Talking punishment: How victim perceptions of punishment change when they communicate with offenders

The myth that restorative justice is the opposite of retributive justice persists, despite a long history of rhetorical challenges. Only empirical evidence can advance the debate, so this article investigates the relationship between punishment and victim–offender communication from the victim’s perspective. Interviews with 40 victims of crime established that some victims saw victim–offender communication and punishment as alternatives, and others saw them as independent. However, more than half the participants expected that communicating with the offender would increase their satisfaction with the offender’s punishment or reported afterwards that this was in fact the case, suggesting that some victims fulfil punishment objectives through communication with the offender. The changes occurred when victims received information about the offender’s punishment, received feedback from the offender or used communication with the offender to impose a mild punishment of their own. Victims were not excessively punitive, but this study demonstrates the existence of an association between punishment and victim–offender communication from at least some victims’ perspectives. This article argues that we should not ignore or attempt to eliminate this relationship. Rather, acknowledging and examining the existence of punishment within victim–offender communication would improve practice and generate better outcomes for victims, offenders and society.

Some early proponents of restorative justice (RJ) contrasted it with retributive justice, claiming that their foundations and objectives are diametrically opposed (Barnett, 1977;Braithwaite, 1989;Eglash, 1977;Fattah, 1998;Zehr, 1990). Others maintained that the relationship between RJ and retributive justice was more complex (e.g. Daly and Immarigeon, 1998;Davis, 1992;Hampton, 1992). In the ensuing decades many arguments have been put forward on both sides of what Daly (2012) calls the 'punishment debate'.
In the 1990s, the punishment debate focused on the relationship between retribution and reparation. Duff (1992), for example, considered reparation as an alternative punishment (rather than an alternative to punishment), and Zedner (1994) argued that offender activities designed to make amends to the victim or community can also deliver retributive goals. In the 2000s, the debate shifted to whether victims should have any type of sentencing authority. Retributivists expressed concern that this could test both the upper and lower limits of proportionality (Ashworth, 2002;Robinson, 2003;Wasik, 1999), but others argued that the punitive and restorative aspects of victim-influenced sentencing are largely compatible if appropriate safeguards are in place (Roche, 2003;O'Hear, 2005;Young and Hoyle, 2010).
While the debates about reparation and victim-influenced sentencing continued, some theorists turned their attention to the relationship between retributive justice and a third type of RJ practice: communication between victims and offenders. In its infancy, communication was generally offered as an alternative to state-administered punishment (e.g. Umbreit et al., 1994), so there was little reason to suspect anything other than an inverse relationship between the two. However, researchers soon started to consider whether a meeting with a victim might itself constitute punishment for the offender, observing that such meetings are likely to entail suffering on the part of the offender (Robinson, 2003), that making an apology is 'burdensome' (Duff, 2002), and that calling the offender 'to account' is inevitably an imposition (Von Hirsch et al., 2003).
Could we not just agree with Daly then that the dichotomy between restorative and retributive justice is a 'nonsense' (Daly, 2016: 15), and consider the case closed? Although Daly and many others have attempted to dispel the myth that RJ is the opposite of retributive justice (Daly, 2002), it continues to permeate both the academic literature and RJ practice. For example, in social psychology experiments participants are asked to choose between 'RJ' and 'punishment' (Moss et al., 2019;Rasmussen et al., 2018), in public discourse RJ is often described as 'non-punitive' (e.g. Donnelly, 2015;McKend, 2020), and in RJ practice some facilitators say they would not proceed with a victim-offender meeting if the victim had 'punitive' motivations. 1 Why does this myth that restorative and retributive justice are opposites persist? Despite a long history and many pages of argumentation, there is a lack of empirical evidence regarding the nature of the relationship from the perspective of stakeholders in the justice process. A full investigation would ideally incorporate the views of offenders, criminal justice practitioners and the public, but I argue that the victim's perspective is especially informative for two primary reasons. First and foremost, victims and offenders each have a degree of control over the other's experience when they communicate with one another, and most RJ schemes explicitly endow all the parties with some degree of power and control over the process (Dignan, 2004). Thus, if the victim wishes the offender to suffer for their crime, the victim has somealbeit limitedpower to impose such suffering. Secondarily, if communication with the offender and state-imposed custodial sentences both contribute to the fulfilment of victims' desire for punishment, this would suggest that they have something in common. It would not necessarily mean that victim-offender communication constitutes punishment, but it would suggest that they are positively related, and it would challenge the premise that they are mutually exclusive.
Few existing studies have specifically investigated the relationship between punishment and communication with the offender from the victim's perspective, but studies of victim motivation and RJ outcomes provide some indications about its nature. Some studies indicate that there is an inverse relationship between victims' desire for punishment and their willingness to communicate with the offender, corroborating the claim that they are opposites. For example, when offered a restricted choice between RJ and punishment, some victims choose RJ as an alternative (Aertsen and Peters, 1998;Doble and Greene, 2000;Moss et al., 2019;Pereira, 2017;Rasmussen et al., 2018;Umbreit et al., 1994), and they are more likely to do so when punishment goals have already been met Darley, 2006, 2009). Other studies indicate that there may be no relationship between victim willingness to communicate with the offender and perception of the offender's punishment, suggesting that they are independent. Paul (2015), for example, found that the importance participants placed on punishment did not predict their (hypothetical) willingness to meet the offender, whereas other types of justice goal did (e.g. offender rehabilitation).
There have been no explicit tests of whether victims expect and experience fulfilment of punishment goals through communication with the offender, although Gade draws on his own experience as a mediator and knowledge of the victims' perspective in two case studies to argue that RJ does sometimes constitute punishment (2021). Many have found that victims who communicate with the offender are less angry or less vengeful afterwards (Sherman et al., 2005;Strang et al., 2006;Umbreit et al., 1994), but it is difficult to distinguish whether such outcomes are derived from fulfilment of punishment objectives or forgiveness (Strelan et al., 2020). Victims who communicate with offenders are more likely to say the offender has been 'held accountable' (Poulson, 2003), but this does not necessarily entail the view that they have been punished. Van Camp and Wemmers (2013: 130) cite a single case in which the victim chose monetary compensation over an apology because she knew it would be a good way 'to get back at' and 'hurt' the offender. However, the authors imply this is an unusual motive for RJ, and they make no further comment on it.

Aim of the study
The aim of this study is to understand the relationship between victim-offender communication and punishment from the victim's perspective. Do victims expect communication with the offender to fulfil their punishment objectives? Do victims experience fulfilment of punishment objectives through communication with the offender? If so, how?
By communication between victims and offenders I mean both direct, face-to-face communication (e.g. meetings or conferences), and indirect communication (e.g. via a facilitator or by letter). In legal terms, punishment tends to mean censure and sanction (Zedner, 2016) inflicted upon an offender (Hart, 1968), by a legal authority (Flew, 1954). However, as the communication between victim and offender under investigation here takes place at least partly outside of the justice system, and it is the victim's perception of punishment that is of interest, I instead draw on Bilz and Darley's conception of punishment as the offender's experience of 'harm' arising from the crime (2004). This aligns more closely to lay conceptions of punishment, as illustrated by a common definition of 'to punish', which is, 'to cause someone who has done something wrong or committed a crime to suffer' (Cambridge Advanced Learner's Dictionary, 4 th Edition). This therefore includes harm or suffering which was not intentionally imposed on the offender, or harm which was imposed by someone other than a legal authority.

Participants, recruitment and data collection
Semi-structured interviews were conducted in 36 cases, involving 40 victims. The 40 interviewees were aged between 16 and 73, the average age was 42 years. Two participants identified as Asian British, two as Black British and the rest as White (36). Most of the participants identified as women (34 cisgender women), six identified as men (5 cisgender men and 1 transgender man). Just under half the participants (19) considered meeting offenders of sexual offences against them, ranging from a single incident of voyeurism to sustained childhood abuse. The other people were directly or indirectly affected by burglary and theft (8), non-sexual physical assault (8), harassment (1), slavery (1), fraud (1), manslaughter and murder (2). All the crimes had been reported to the police except one (Faye). The RJ process took place after the offender had been cautioned or received a community sentence (7 offenders), or a custodial sentence (28 offenders, mean duration of 7 years). There was no formal mechanism by which the RJ process could influence the offender's sentence, although the parties may have believed that offenders could use it to demonstrate good character at later parole hearings, and this may indeed have been the case.
The participants were referred for the research by facilitators from a charity commissioned with RJ provision across a single police force area in South-East England, with one additional case referred by a Youth Offending Service. Facilitators were instructed by their manager to inform all victims about the research, unless there was reason to believe it would be inappropriate. The decision to offer participation to service-users was therefore left to facilitators' discretion, and unfortunately no records were kept of the number of service-users who declined. Fifty-seven people gave the facilitators verbal consent to participate in the research, but 17 subsequently changed their minds. The names used below are pseudonyms and identifying details have been removed.
The study had university ethics approval, and support was offered to participants in case they found the interviews distressing.
All participants had at least considered communicating with the offender. The extent of the communication then ranged from none (e.g. in one case the victim only had a conversation with the facilitator about the possibility of communication with the offender) to multiple meetings between the victim and offender. Details of each case can be found in Supplemental Materials. Interviews with victims before and after communication with the offender have been labelled Time 1 (T1) and Time 2 (T2), respectively. 2 Ten people were interviewed only at T1, 14 were interviewed only at T2, and 16 were interviewed at both time points. T1 interviews took place on average just over 8 years after the offence, after the victim had been referred to the RJ service but before communication with the offender had taken place. The T2 interviews took place on average just over 10 years after the offence, after communication with the offender had taken place or when the offender had declined to participate. 3 The sample size was not pre-determined but was a result of collecting the maximum amount of data in the allotted timescale: April 2016 to September 2018. The interviews lasted on average 55 min, ranging between 15 min and 2 h. Most were conducted in person, eight were conducted by telephone. The interviews were digitally recorded, 4 and transcribed.

Analysis
This study was part of a larger research project, whose aim was to understand the psychological changes expected and experienced by victims who communicate with offenders. Participants were asked broad questions about the crime and the offender, such as 'How has the crime affected you?' At T1 participants were asked about their desire to communicate with the offender, for example 'How would you describe your expectations of the process?' At T2, they were asked questions about what had changed, especially regarding any expectations they had mentioned at T1. I did not set out to investigate participants' attitude towards punishment, and so the only questions in the interviews specifically about punishment (or related words and concepts such as revenge and retribution) were supplementary questions after the concept was first spontaneously raised by the participant. Braun and Clarke (2006) note that thematic analysis is not constrained by a particular theoretical approach, and it is possible to combine 'top-down' analysis, that is, a deductive, theoretical approach, with some degree of 'bottom-up' analysis, that is, an inductive, data-driven approach. Accordingly, my purpose in the larger research project was both to test existing theories and to build theory from participants' reflections. Using the qualitative data analysis software Nvivo, I identified and analysed participants' expectations and experiences of change. I had not anticipated that victims would expect and experience a change in their perception of the offender's punishment through communication with them, and so the theme that is the focus of the current study emerged entirely from the data.

The relationship between victim perceptions of punishment and communication with the offender
In support of the claim that communication with the offender is opposite, or an alternative to punishment, some victims in this study appeared to believe that communicating with the offender involved sacrificing punishment objectives. Beatrice and Carol spontaneously wrote to the judge to ask for lesser sentences, saying they wished to communicate with the offender instead. Faye explicitly told the offender she would not report the crime to the police in exchange for a meeting. Razik and Brenda suggested that meeting the offender prevented them from exacting some form of personal retribution. These victims were willing to exchange seeing the offender suffer appropriately (where 'appropriate' could be determined by a legal, moral or emotional judgment), in return for the chance to communicate.
In support of the claim that punishment and communication with the offender are simply independent aspects of the justice process, other participants appeared to make no connection between their choice to communicate with the offender and the offender's punishment. Lydia, Megan, Marie and Naomi did not mention punishment in relation to communication with the offender at all. Tasha, Dorothy and Michelle said that their motivation for communicating with the offender was in part to express interpersonal forgiveness, but they clarified that this was not connected to any wish for the offender to receive a lesser sentence or to suffer any less. Rachel, Owen and Philip explicitly said that the offenders' sentences were sufficient, and that the purpose of communication was unrelated to punishment.
Some victims appeared to fulfil their punishment objectives through communication with the offender; more than half the participants expected that communicating with the offender would increase their satisfaction with the offender's punishment or reported afterwards that this was in fact the case. Participants' initial dissatisfaction with the offender's punishment fell broadly into three camps: guilt that the state-imposed punishment was excessive or harmful to the offender, anger that the punishment was too lenient and some people who were concerned on both counts. I describe each of these views in more detail below, along with the changes victims expected or experienced that led to increased satisfaction with the offender's punishment.

What changes did victims expect and experience through communication with the offender?
Concern about leniency. Some participants appeared to be primarily concerned that the state-administered punishment the offender received was too lenient, and they tended to feel angry as a result. Hannah, Mike, Oliver and Gemma implied that communicating with the offender enabled them to get feedback about the 'pains of imprisonment' (Sykes, 1958). Hannah, who was sexually abused as a child by her mother's partner, described meeting with the offender as an opportunity to witness his punishment. She said it was satisfying to have met him in prison, and then been the person who was able to leave: 'I needed to see that I've put him there…. And I was able to walk out and go home'.
She described feeling satisfied that he was suffering, while implying that such feelings were somewhat unacceptable: He's not a man anymore, he's dead in his eyes, there's no person there anymore. He's just a body.
Interviewer: How did it feel seeing that?
It feels really horrid to say… but it felt awesome.
Mike seemed pleased to learn that the offender's experience of prison was at least somewhat equivalent to his own experience of slavery and forced labour: 'You take his freedom, away from his family, and then he realises what other families and victims are going through'. Oliver, who met the man who stabbed him more than 17 years after the incident, explained that, 'to know that he wasn't having the time of his life in prison did feel good for me', and Gemma, whose son was physically assaulted resulting in serious injuries, similarly said that she felt less angry because the offender 'didn't look like he was having a good time [in prison]'. Gemma's views of the sentence also changed after she visited a prison museum, saying 'it was awful in there. But it was good to see'.
Some victims derived satisfaction from learning that the offender had suffered beyond the punishment imposed by the state. Bridget, who was raped by a stranger almost 20 years previously, said, 'This sounds really horrible, but I kind of got a bit of satisfaction as well, and power back, about how much his life had been destructed (sic) by it all coming up again', and 'sorry I feel like a right bitch for thinking it'.
Francis and Nita sought to remedy their belief that the sentence was too lenient, but meeting the offender failed to alleviate their concerns. Both seemed to expect the offender to recognise that the sentence was too lenient, but the offenders were unwilling to do so. A victim of a hate-motivated violent attack, Francis said: 'they're out in 5 minutes, realistically, for what they've done'. The offender conveyed that his time in prison had been very hard, but Francis said this brought him no comfort and he continued to feel that the offender's suffering was disproportional: How can somebody make another person feel belittled, undermined, battered, assaulted, never be the same person again. And you feel like you're hard done by because you got two years?
Francis appeared to manage his feelings of dissatisfaction by reminding himself of the offenders' punishment: When I'm out with the dogs, and I'm up in the countryside, I do say to myself "I've got this. They've not got this". Nita, a victim of voyeurism, also remained unsatisfied with the offender's punishment after she met with him, saying, 'he has got away with it' because the community sentence was too light, 'thirty-five days of sitting having cups of tea, chatting'. Knowing the offender was an American citizen, Nita had insisted in the meeting that his sentence would have been heavier if he had committed the same crime in the USA. This developed into a technical conversation about the legal system in both countries, which Nita described as confrontational and unsatisfying.
Several of those who were concerned that the state-imposed punishment was too lenient indicated that the communication might itself constitute a punishment (Sam, Denise, Bridget, Nita, Gemma, Hannah). Nita said at T1 'I want him to feel a bit of what I felt'. Although she was unable to extract agreement from the offender that he should have got a more serious sentence, she said, 'he may have found [the RJ meeting] hard' and this seems to have been a source of satisfaction. Hannah said she could see that the offender was nervous in the meeting as he was wringing his hands, and she was pleased to hear through the facilitators that 'apparently [the meeting] destroyed him for about a week after'. Bridget stated that her renewed sense of power derived partly from her personal involvement in making the person who raped her feel uncomfortable as part of their meeting: Because he must have been as nervous as me… So, to have him wait, have him waiting, worrying, to take my time going into the room, that kind of gave me a bit of power.
Gemma described meeting the person who attacked her son as 'going face to face', and said, 'I just wanted to rub it in. That he's missed a year and half of his kid's life, for being an idiot'. She also seemed satisfied that the offender seemed to suffer during their meeting, saying, 'I thought he was going to start crying, and that was the best bit for me. It sounds evil, but that was the best bit'. Although none of the many victims who told the offender about the impact of the crime suggested they did so to make the offender suffer, Zoe articulated the possibility that describing the impact could also induce suffering, whilst explaining her reasons not to do so: [The facilitators] asked me to go through what had happened to me afterwards, and I didn't go into too much detail. Because he'd already suffered… I didn't think the gruelling details were necessary. This raises the possibility that some aspects of victims' communication with offenders might not be immediately recognisable as punitive but nevertheless could contribute to the fulfilment of victims' punishment objectives. Sam, for example, who was sexually abused by her father in childhood, said she wanted to forgive him but described forgiveness partly as a punishment: The worst thing to do [for the offender] is actually to tell the person face-to-face that you forgive them. Because that's you releasing their control, and they've got no control of that. …Because it's the last thing they think you're going to do. Especially sex offenders. Finally Denise, who had been a victim of harassment by two teenage offenders, was unhappy that a meeting with her had not been compulsory, and she described it in terms of the offenders having evaded punishment. Denise said 'I still think that she got off very lightly…. I was hoping that she would actually have to face me'. She further explained that in being given the choice, the two girls involved 'weren't made to do something they didn't really want to do'.

Concern about excessive punishment
Some participants appeared to be primarily concerned that the sentence the offender received was excessive and they felt guilty or responsible (e.g. because they had called the police). Unlike the group who were willing to substitute punishment for communication, these victims did not wish to see the offender's punishment reduced. Rather, communication with the offender enabled a change in these victims from concern about the offender's punishment to satisfaction with it, much like those who were concerned that the punishment was too lenient.
Some victims became more satisfied that the punishment was not excessive when they received feedback that the offender was unremorseful and/or unwilling to meet them (Lisa, Kaitlyn, Terri, Casey, Willow, Rose). These victims appeared to infer from a lack of remorse that the offender was more culpable than they had previously thought, and thus they were able to conclude that the punishment was not excessive but was proportionate. For example at T1 Rose, who had been a victim of sexual violence at knife point, was motivated to communicate with the offender to help him: I am for [meeting him] mainly from his point of view. I think I've got closure as much as one can… I thought if I say I forgive youbut I don't want to see you againthat it will help him in his recovery.
However at T2, after the offender declined to meet with her, she said she stopped concerning herself with the extent to which he might be suffering and instead expressed satisfaction that he got the highest available sentence: I think I'm quite -pleased is the wrong word, I can't think of the word -that he actually got the highest sentence he could get. So, the judge realised.
Others gained reassurance that the sentence was appropriate through finding out that the offender had not been harmed by their sentence (Barbara, Kathy, Zoe). Kathy said she had felt guilt about involving the police, after a man who had been her friend almost killed her by asphyxiation. After meeting with him, she felt less guilty and said she was pleased that being in prison had helped him: He's given up smoking, you know, and hopefully will stay off drink and drugs and one thing and another. So yeah, he kind of implied that that's what he needed.

Concern about both excess and leniency
In some cases, participants were concerned that the sentence the offender received was disproportionate, and they seemed to vacillate between guilt that it was excessive and anger that it was too lenient (Emma, Karen, Sadie). Emma, who was a victim of serious fraud by an intimate partner, said one reason for wanting to communicate with him was to try and find out why his sentence was so short in comparison to the harm she had experienced: So that has totally rocked me really. …[That he served] five months of a 20-month prison sentence, I don't think is justice. You know, I, I'm still fighting, financially, for the shit that he's left me in. (Emma) However, she also expressed empathy for the offender, and said she wanted to explain why she had caused him the 'pain' of prosecution.
Similarly, Karen, who was burgled by a long-term lodger, said she derived satisfaction from reminding the offender of the consequences of his actions (or as Gemma called it above, 'rubbing it in'): I kind of said to him, "Well look where it's got you -I'm going home, where are you going?" I kind of said it like that.
Yet at the same time, Karen was concerned about being perceived as 'the wicked witch of the west' for reporting to the police someone for whom she had caring responsibilities. She said she had explained to the offender at their meeting her reasons for calling the police, and she consequentially felt less guilty about it.
Finally, Sadie, whose father raped her sister, said she did not know how to make sense of the situation and was concerned about what might be happening to her father in prison: I didn't understand. I didn't know what was happening to him in prison. And I was hearing a lot of "oh he'll get accidental hot water poured over him" and I thought, I don't want him to die! … You either get they're playing x-boxes, or they're being beaten black and blue.
This position encapsulates the range of feelings victims had about the offenders' punishment and highlights the common desire to feel that the punishment was neither too lenient nor excessive but proportionateaccording to their own understandings of proportionality. At T2 having received a letter from her father, Sadie seemed reassured to have heard that he was not being 'beaten black and blue', but reiterated that his loss of liberty and small deprivations meant he was being punished sufficiently. Throughout the interview she appeared to be processing aloud the extent and meaning of her father's punishment: Not being able to go outside, and not being able to see his loved ones, and not being able to go to bed with his girlfriend at night and waking up in the morning together, or going for a walk. I think that is such a punishment.

How communication fulfilled victims' punishment objectives
Having collated in the previous section the changes in victims' perceptions of punishment that were wrought by the victim-offender communication process, in this section I revisit the cases and identify three primary components of the RJ processes which led to increased satisfaction with the offender's punishment: receiving information, receiving feedback and imposing punishment.
Some victims' desire for punishment was fulfilled through receiving information about the offender's sentence. In most cases, victims had already received at least some information about the offender's sentence through the criminal justice system, so this was not the dominant method by which communication with the offender fulfilled victims' punishment objectives. However, several victims indicated that they were at least partly motivated by receiving information about the offender's sentence, or about the process by which it was decided upon (e.g. Emma, who wanted to know why the offender only served 5 months of his 20-month sentence). This is consistent with decades of research suggesting that victims consistently want to be better informed throughout the criminal justice process (Boom and Kuijpers, 2012) and that informational justice is key to victims' perceptions of justice in a variety of contexts (Kernan and Hanges, 2002). While the RJ process provided information for some victims that they would not have been able to acquire elsewhere, others should have been able to access the information they sought without communicating with the offender, for example in the UK victims have a right to ask questions of the police or the Crown Prosecution Service (Ministry of Justice, 2020). Some victims also access information about punishment from other sources (e.g. Gemma's visit to a prison museum).
Many victims' desire for punishment was fulfilled through receiving feedback from the offender. This was primarily feedback about the state-imposed punishment; what it was like (e.g. Sadie), and how the offender was responding, for example knowing that they were 'not having a good time' (Oliver), or that the sentence had helped them reform (Kathy). Some victims seemed to consider seeing the offender in prison (e.g. Hannah, Mike) or after they had been in prison (e.g. Zoe), to be visual feedback about the offender's response to the punishment. Others received feedback about the extent to which the offender had suffered outside of the criminal justice system (e.g. Bridget), or received feedback about the offender's attitude that changed their perception of the offender's punishment, for example those who surmised the punishment was justified because the offender was unremorseful (e.g. Rose). Laboratory studies in psychology have shown that victims benefit from receiving feedback about punishment from the offender because they seek evidence of change in the offender's attitudes and behaviour (Funk et al., 2014;Gollwitzer et al., 2011;Gollwitzer and Denzler, 2009;Sarin et al., 2021). While it was beyond the scope of this article to investigate whether victims desired feedback about punishment primarily for retributive or consequentialist reasons, this study confirms its importance to victims outside the laboratory in the context of a range of real offences, including very serious ones.
Finally, some victims' desire for punishment was fulfilled through imposition of suffering during communication with the offender, for example either through the meeting itself (Nita), or some aspect of the dialogue (Gemma, Karen). In the context of such serious offences, imposition of the hardships described here may appear almost ludicrously minor; a victim of rape taking satisfaction in making the offender nervous by having to wait a few minutes (Bridget) or a victim who suffered for many years from the effects of sustained childhood abuse being pleased that the offender was upset for a week (Hannah). Nonetheless, the participation of victims in the offenders' punishment poses important questions, to which I turn in the next section.

Discussion
Most contemporary criminal justice systems are designed precisely to prevent victims from taking punishment into their own hands (Ashworth and Wasik, 1998). Therefore, the thoughts and feelings of some of the victims in this study pose a dilemma. None wanted to meet the offender with the proverbial 'baseball bat' or mete out any extreme punishment, supporting research showing that victims tend to be no more punitive than the public (Doak and O'Mahony, 2006;Maruna and King, 2009). Nevertheless, even if the punishment is not severe, could victim participation in its delivery undermine the principles that criminal justice systems are built upon? Could acknowledgement and acceptance of such punitive attitudes within RJ result in a slippery slope towards vigilantism? I argue that there will always be some degree of victim participation in offender punishment, and that therefore the acknowledgement and acceptance of punishment within RJ processes is necessary both to mitigate the risks involved, and to enhance the process for all concerned.
Victims' sense of participation in the offender's punishment has survived in two environments explicitly designed to eliminate it. Within the RJ process, interviewees in this study described varying degrees of participation in the offender's punishment, despite the context of a self-proclaimed 'non-punitive' process. Even more tellingly, within the confines of the UK criminal justice system which takes full ownership of prosecution and sentencing, many of the victims in this study were acutely aware that offenders may not have been punished at all if they (the victim) had not called the police, pressed charges or given evidence.
If we accept that some degree of participation in the offender's punishment is likely or even inevitable, we might nevertheless want to ask whether acknowledging and accepting it within RJ could encourage descent into vigilantism. There is, for example, a risk that some victims would interpret this as a licence to be vengeful or cruel in meetings with offenders. However, like performing surgery in the dark, drawing a boundary between the acceptable and unacceptable levels of victim participation in offenders' punishment without explicit recognition and investigation of its nature, seems to be by far the bigger risk. It would be highly arbitrary, for example, to declare the boundary to lie between victim participation in the formal criminal justice system and participation in informal RJ processes, especially in light of evidence that court processes can be harmful to both victims and offenders (Herman, 1992), and RJ processes can be beneficial (Poulson, 2003;Umbreit et al., 1994). Instead, we should monitor the specific types and extent of participation which are beneficial or harmful to the parties involved, and then make carefully considered, evidence-based decisions about where the boundaries should lie.
Acknowledgement and acceptance of punishment within RJ may also mitigate some of the risks arising from power imbalances among participants, an important area of concern about RJ practice (Delgado, 2000;Gavrielides, 2014;Herman, 2005;Willis, 2018). Delgado (2000: 752), for example, worries that 'a vengeful victim and a middleclass mediator will gang up on a young, minority offender' and 'exact an apology'. Willis (2018) notes that participants from different socio-economic and ethnic groups have different styles of communication, and RJ may exclude those who do not communicate in a certain manner. Under cover of a 'non-punitive' process, some victims may be given free rein while others are excluded by facilitators who pronounce them 'too angry'. Acknowledging the punitive elements of victim-offender communication would enable RJ facilitators to be more accountable about their decision making, thus somewhat protecting against the impact of individual and systemic biases.
As well as mitigating against some of the risks, acknowledging and accepting the punitive elements of RJ could potentially also significantly enhance its delivery. Perhaps one of the most well-known critiques of RJ was put forward by Acorn, the main thrust of which is encapsulated in the title of her book Compulsory Compassion (2004). While some argue that she unfairly generalises about the field (Braithwaite, 2006;Ribet, 2006), some victims do believe they must abandon their desire for punishment to participate in RJ. Take the victims in this study, for example, who thought it unacceptable to express satisfaction with the offender having been punished. Bridget apologised in the interview for being pleased that her rapist had suffered, and Gemma said 'it sounds evil' that she was pleased to see her son's attacker cry.
Victims might be more willing to communicate with offenders if they were made aware that it does not require them to abandon their punishment objectives. Moreover, honest discussions about victims' punitive motivations would also enable both victims and offenders to make better informed decisions about whether to participate. They would embark on the process better prepared, with realistic expectations. For example, Francis and Nita in this study wanted agreement from the offender that the sentence was too lenient, but they had not clearly articulated this desire before the meeting. The offenders, though remorseful, were unwilling to agree that their sentences should have been longer. More explicit discussion of Francis and Nita's punitive motivations and the offenders' responses in advance of the meeting may have enabled them to anticipate the extent to which their punishment objectives would be fulfilled, and to derive the most benefit from the meeting.

Limitations
Serious and sexual offences were overrepresented in this study. This means it complements previous research on RJ which has often been conducted with victims of minor crimes, but it makes comparison harder. Broadly speaking, victims in this study who saw communication with the offender as a means of fulfilling punishment objectives had experienced a range of crime types. However, only quantitative analysis of a larger data set could assess the effect of potential moderators of victims' perceptions of punishment, such as the severity of the crime, the degree of contact between the offender and victim during the crime, the offence being of a sexual nature, and victims' previous relationships with the offender. Without quantitative analysis of a bigger data set, we also cannot generalise these findings to establish the proportion of victims in the wider population who might communicate with the offender to fulfil punishment objectives. Nevertheless, this study constitutes evidential proof that at least some victims expect and experience changes in their perceptions of punishment through communicating with the offender, and it has documented some of the ways in which this occurs.

Conclusion
Nils Christie famously argued that state criminal justice systems have 'stolen' conflict from the stakeholders, and that we might consider returning it to them (Christie, 1977). We must acknowledge that if conflict is 'returned' to victims, their desire for punishment is likely to become apparent during victim-offender communication. This is neither a reason to abandon RJ altogether, nor is it a reason to abandon the limiting principles of our justice systems and give victims free rein to punish offenders. Instead, it highlights the importance of openly discussing the spectrum of victims' perceptions of punishment and the resulting behaviours, investigating their effects on all parties involved, and explicitly deciding where the boundaries of acceptable participation should be drawn. The boundaries, of course, are likely to differ from one case to another, and will often be negotiated with the parties involved, but they should not be left to chance.

Declarations of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding
The authors disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This work was supported by the Economic and Social Research Council [ES/J500112/1 and ES/V011626/1].

Supplemental material
Supplemental material for this article is available online.