1 Introduction

Restorative justice attempts to improve communities by engaging offenders and victims in a shared process of both repairing harms caused by criminal wrongdoing and establishing relations of mutual recognition and respect among those who were affected by the wrongdoing – what Lode Walgrave (in this issue) calls, respectively, reparation and restoration. When it lives up to its promise, victims are actively involved in the process and their specific needs are addressed, often with an apology from the offender and material restitution; and offenders try to correct things rather than being treated punitively, with harm inflicted on them for its own sake. This is more satisfying for everyone than the normal procedure of criminal justice, where victims and offenders are mostly passive observers in what often seems like a contest among lawyers.
Re-establishing a just balance of benefits and burdens is a laudable aim, and, since we ought to achieve justice not only in the context of criminal offending but in society more generally, some people have sought to extend restorative justice to other areas – primarily as a process-based approach to address all structural injustices, everything from racial disparities to economic oppression. As Walgrave puts it, some people think of restorative justice as ‘the omnipotent magical formula to resolve all social problems’ (2023a, this issue). He says that this is problematic for several reasons, mostly because the concept of restorative justice becomes so vague that specific outcomes cannot be defined. And if its outcomes cannot be defined, then we cannot assess its effectiveness through empirical research. Restorative justice is a victim of its own success.
I agree with Walgrave’s attempt to narrow the scope of restorative justice, to characterise it essentially as a way of responding to criminal offences. His argument for the thesis is mostly pragmatic: the broad definition undermines its ‘social credibility’, which slows its more widespread adoption, and it is ‘detrimental for research’, which inhibits efforts to improve it (Walgrave 2023a, this issue). In this response, I propose a different justification of his conclusion – namely, that the restorative justice approach is a mechanism of criminal justice because it achieves the aims of criminal justice. Making this argument will also reveal some points of disagreement between us, specifically regarding his rejection of retribution and ‘the punitive premise’ (2023a, this issue).

2 Retributivism and consequentialism

Historically, philosophers have taken two opposed theoretical approaches to justifying punishment: retributivism and consequentialism. Retributivists think that offenders ought to be punished because they acted wrongly. Immanuel Kant provides a clear expression of this view:

Punishment by a court … can never be inflicted merely as a means to promote some other good for the criminal himself or for civil society. It must always be inflicted upon him only because he has committed a crime (1996: 6:331 [473]).

Retributivism is backward-looking in the sense that what matters is what the offender did, not what will be achieved as a result of the punishment. Within this broad category, there are many varieties of retributivism (Cottingham, 1979; Walker, 1999). The most common form of retributivism is desert-based: because offenders did something wrong, they deserve to suffer, so they ought to be punished. Some other forms of retributivism claim that punishment ‘annuls’ the wrong (G.W.F. Hegel’s annulment theory), that it corrects the unfair advantage that offenders gain by breaking a law (Herbert Morris’s fair play theory), or that it censures wrongdoers for conduct of which the community disapproves (Antony Duff’s communicative theory). For the retributivist, suffering is intrinsically good as long as it is deserved; no further justification is necessary.
By contrast, consequentialists believe that offenders ought to be punished to produce some valued state of affairs. For the consequentialist, suffering is intrinsically bad and ought to be minimised. Causing someone to suffer can only be justified if, compared with the alternatives (including not punishing), it optimally reduces overall suffering: as Jeremy Bentham claims, ‘all punishment is mischief: all punishment in itself is evil. … it ought only to be admitted in as far as it promises to exclude some greater evil’ (1970: 158). Consequentialism is forward-looking in the sense that what matters is the consequences that follow from punishing offenders, including deterring potential criminals (general deterrence), making it less likely that actual criminals will recidivate (specific deterrence and rehabilitation), and separating criminals from potential victims (incapacitation). The different versions of consequentialism are distinguished by which consequences are valued and ought to be maximised, such as happiness (John Stuart Mill’s utilitarianism), the defence of core rights (Hsin-Wen Lee’s rights-protection theory), and fairness (Erin Kelly’s harm reduction theory). For the consequentialist, imposing burdens on people because they acted wrongly in the past – and for no other reason – is gratuitous.
These are not the only theories of punishment, but historically they have functioned as twin poles in this subfield of the philosophy of law. That is, it is hard to define a theory such that it is not ultimately backward- or forward-looking, concerned either with what we owe to the offender or what good will come from punitive treatment (or both, as in my own mixed theory of punishment [Altman, 2021]). And, although the two theories contradict one another (especially around whether suffering can be intrinsically good), many countries – the US, UK, and Australia among them – hold them simultaneously: punishment ought to give offenders what they deserve, restrain them and reduce future crimes.

3 Justifying restorative justice

Under these circumstances, what is the value of restorative practices, or why should we transform criminal justice systems in the way that advocates of restorative justice propose? In short, why is restorative justice a good thing? According to Walgrave, those who define restorative justice too broadly appeal to values such as respect, individual dignity, and inclusion, which are valuable but ‘are not the monopoly of restorative justice’. Instead, he says, we ought to identify the more specific aim of restorative justice, which is ‘to restore/repair as much as possible all harms and damages that have been caused by injustices’. That is the end, and restorative practices are a means to that end because they, rather than punitive practices, ‘facilitate (expressions of) respect and support, mutual understanding, regret, willingness to make up and willingness to accept it’ (Walgrave 2023a, this issue).
I doubt that Walgrave’s characterisation narrows the scope of restorative justice as much as he thinks it does. There are systematic harms and injustices that also need repairing, which is exactly why Jennifer Llewellyn and John Braithwaite – as referred to by Walgrave – demand that restorative justice be extended beyond the criminal justice arena. Consider, for example, the response to slavery in the United States. In 2008, the US House of Representatives formally apologised for the institution of slavery. And there are ongoing debates at the federal, state and local levels about whether to pay reparations to Black Americans to provide some measure of remedy for the exploitation of their ancestors and its long-term social and economic effects. This seems to fit Walgrave’s characterisation of restorative justice, pursued on a broad scale. One can imagine similar attempts to address other inequalities, as Llewellyn and Braithwaite suggest. Given the values that Walgrave draws upon, restricting restorative justice to criminal wrongdoing seems ad hoc.
As an alternative, I propose a principled basis for excluding restorative justice from Llewellyn and Braithwaite’s ‘extended vision’ (Walgrave 2023a, this issue). I argue that the scope of restorative justice ought to be restricted to criminal wrongdoing because it is justified on the same consequentialist and retributivist grounds as traditional punitive practices. I thus arrive at Walgrave’s conclusion, but in a different way.

3.1 Consequentialism

It is not difficult to show how restorative justice is or could be justified by its consequences. In fact, Walgrave’s own version of restorative justice is not only ‘consequential’ but consequentialist. He claims that restorative justice can better repair harms to social life than merely throwing offenders in cages. This is a forward-looking approach.
Consider one of the primary harms caused by crime: harm to the victims. There is ample evidence that, currently, victims feel alienated from criminal justice proceedings, whether they result in plea bargains or trial verdicts. Criminal offences are defined as public wrongs that violate the state’s legal norms, with reference to the offender’s action and state of mind. Victims are involved only to provide testimonial evidence about what the offender did. Victims’ interests are irrelevant.
Insofar as restorative justice addresses the specific needs of crime victims and communities, it does a better job of repairing the harm caused by criminal wrongdoing. For example, in victim-offender conferencing, wrongdoers may express remorse and ask the victim for forgiveness, and they may offer monetary compensation, perform community service that is specific to the offence, or participate in drug/alcohol abuse treatment programmes. As a result, restorative practices have been shown to improve victim satisfaction and increase offender accountability (Latimer, Dowden & Muise, 2005; Poulson, 2003; Strang & Sherman, 2003; Strang, Sherman, Mayo-Wilson, Woods & Ariel, 2013).
Restorative justice is consistent with many versions of consequentialism, which are defined by the ends that are valued and ought to be maximised. For example, restorative practices may maximise rights protections. Nils Christie (1977) has argued that crimes are best construed as conflicts between offenders and victims, that these conflicts ‘belong’ to them, and that they have a right to participate in their resolution. By allowing victims and offenders to participate actively in resolving the situation, restorative practices protect their rights. And Erin Kelly herself has claimed that victim-offender conferences can both allow for interpersonal blame and moral repair while also achieving the political aim of enforcing compliance with a shared standard of conduct (2021: 237-240). Although the impact of restorative practices depends on many factors – whether it is court-mandated, which methods are used, whether offenders are low-risk or high-risk, whether they are adults or juveniles, and so on – empirical research shows that such practices can be effective in reducing recidivism more than incarceration and probation (Bonta, Wallace-Capretta, Rooney & McAnoy, 2002; Latimer et al., 2005; Piggott & Wood, 2019; Sherman & Strang, 2007; Sherman, Strang, Mayo-Wilson, Woods & Ariel, 2015; Sherman, Strang & Woods, 2000; Strang et al., 2013; Umbreit, Coates & Kalanj, 1994; Umbreit, Coates & Vos, 2002).

3.2 Retributivism

If punishment had an exclusively consequentialist basis, this would be enough to show that restorative justice is justified as a response to crime and thus that it is best defined as a response to criminal wrongdoing, since it serves the essential function of punishment. But many philosophers (such as Andreas von Hirsch, Michael Moore, and Antony Duff) argue that retribution is crucial to justifying punishment – either the institution as a whole or its distribution to particular criminals – and social scientific research shows that, in general, people tend to be intuitively committed to the notion that punishment must be deserved: our assessments of appropriate punishments are highly convergent, considering aggravating/mitigating factors and background conditions, and assigning similar sentences for specific crimes based on what offenders did (see esp. Robinson & Kurzban, 2007). There are reasons to think that punishment should be backward-looking, at least in part (see Altman, 2021: 67-83; 2023: 198-203). If retribution is essential to punishment, however, this would seem to be at odds with the aims of restorative justice. For the retributivist, whether our treatment of offenders repairs harms and restores relations of mutual respect is morally and legally irrelevant. As a result, Walgrave (2008: ch. 2) claims that restorative justice is an alternative to the retributive response, and Howard Zehr (2015) sees them as two opposing ‘lenses’ through which to view criminal activity and our response to it.
The problem with this view is that it characterises retributivism too narrowly; it ignores the varieties of retributivism. One could easily argue that restorative practices and reparative actions – allowing victims to participate actively in the process, material restitution, and (sometimes) acts of contrition by the wrongdoers – all serve to annul the offender’s wrongdoing and correct the unfair advantage. As Jean Hampton notes,

The demand for a wrongdoer to ‘make amends’ to his victim is a retributive idea, arising from the retributive claim that repairing diminishment requires, among other things, repairing the wrongdoer’s damage to the victim’s entitlements (generated by her value). A punishment can have built into it actions or services that constitute such amends (Hampton, 1992: 1697).

Defenders of restorative justice often assume that, on the retributivist view, the suffering of wrongdoers accomplishes nothing. This is one of the problems with the consequentialist/retributivist distinction: it incorrectly implies that retributivism cannot value consequences (Davis, 2009). In fact, retributive punishment is also forward-looking in the sense that it attempts to produce justice, and that may be accomplished through restorative practices.
Expressivist retributivism in particular is consistent with restorative justice, and indeed may be better served by restorative practices than by more traditional, punitive responses such as incarceration. Duff’s communicative theory, for example, says that punishment is necessary to publicly censure offenders so that they recognise that what they did was wrong; censure has the aim of ‘repentance, reform and reconciliation’ (Duff, 2001: 107). Although prison can function as a kind of ‘secular penance’ (1986: ch. 9; 2001: 106-115), Duff also says that censure can be communicated through restorative practices: ‘criminal mediation and reparation [are] a kind of secular penance’ (2003: 53). According to my own theory (Altman, 2021), punishment of offenders expresses the individual’s and the community’s collective sense of resentment at having been wronged. This may be better accomplished through restorative justice. As opposed to criminal trials, where defendants are incentivised to deny or avoid legal culpability, restorative practices are designed so that wrongdoers can take responsibility for their actions. Accountability need not be achieved through harsh treatment but by making offenders understand the consequences of their actions – the suffering victims have experienced and the offenders’ responsibility for it. As an ideal result, offenders would take on obligations to repair the harms as much as possible (Altman, 2021: 238-253). Kathleen Daly concludes that retribution and reparation are actually ‘dependent’ on one another:

Retributive censure should ideally occur before reparative gestures (or a victim’s interest or movement to negotiate these) are possible in an ethical or psychological sense. … and both censure and reparation need to occur before a victim or community can ‘reintegrate’ an offender into the community (Daly, 2002: 60).

The supposed contrast between retribution and reparation depends on a caricature of retribution.
There is one sticking point that remains. According to Walgrave (2008: 65; 2023b: 622) and others (e.g. Boonin, 2008: 6-21), punishment essentially involves the intentional infliction of suffering. Retributivists claim that deserved suffering is intrinsically good, or good for its own sake. It follows that intentionally harming wrongdoers is a good thing. That is what Walgrave calls ‘the punitive premise’ (2023a, this issue). This seems morally problematic on its face. The intentional infliction of suffering is wrong in just about any other context or, at best, is considered a necessary evil (as it is in consequentialism) rather than a positive good. It is also contrary to the aims of restorative justice, since we cannot restore justice by producing more harm.
There is by now a voluminous literature on whether punishment must involve the intentional infliction of suffering, whether restorative practices impose suffering (intentional or otherwise), and whether restorative justice is a form of punishment or is an alternative to punishment. For his part, Walgrave is clear: ‘It is neither an alternative punishment nor complementary to punishment’ (2008: 65; see also 2023b: 623-624). Although offenders may suffer in victim-offender conferencing, reintegrative shaming, and sentencing circles, the purpose is not to inflict harm. For people like Daly (2000; 2002) and Duff (2001: 92-99; 2011), however, restorative practices are alternative forms of punishment. They define suffering more broadly to include burdensome obligations. If we subject offenders to burdens that we can foresee, then that is hardly different from intentional harm. On my view, critics such as Walgrave and Zehr characterise punishment and retributivism too narrowly. Properly construed, retributive justice and restorative justice can achieve the same thing: ‘Both serve expressive functions in communicating reactive attitudes, including resentment’ (Altman, 2021: 245). I cannot settle this disagreement here. One thing is clear: excluding retribution from restorative justice simply by making definitional claims and arbitrary distinctions begs the question.

4 Conclusion

If, as I have argued, restorative justice is consistent with – in fact, justified by – the aims of the criminal law, my argument supports Walgrave’s claim that we should define restorative justice narrowly as a way of responding to offences. This appeal to widely held consequentialist and retributivist principles provides a more solid foundation for this conclusion than his worries about ‘social credibility’ and ‘blurred research’ (Walgrave 2023a, this issue). It also justifies the idea that we ought to change existing criminal justice practices: if the retributivist and consequentialist purposes of the criminal law can be better accomplished with restorative justice, then we are morally obligated to pursue it.