forgiveness, they did not refer to the term apology in their model and used the term remorse. They wrote, The ideal outcome from the point of view of symbolic reparation, is constituted by two steps: the offender first clearly expresses genuine shame and remorse over his or her actions. In response, the victim takes at least a first step towards forgiving the offender for the trespass. These two steps are the core sequence. (Retzinger and Scheff, 1996: 316, emphasis in the original) Consistent with the benefits of the apology-forgiveness cycle, this core sequence, if it happens, may have substantive positive impacts not only on victims and offenders but also on a wider community: The core sequence generates repair and restoration of the bond between victim and offender, after this bond had been severed by the offender's crime. The repair of this bond symbolizes a more extensive restoration that is to take place between the offender and the other participants, the police and the community. (Retzinger and Scheff, 1996: 316) While Retzinger and Scheff modelled the apology-forgiveness cycle in RJ, a few problems remain. First, their model may be too simplistic as it does not take into account the case where forgiveness precedes an apology as well as the case where only either apology or forgiveness is offered. Relatedly, their model may need more elaboration as it does not incorporate factors germane to the RJ process, such as it is a multi-faceted process that involves not only a victim and an offender but also their supporters and a facilitator among others. Taken together, the apology-forgiveness cycle in RJ needs a more careful examination. It is necessary to consider what factors of RJ function as a facilitator or a challenge for the apology-forgiveness cycle. In what follows, we seek to identify possible key elements and impediments of the apology-forgiveness cycle in RJ while considering possible patterns of a sequential relationship between apology and forgiveness.

Key elements
In cases where an apology precedes forgiveness in the apology-forgiveness cycle, central to this cycle may be that the apology matches the needs of victims. This is demonstrated by research by Cowden et al. (2019). To examine what kind of an apology may facilitate forgiveness, they asked participants to recall a past transgression they experienced and evaluate the apology offered by their transgressor. Their findings revealed that the apology that was congruent with the 'victims' preferences for empathy, compensation, and acknowledgement of violated norms' was more likely to promote forgiveness among victims (Cowden et al., 2019: 411). Since an apology has a variety of components such as taking responsibility, this finding leads to a question as to how offenders should apologise in a way that is consistent with the victims' needs. Indeed, by exploring how and why an apology can contribute to conflict-solving, Kirchhoff et al. (2012) showed that whereas the significance of the components in an apology such as conveying emotions and admitting fault is not equal, the apology may need to be more 'complete', especially after a severe transgression. Therefore, it may be necessary to examine what components of an apology can facilitate forgiveness in RJ contexts because RJ deals with crime, which is usually more serious than daily transgressions discussed in psychology literature.
First and foremost, when offenders apologise, they need to take responsibility for what they have done to the victims. To be effective, an apology needs to consist of multiple elements such as admittance of responsibility, expression of remorse, and offer of reparation (Carroll et al., 2017;Gill, 2000;Govier and Verwoerd, 2002). If any component is lacking, an apology is considered 'partial' rather than 'full', becoming less effective (Dhami, 2017). Accordingly, if an apology lacks the responsibility component, it is less likely to facilitate forgiveness. Allan et al. (2006) examined the association between apology and forgiveness in the South African Truth and Reconciliation Commission (TRC). Drawing on a survey distributed to the people who were the victims of a gross human rights violation, their research suggested that what matters to forgiveness was how the transgressors behaved after their wrongdoing. While not statistically significant, when the transgressors apologised, the victims were more likely to be forgiving. On the other hand, while not statistically significant, when the transgressors made excuses for their wrongdoing, victims were less likely to be forgiving. In a nutshell, for the apology-forgiveness cycle to take place, offenders may need to acknowledge their responsibility for what they have done when they offer an apology.
There are two possible reasons why an apology that entails taking responsibility may be effective in facilitating forgiveness and why the RJ meetings may be an ideal environment that can promote such a type of apology. To begin with, an apology that entails taking responsibility can convey offenders' remorse to victims. According to Davis and Gold (2011), remorse perceived through apology may facilitate forgiveness because it signals that offenders are less likely to repeat the wrongdoing, resulting in an increase in empathy and forgiveness. By treating offenders with respect and fairness, having them listen to the victims' stories, and surrounding them with people who care for them, RJ encounters can nurture repentance within offenders (Braithwaite, 2000). Alternatively, an apology that entails taking responsibility may facilitate forgiveness because it conveys shame and guilt from offenders. Hareli and Eisikovits (2006) demonstrated that if the victims knew that the apology was motivated by shame and guilt, they were more likely to forgive their perpetrators. As is evident in Braithwaite's (1989) reintegrative shaming, which is considered 'one of the theoretical underpinnings' of RJ (Strang, 2020: 23), both emotions, especially shame, are critical in the RJ processes. Rather than stigmatising offenders through negative labelling as a 'bad' person, the reintegrative shaming theory suggests that RJ encounters are able to separate the deed and self in offenders and accept the offenders as a 'good' person while condemning their offending behaviours (Strang et al., 2011).
The other essential component of an apology for the apology-forgiveness cycle that is pertinent to the RJ processes may be a reparation offer. According to Witvliet et al. (2020), apology and restitution can independently promote forgiveness. Given this finding, an apology may be more effective in communicating the sincerity of offenders when it is combined with a reparation offer. In fact, Jeter and Brannon (2018) revealed that such a 'costly apology' may be more impactful on victim forgiveness than other types of apologies such as a mere verbal apology. The apology-forgiveness cycle may be more likely to be present if an apology is combined with a reparation offer. In this regard, RJ meetings may help to achieve such a costly apology because it is expected to generate a reparative agreement plan to restore the harm caused by crime. It is common to discuss and include apology as well as conciliatory behaviours in the agreement plan (Dhami, 2016), and there is a high completion rate for this agreement plan (Latimer et al., 2005).
Thus far, we have focused on the key elements of the apology-forgiveness cycle in RJ, particularly the case where apology facilitates forgiveness. However, there is a sequence of the apologyforgiveness cycle in RJ where forgiveness precedes apology. This sequence may be possible because, for example, various faith traditions require the faithful to forgive others -to, in effect, turn the other cheek, in order to lay claim upon 'divine' forgiveness. For such religious devotees, giving forgiveness in the aftermath of wrongdoing is considered an ethical obligation and a religious duty to God (Newman, 1987). It could be argued that a religious construction of forgiveness would predispose victims to forgive even without an offender's apology. Forgiveness without a preceding apology is also consistent with the view of forgiveness as a gift given unconditionally and one that does not require repentance on the part of the offender. North (1987) states, If I am to forgive I must risk extending my trust and affection, with no guarantee that they will not be flung back in my face or forfeited again in the future. One might even say that forgiveness is an unconditional response to the wrongdoer, for there is something unforgiving in the demand for a guarantee. (p. 505) This 'reversed' sequence may occur in RJ encounters because victims often participate in the RJ meetings for 'other-regarding' reasons such as helping offenders (Van Camp, 2017;Van Stokkom, 2008). Research suggests that such victims' generous attitudes, which involve taking the perspective of offenders, are linked with forgiveness (Takaku, 2001). Some victims may verbally offer or may otherwise show their forgiving attitude without receiving an apology in the RJ process. Forgiveness before apology may contribute to the apology-forgiveness cycle because forgiveness can open the door for an apology from offenders (Exline et al., 2007). Indeed, research by Leunissen et al. (2012) suggested that given the risks associated with offering an apology including being blamed for the offence, perpetrators may strategically offer an apology, that is, they may be more likely to apologise when they perceive a likelihood that victims may forgive them in response to their apology.
Putting aside the argument as to whether it is appropriate in RJ, an apology driven by forgiveness may not be without benefits for victims. This is evident in the effect of what is called a 'cued' apology. According to Peyton and Goei (2013), a cued apology is one that is offered as a result of victims' explicit demand for an apology. Their research found that a cued apology may unexpectedly be more effective in increasing perceptions of sincerity and eliciting forgiveness than a spontaneous apology. Therefore, when examining the micro-mechanism of the apology-forgiveness cycle in the RJ processes, it may be worthwhile to examine cases where forgiveness precedes an apology to unpack its micro-mechanism.

Impediments
In addition to the elements that can facilitate the apology-forgiveness cycle in the RJ processes, there are impediments that hinder its occurrence. A substantial impediment is the utilisation of various neutralisation techniques. Offenders who seek to deflect blame or justify their behaviour may not take responsibility for their actions when offering an apology in the RJ processes. Drawing on Tavuchis' (1991) sociological theory of apology, Hayes (2006) aptly investigated this possibility. He noted that offenders may be encouraged to deny their offending in RJ because of 'competing demands'. As part of the RJ process, offenders are expected to accept responsibility for what they have done as well as to explain what led them to commit the offence. In the course of this speech act, offenders 'may drift from apologetic discourse to mitigating accounts and back again' as they offer 'various excuses and justifications' for their offending while acknowledging responsibility and offering an apology (Hayes, 2006: 378, emphasis in the original). Such an apology, Hayes (2006: 378) argued, is less likely to 'convince victims of offenders' "worthiness" (Tavuchis, 1991) but instead may be seen as attempts to acknowledge blame but deflect shame', thereby neutralising the responsibility of offending. An apology offered with neutralisation is less likely to promote forgiveness in victims.
Another impediment may be coercion. While RJ aims to treat every participant equally, there may be an inherent power imbalance between participants (Suzuki and Wood, 2018). This power imbalance may be the most salient for offenders because they are sometimes coerced to participate in the RJ processes under the fear of more severe punishment (Abrams et al., 2006). Yet, the problem is not limited to the voluntariness of offender participation. More problematic may be the risk of coercion on offenders' behaviours both during and after RJ meetings. In fact, research suggests that some offenders feel pressured to not only accept an agreement plan that they were not in complete agreement with (Beckett et al., 2005) but also offer an apology to their victims (Hoyle et al., 2002). Such a coerced apology may have a detrimental impact on the apology-forgiveness cycle because victims look less favourably on a coerced apology than an apology that is offered voluntarily (Jehle et al., 2012). As a result, an apology offered as a result of coercion may not elicit forgiveness from victims.
Finally, the very environment of RJ meetings itself may function as an impediment to the apology-forgiveness cycle. While the RJ process is supposed to be a supportive environment where every participant can freely express their opinion, this is not always the case. Occasionally, participants may lecture offenders (Maruna et al., 2007), and some victims are too defensive towards offenders (Kenney and Clairmont, 2008). Hostile environments do not elicit apologies as offenders may need the assurance that their apology will be 'rewarded'. According to research, perpetrators tend to be reluctant to offer an apology due to an overestimation of its negative effects (Leunissen et al., 2014). When RJ meetings become a hostile environment, there may be an incentive for offenders to not offer an apology, even if victims offer forgiveness.

Typology of the apology-forgiveness cycle in restorative justice process
In the preceding section, we have identified the possible key elements and impediments that are relevant to the apology-forgiveness cycle during the RJ meetings. Following this, we now turn to the theoretical model of the apology-forgiveness cycle within the RJ processes. We discuss how these key elements and impediments are intertwined with the RJ process and identify patterns of the sequential relationship between apology and forgiveness in RJ encounters.
To this end, we draw on the existing typologies of the apology-forgiveness cycle put forth by Lazare (2004) and Toney and Hayes (2017). Lazare (2004) identified four comprehensive categories in the apology-forgiveness cycle that this study will build on to develop the typology of the apology-forgiveness cycle: '1) forgiveness without apology; 2) no forgiveness regardless of the apology; 3) forgiveness that precedes apology; and 4) apology that precedes forgiveness' (p. 131). Toney and Hayes (2017) offered both a typology and a sequential model of the apology-forgiveness cycle, in which they described how the path of the apology-forgiveness cycle may be altered as a result of communication between victims and offenders. Building from their research, we discuss how the key elements and impediments intervene in the RJ process, subsequently facilitating, blocking, and changing the apology-forgiveness cycle.
Our theoretical model has limitations in its scope. As mentioned at the outset, this paper will not address the debate over the appropriateness of apology and forgiveness in RJ, and in our typology, it takes for granted that either or both apology and forgiveness are provided in the RJ process. Also beyond the scope of our model is an explanation as to why the key elements and impediments for the apology-forgiveness cycle appear in RJ encounters. The focus of this study is not about why apology and forgiveness take place in the RJ process individually (e.g. why some offenders offer an apology in the RJ meeting), but about how the existence and/or lack of apology or forgiveness influence the apology-forgiveness cycle alongside the RJ process. Hence, we solely focus on the sequential relationship between apology and forgiveness in RJ encounters. Finally, and perhaps most importantly, we focus on the possible occurrence of the apology-forgiveness cycle during RJ meetings. We acknowledge that apology is sometimes included in the agreement plan offenders are expected to fulfil after the RJ process (Dhami, 2016;Rossner et al., 2013), leading to the possible occurrence of the apology-forgiveness cycle after the RJ processes. However, given the relatively high prevalence of both apology and forgiveness during the RJ meetings that has been mentioned at the beginning of this paper, the apology-forgiveness cycle in RJ may take place within RJ meetings. Therefore, our focus is how the interactive dynamics within RJ encounters are linked with the apology-forgiveness cycle.
Our theoretical typology of the apology-forgiveness cycle consists of the following four cycles (Table 1): (1) apology facilitating forgiveness, (2) apology without forgiveness, (3) forgiveness promoting apology, and (4) forgiveness without apology. We further discuss how the key elements and impediments of the apology-forgiveness cycle during RJ meetings are intertwined with the respective cycle. Table 1 describes the four cycles including their attributes. With the abovementioned limitations in mind, we elaborate on each cycle.
Cycle 1, apology facilitates forgiveness, is likely to be what first comes to mind when one thinks of the apology-forgiveness cycle. For this cycle to occur, offenders need to take responsibility for what they have done when they apologise to their victims. Existing research demonstrated that RJ encounters can encourage offenders to take responsibility for their behaviour (Boriboonthana and Sangbuangamlum, 2013;Calhoun and Pelech, 2013). RJ meetings can facilitate this process as offenders receive support from a trained facilitator, family members, and close friends whose presence and care can help to reduce negative feelings such as shame (Ahmed et al., 2001;Scheuerman and Keith, 2015). The reparation offer, a key element in the RJ processes, can further aid this process as such a costly apology that entails restitution is effective in eliciting forgiveness . Taken together, an apology that entails taking responsibility in conjunction with a reparation offer may facilitate forgiveness in the RJ processes. Offender apologies do not necessarily result in victim forgiveness. Cycle 2 of this typology suggests that the use of neutralisation and coercive strategies may impede forgiveness, even in the presence of an apology. Apologies offered in the context of the offender's attempt to deny or minimise responsibility for their offending (Karp et al., 2004) or in cases where offender supporters defended or made 'excuses for the offender's actions' or did not 'allow him or her to take responsibility' (Rossner and Bruce, 2018: 509) may be perceived as 'partial' by victims and are less likely to facilitate forgiveness in RJ. Alternatively, in the RJ meetings, offenders may offer an apology with a statement of their responsibility, but they may be forced to do so rather than offer the apology spontaneously (Beckett et al., 2005;Hoyle et al., 2002). Such coerced apologies are less likely to be perceived as sincere by victims (Saulnier and Sivasubramaniam, 2015), which may result in victims becoming reluctant to offer forgiveness in RJ encounters.
Cycle 3, forgiveness promoting apology, is the sequence in the apology-forgiveness cycle that appears most counterintuitive as apologies are generally viewed as preceding or eliciting forgiveness. Nonetheless, this cycle is important to consider given that offenders do not necessarily participate in the RJ processes voluntarily (Abrams et al., 2006;Choi et al., 2010a;Zernova, 2007). As Suzuki (2020) demonstrated through his concept of 'readiness', some offenders attend RJ meetings for 'self-regarding' reasons such as avoiding harsher punishment. Even when offenders participate in the RJ processes for self-serving motivations, the victim's generous gift of forgiveness may alter the offender's attitude to the degree that they may be motivated to repay the victims (Kim and Gerber, 2012). The offender, emotionally touched by the victim's forgiveness and perceiving that the victim's forgiveness indicates that there is a low risk of being blamed as a result of an apology (Leunissen et al., 2012), may be inspired to apologise.
Finally, Cycle 4 is the sequence of the apology-forgiveness cycle wherein victims offer forgiveness, but offenders do not apologise. An RJ meeting environment that is hostile to the development of a sincere apology may explain this sequence. Since victims have already offered forgiveness in this cycle, the problem may lie in other participants in the RJ processes. For instance, research suggests that offenders' supporters, particularly parents, who were defensive of their child, discouraged them from taking responsibility for their offending (Hoyle and Noguera, 2008;Kenney and Clairmont, 2008). In addition, facilitators or police officers who addressed the offenders 'in a moralizing or scolding tone' (Campbell et al., 2006: 79) may also create an environment that is inhospitable to the offer of an apology. Even if victims explicitly or implicitly offer forgiveness, offenders may be reluctant to offer an apology as they may not perceive apologising as beneficial (Leunissen et al., 2014). Rather, offenders may view not apologising as more rewarding because a refusal to apologise is linked with an increase in self-esteem (Okimoto et al., 2013).

Conclusion
Given the fair prevalence of both apology and forgiveness in the RJ meetings, the apology-forgiveness cycle may take place during RJ encounters. However, no research to date has examined the relationship between apology and forgiveness in the RJ processes. We sought to fill this gap in the literature by developing a theoretical model of apology-forgiveness in the RJ processes that takes into account the key elements and impediments that are germane to the apology-forgiveness cycle in RJ. This four-part typology of the apology-forgiveness cycle in the RJ meetings contributes to the literature in three ways. First, by focusing on the RJ meetings that deal with the aftermath of law-breaking, this model extends the existing literature on the apology-forgiveness cycle that tends to focus on everyday transgression in a close relationship. Relatedly, by situating key elements and impediments of the apology-forgiveness cycle in the RJ processes, our model extends the existing typology of the apology-forgiveness cycle that does not take into account factors beyond the components of apology and forgiveness. Last but not least, our typology also extends the model by Retzinger and Scheff (1996) because our model is the first that not only incorporates the prevailing notion that apology facilitates forgiveness but also addresses the other possibilities including no forgiveness despite the apology, forgiveness preceding apology, and forgiveness with no apology. While the last three cycles are probable, existing literature on RJ hardly considers them. Hence, our typology offers a new understanding of the occurrence of the apology-forgiveness cycle during RJ encounters. Given the benefits of the apology-forgiveness cycle, this typology contributes to addressing the question of how RJ 'works'.
Despite the significance, our theoretical typology is not without limitations. For one, since our model explores the possibility of the apology-forgiveness cycle during the RJ meetings, it does not embrace the possibility of the apology-forgiveness cycle that takes place after the RJ encounters. Accordingly, by testing our theoretical typology, it may be necessary to modify the model to account for the occurrence of the apology-forgiveness cycle outside the RJ processes. Relatedly, although the typology is based on empirical evidence on the apology-forgiveness cycle, it remains theory-based, requiring empirical scrutiny. Empirical research of this subject is critical because some literature we drew on largely focused on the apology-forgiveness cycle in a close relationship such as a romantic relationship (e.g. Davis and Gold, 2011), and other literature utilised relied on a hypothetical scenario to explore the apology-forgiveness cycle (e.g. Ohtsubo and Watanabe, 2009). Hence, the typology needs to be tested in RJ in a real-world situation that deals with a crime that often takes place between strangers. In this respect, a qualitative approach may be useful for this purpose as it allows an exploration of the subjective dimensions of the apology-forgiveness cycle (Stanfield, 2006). Since both apology and forgiveness, particularly the latter, may not explicitly be expressed in words during RJ meetings (Shapland, 2016;Umbreit et al., 2015), observation of non-verbal communication between participants during the RJ process, as well as interviews about their experiences and perceptions towards RJ encounters, may be crucial to ascertain the presence of the apology-forgiveness cycle during RJ meetings.
In conjunction with the abovementioned research implications, there are other external factors that need to be attended to in research because they may influence the apology-forgiveness cycle during the RJ processes. Age is one important factor to consider as research suggests that older people are more likely to forgive (Allemand, 2008). Recently, some scholars have started to explore the impact of the youthfulness of offenders in the RJ process and outcomes (e.g. Suzuki and Wood, 2018). In contrast, with a few exceptions such as Gal (2011), scant attention has been paid to the age of victims in RJ literature. If the findings of the relationship between age and willingness to forgive also apply to RJ meetings, the apology-forgiveness cycle may be more likely to occur when the victim is an adult. This warrants future research. Similarly, the gender of both offenders and victims needs to be taken into account. According to the literature, women may be more likely to offer an apology (Schumann and Ross, 2010) as well as forgive (Miller et al., 2008) than men.
It is therefore necessary to examine whether the apology-forgiveness cycle may be more likely to appear when either or both the offender and the victim are female. The final external factor that may need to be considered in research on the apology-forgiveness cycle during RJ encounters is offence type and severity. Research indicated that the more serious the offence, the more likely the apology is to be rejected (Bennett and Earwaker, 2001). Yet, simply examining the relationship between the apology-forgiveness cycle and severity of offence may not be sufficient as this relationship may be more complex than that. As Daly (2008) revealed, what matters in the RJ processes may be the context of how the offence happened and how it impacted the victims. Indeed, research by Komiya et al. (2018) showed that compensation for irreplaceable losses was less effective in promoting forgiveness than that for replaceable losses. Hence, what is needed may be to examine how the offence type and severity influence the apology-forgiveness cycle while exploring the context and impact of offending on victims.