1 Introduction

Looking back across the last 35 years – since landmark publications, including John Braithwaite’s Crime, shame and reintegration (1989) and Howard Zehr’s Changing lenses (1990) and the path-breaking New Zealand Children, Young Persons and their Families Act (1989), that ushered the legislative frame for family group conferencing – restorative justice has come a long way, developing rich conceptual insights and prompting innovative practices. Yet many of the ambitions and ideals held out for a fundamental shift in how contemporary societies respond to harm, conflict and victimisation remain as elusive as ever. For many, restorative justice was hailed as a new paradigm of justice with the scope to replace, or at least dramatically shrink – in what Walgrave (2003) described as its ‘maximalist’ guise – the conventional criminal justice system and its punitive logics. Restorative justice presaged a new dawn for victims of crime – so often ignored, poorly treated and further harmed by formal justice processes – and an alternative to the ever-revolving doors of punishment and retribution that lie in store for generations of marginalised peoples. As Foucault (1977) evocatively reminded us, the latent success of the carceral archipelago of penal institutions is to perpetuate itself and its own store of recipient bodies. Today, the obdurate hold of these institutions remains as prominent as ever, while the gulf between the empirical realities of what justice is and how it is experienced, on the one hand, and restorative justice’s normative theories of what is ought to be, on the other hand, remains stubbornly wide.
Commentators have long raised warnings about how the processes of incorporation and institutionalisation can warp, dilute and corrupt restorative values, principles and potential, undermining its power to elicit significant change (Aertsen, Daems & Robert, 2006; Levrant, Cullen, Fulton & Wozniak, 1999; Rossner & Taylor, 2024) and all too often ending up reaffirming the dominant logics of state power (Pali & Pelikan, 2014). In stark contrast to the abject failings of traditional models of criminal justice, restorative justice was often presented in a ‘dichotomous, oppositional picture of different justice forms’ (Daly, 2002: 72), in which restorative justice occupied a romanticised position as a superior form of justice that should stand uncontaminated by the corrosive dynamics of punishment, coercion and control, in order to elicit the range of societal benefits associated with it. Early advocates held out the promise that restorative justice provided an ambitious way of changing not only how society achieved justice after a crime but also the ways in which people build community and human relationships (Umbreit & Armour, 2011; Zehr, 1990). In this, voluntary participation and the absence of coercion became something of a sacred cow. Rather like the unwillingness of some to accept that punitiveness may play a role in restorative processes (Daly, 2003; 2013), the possibilities of empowerment through coercion, whether explicit or implied and held in the background, were summarily rejected. Coercion was conceived as diametrically opposed to empowerment – such that it could not be possible to empower through coercive methods or ones that implied constraints on voluntariness. Coercion – so evidently at the heart of criminal justice – became something undesirable to be avoided at all costs in rendering restorative justice uncorrupted.
Optimistic visions and sacred values may have brought with them a lack of awareness of the need for tempering the passionate rejection for coercion. To some extent, the movement’s evangelism elicited a certain blindness to the fact that punishment and coercion have always been necessary elements of every democratic society. While advocating for victims’ rights, initial enthusiasm made it harder to see that getting rid of punishment would have meant disposing of something that some victims may consider beneficial or necessary (London, 2003; McGlynn, 2022). Hence, associating restorative justice with practices that entailed (explicit or implicit) coercion or that sought to accommodate punitive sentiments and sensibilities was quickly and widely rejected. Only through voluntary participation was it anticipated that the offender would express sincere emotions that would have them, the victims and the community restored. Hence, the possibility of requiring or implicitly coercing offenders to participate in restorative process became an undesirable avenue. Moreover, for some the use of coercion would completely rule out the possibility of a practice to be called restorative justice (McCold, 2000). Similarly, there was an understandable reluctance to engage restorative practices where power imbalances between the parties are most acute.
However, this was not uniformly true for all restorative justice advocates. The nuanced role and place – in the background – of coercion and punishment was long recognised by some commentators. Most prominently, Braithwaite (2005) highlighted the role of coercion in his discussions about the Republican duality of ‘speaking softly and carrying big sticks’ and in his wider location of restorative justice within ‘responsive regulation’, as nested within regulatory pyramids that explicitly acknowledge coercive forces (Braithwaite, 2002b). This recognition that coercion and volition are neither absolutes nor always inversely related is important for our purposes here but is often missed in contemporary accounts. Importantly, Braithwaite (2003: 8) and others (e.g. Walgrave 2008) have argued that competing values within restorative justices can be negotiated and rendered commensurable when evaluated with reference to ‘how they contribute to advancing dominion or freedom as non-domination’. Braithwaite (2002a: 565) recognised:

We do see a lot of domination in restorative processes, as we do in all spheres of social interaction. But a programme is not restorative if it fails to be active in preventing domination.

Nonetheless, the corrosiveness or compatibility of such forces is rarely explored or critically analysed in terms of its interpersonal implications and interactional expressions rather than at the more abstract level of institutional design. The realities of institutionalisation create some awkward accommodations and complex cultural, practical and administrative barriers that need to be understood, managed and, where possible, overcome. We know that designing and communicating the offer of restorative justice to the relevant parties is intrinsically challenging (Shapland, Burns, Crawford & Gray, 2020). In designing restorative justice practices that engage with and sit within the logics of criminal justice, Braithwaite’s (2002a: 569) distinction between ‘constraining standards’ and ‘maximising standards’ is both pertinent and useful. Some values and principles should serve as bounding mechanisms that constrain and limit institutional processes – such as respect for substantive and procedural rights and due processes – while others should be maximised where possible without exceeding or breaching the constraints.
One of the findings from the somewhat fraught attempts to institutionalise restorative practices within the predominantly punitive English youth justice system in the early 2000s through referral orders relates directly to this issue. Reflecting on the findings of the evaluation of the national pilots:

Unlike most initiatives that deal with very small caseloads and remain peripheral to the coercive system, the referral order moved centre stage almost overnight. Coercion provided the capacity to move certain restorative values to the very heart of the youth justice system, and the loss of voluntariness was the price paid. The coercive nature of the referral order undoubtedly constrains the work of youth offender panels. However, the evidence … suggests to us that they were nevertheless able to engage young people and their parents in a very different, and more positive, process of communication and reasoning from that found generally in the criminal courts. Thus one of the positive lessons for restorative justice is that despite the coercive context, and possibly partly as a consequence of it, change in the direction of delivering a more deliberative process can be realised (Crawford & Newburn, 2003: 239-240).

While this early assessment, during the pilot years of the new referral order reforms where practitioners sought to maximise restorative values, may in hindsight appear overly optimistic, it nonetheless points towards the messy compromises of institutionalisation if restorative justice is to move beyond the margins (see also Crawford, 2006). It also reminds us that although institutional reform is paved with good intentions, restorative justice reforms have historically tended to raise unrealistic expectations and set unachievable goals (Dignan, 2003), which can end up disappointing many across diverse fronts (Crawford, 2007).
The challenges of how restorative justice is accommodated in relation to established criminal justice remains its Achilles heel. Evidence now shows that despite its theoretical development and the fact that restorative justice has become more a part of mainstream policy discourse (Shapland, 2014; Wood & Suzuki, 2016), its widespread use remains constrained to minor offences committed by young offenders, largely at the ‘shallow end’ of criminal justice. Fears of contamination and corruption by punitive logics have tended to leave restorative justice operating at the margins, engaged primarily in what Karp and Drakulich (2004) describe as ‘minor crimes in quaint settings’, most notably in relation to young offenders. Accordingly, some have raised alarms about restorative justice risking becoming just another good reform gone bad (Greene, 2013). Moreover, for victims, the chances of any restorative resolution remain largely dependent upon whether their offender is legally defined as a young person who is motivated to consider restorative processes. This serves as a further reminder of their marginal status in the resolution of their own matters.
In the last few decades, despite the consistent growth of interest in and debate on restorative justice throughout the world, its early goals of societal transformation have given way to more instrumental and institutional practices (Wood & Suzuki, 2016). What is more, some of the deeply held early principles have started to be adapted in the light of evidence and institutionalisation challenges. As the engagement of restorative justice within criminal justice systems has become further muddied, some commentators have interpreted restorative justice’s current prospects as an invitation to renew the way we think about it and conceptualise its core principles and the ongoing challenges of institutionalisation (Daly, 2016; Rossner & Taylor, 2024). To this, we wish to add our voice. For better or for worse, all criminal justice systems are coercive state apparatuses. So if the aim is to expand the presence and relevance of restorative justice within the criminal justice system context, we might need to create better ways of engaging with coercion. As Shapland (2014: 120) astutely notes,

The growth of restorative justice and its development into forms with a closer tie to criminal justice bring in their train, in my opinion, the need to re-examine the theoretical underpinnings of restorative justice. It is not that the insights of the pioneer theorists for restorative justice, such as Braithwaite, Zehr, Christie and Walgrave, have diminished in their force. However, their ideas concerned what had already happened by that time and were shaped by the contexts of those times – and many of these theoretical developments specifically refer to those contexts.

If restorative justice aims at creating less violent, less punitive and more fair societies, the way forward must be to improve the way it contributes to reducing societal evils associated with the specific way punishment has been historically used while moving one step further towards ‘a criminal justice system aimed at restoration’ (Braithwaite, 1996: 24). We suggest that a lack of nuance may have pushed certain restorative principles a little too far, obscuring a more accommodating ‘messy’ relationship between restoration and coercion in the criminal justice system context – in the way Nowotny (2017: 49) uses the term ‘messy’ to emphasise ‘contingency, indeterminacy, sense-making and openness to change’.

2 Voluntariness in restorative justice

Party empowerment is central to restorative values, but the extent to which this is tied to voluntariness is less clear. It may be self-evident that restorative justice practices are more effective when people enter voluntarily and that the quality of restoration will naturally improve under free cooperation from both parties (Walgrave, 2003). Similarly, where voluntariness is not apparent, motivation may be weak and accountability undermined. If restorative justice aims to protect the victim from further anger and distress, offender participation needs to come from a sincere feeling of wanting to acknowledge wrongdoing and make amends to their victim. If coercion enters the frame, victims might be at risk of being manipulated and further harmed.
Yet as Johnstone (2020) has argued in this journal, offender participation in restorative justice is rarely as voluntary as advocates tend or would like to believe. If most human actions, he contends, take place somewhere between two distant points of a continuum between fully voluntary and fully coerced actions, analysing the specific context and background conditions on which someone decides will then enable us to assess whether what that person is doing has been freely entered into. Agency and choice will be circumscribed by contextual constraints. Placed in a criminal justice context, any offer of restorative justice will be conditioned by the binary ‘choice’ between entry or exit, where exit implies that the routines of traditional criminal justice hold sway. As Johnstone (2020: 163) notes,

[T]he threat … is implicit in all ‘invitations’ to offenders to take part in restorative justice where the background context is clear to all: the alternative is ‘to let formal justice take its usual course’.

Given the array of incentives and disincentives at play across criminal justice systems and procedures, ‘free choice rarely exists in any absolute sense in relation to most forms of restorative justice’ (Crawford, 2007: 58). Conversely, we know that the ‘offer’ of restorative justice often falls on deaf ears – for victims and offenders – due to misunderstanding of restorative process, a lack of understanding of the needs of the parties, challenges of communication and a host of cultural and organisational barriers. This renders ‘making the offer’ a fundamentally important yet highly complex process. As Shapland and colleagues (2020: 212-213) note,

until the day ever comes that the general public in each country are themselves well aware of the nature of restorative justice and its potential for them, should they become victims, the task of making the offer will remain a core challenge.

Furthermore, when a system offers a restorative avenue to offenders and they welcome it, there is no real way of knowing if this acceptance comes from a sincere inner desire to accept responsibility and make amends to the victims. Nor can it be easily determined if this is just the result of the fact that the alternative option for the offender is facing the conventional criminal justice and its machinations. Although a genuine embrace of all the values of remorse, accountability and restoration might arise among some offenders in some instances, it is highly unlikely that offenders will elect to participate exclusively based upon a sincere desire to do so. Unavoidable, to some extent, their decisions will be mediated by other complex and possibly contradictory factors and forces, as will the motivations of victims.
One of the challenges for restorative justice is to seek to mobilise those higher values through the very processes, emotions, interactions and relations engendered despite, not because of, any initial motivations that the parties may hold. By placing voluntariness as a sacred entry requirement for restorative encounters to occur, we might have transformed it into one of the most determinant stumbling blocks for restorative justice development and growth. In many instances, it will be the restorative process and encounter itself, as well as the dialogue and shared emotions that contribute most directly to desired outcomes and associated sensibilities.
If we aim for restorative justice to become more widespread in mediating the harmful use of state power, we need to think about mechanisms to control and regulate coercion towards restorative aims. In so doing, we might benefit from greater experimentation, empirical knowledge and learning about where, in relation to which offences and for whom different arrangements might work best. Following Daly (2016), we should see restorative justice as a justice mechanism, one among a plurality, rather than a distinct type of justice. In so doing, we should assess and compare a variety of justice mechanisms – ‘residing on a continuum from conventional to innovative’ (Daly 2016: 18). This is not to advocate for coercion as a default route into restorative justice but rather to recognise the reality of latent coercion and seek to minimise corrosive effects on restorative processes (Walgrave, 2023), while at the same time acknowledging the complexities of motivations and the limits of public understanding. This does demand thinking creatively about engendering referral processes out from the criminal justice system that do not require self-referral and that minimise some of the risk-averse features of professional cultures, while at the same time enabling informed decision-making that prepare the parties for the challenges of restorative encounters. It entails keeping coercion’s social role alive while containing its harmful implications. It means challenging the conventional ways of doing restorative justice and reducing harm from within the belly of the beast, acknowledging that with ‘no coercion, no restorative justice’ (Braithwaite, 2002b: 34).

3 Conclusion

We believe that now, more than ever, there is a need to re-engage with these debates about the compatibility and commensurability of coercion and volition with the aim of moving restorative justice further from the margins to the mainstream, as well as to explore opportunities for restorative interventions with more serious offences, notably with adults. Following Garland’s (2023: 452) earlier prompting, we need to explore both empirically and conceptually how ‘this time-consuming, resource-intensive, interactively complex process [can] be reproduced on the mass scale on which criminal justice currently operates?’ As restorative justice debates have moved further into the social realm of transformative forms of relational justice (Llewellyn, 2021), we should not lose sight of important questions about how best to grapple with the challenges of criminal justice itself and deliver on restorative justice’s potential as a mechanism of justice. In re-engaging with these important fault lines and debates, we might do well to draw on insights from complex systems theories (Boulton, 2010; Klijn & Koppenjan, 2014) in considering the feedback loops, path dependencies, non-linearity and opportunities for emergence that inform criminal justice systems and often stymie or thwart the institutionalisation of restorative justice initiatives. This would suggest due consideration should be given to, first, the inevitability of unintended consequences requiring not only experimentation and learning in a context of uncertainty, but also the (political) leadership to embrace learning through experimentation; second, greater definitional clarity in restorative justice policies and practices that grapple with conflicting intended outcomes and a caution against overclaiming across a multiplicity of intended societal benefits; and, third, the need to seek to regulate the self-organising and self-sustaining capacities of criminal justice systems that perpetuate their own dominant logics.
Whereas positive feedback drives change, negative feedback tends towards the maintenance of stability within the system and resistance to change. We know too well from the history of criminal justice reforms the obduracy of dominant logics and how well-intentioned change often goes awry. What is evident from the international research conducted into attempts to institutionalise restorative justice is that policies will not, in general, go to plan, they will have unintended consequences and must handle an often-times fast-changing and complex context. Restorative justice policymaking and practice innovations, hence, need to engage with the complex system dynamics of criminal justice processes and support mechanisms to make policies alive, able to adapt and be refined in the light of experience. Restorative justice policies are made and enacted in and through the processes of implementation rather than imposed pre-determined from above. They need to be attentive to the emergent properties of systems where apparently unreconcilable properties of coercion and volition interact. This is a very different approach to institutionalisation from the assumption that restorative justice values, principles and ideals constitute a higher form of justice that can be fashioned and implemented in a linear manner. And we contend that it is only through recognition of and engagement with these complex dynamics that we can begin to close the persistent ‘gap’ between restorative justice theories and practices (Daly, 2003).