• 1 Introduction

  The problem Lode Walgrave sets out to solve in the stimulating article he has published in this issue of The International Journal of Restorative Justice has to do with the fundamental meaning of restorative justice and how that term ought to be defined for the purposes of research.
  It is clear from his text that Walgrave is a proponent of restorative justice as well as a scholar of it. He views restorative justice – ‘the pursuit of justice through reparation’ – as innovative and normatively attractive. He sees it as a forward-looking, constructive vision of criminal justice, different from, and preferable to, the punitive and welfare approaches that dominate existing criminal justice systems; as a critical alternative, constantly challenging conventional ways of doing justice in the wake of a criminal offence; and as a vision already partly realised in various experimental practices in some progressive penal institutions and on the margins of criminal justice.
  His analysis of restorative justice’s meaning is a critical, ‘reductive’ one that aims to cut the concept down to size, insisting on a meaning that is narrower than many currently circulating. But this reductive account is intended constructively: his aim is to make the term more credible, more serviceable and thereby to advance the cause of restorative justice. It is a recalibration aiming to strengthen the challenge that restorative justice poses to conventional criminal justice by making that challenge more amenable to empirical research.

• 2 Explication

  For Lode Walgrave, restorative justice is a way of doing justice that aims at restoration, reparation, and repair following the disruption of a criminal offence. It seeks to repair the harms caused by the violation; to reconcile those who were party to the offence; and, through these restorative processes, to reinforce community, solidarity and social relations.
  Walgrave observes that, over the last 50 years, interest in restorative justice has exploded. More and more programmes have appeared in more and more countries, in the realm of juvenile justice, problem-solving courts, diversion programmes, prisons and probation orders. Outside of criminal justice, restorative justice has become an important theme in educational settings, in the workplace and increasingly in debates about the proceedings of international courts and tribunals.
  But this explosion of interest brings new challenges and dilemmas. According to Walgrave, all this talk about restorative justice has resulted in a situation where ‘the limits of the restorative justice field have become unclear’. Some continue to view it, as its original proponents did, as ‘a way of responding to crime’ and as a critical current running alongside conventional criminal justice that aims to move the state system towards a more positive, more dialogic, more relational mode. But others have come to view it more broadly, variously thinking of it as: (i) a form of dispute settlement that can be applied in settings such as schools, workplaces and neighbourhoods; (ii) a form of justice that can shape the proceedings of international courts of human rights, truth and reconciliation proceedings and all forms of transitional justice; or (iii) as a value or ideal integral to the pursuit of structural change, reparations and social justice, more generally. As restorative justice has broadened out and scaled up, its meaning has become less clear. Walgrave argues that this more ambiguous, all-things-to-all-people usage undermines restorative justice’s credibility. But he also believes – and this is his main point – that the concept’s new ‘elasticity’ is ‘detrimental to research’, his assumption being that empirical research on restorative justice will produce useful, shared knowledge only if researchers agree about how to define their object of analysis. His article is an attempt to improve the study of restorative justice by respecifying the concept’s meaning.
  Walgrave describes himself as an academic, a scholar concerned with conducting dispassionate social scientific research. He acknowledges the vital role of practitioners in developing the institutions and the ideas that have come to be known as restorative justice, pointing to victim-offender mediations in Canada and Māori-inspired family group conferences in New Zealand and Australia and praising the idealism of activists who continually press to expand restorative justice initiatives. But the role of academics is, as Max Weber observed, to balance passion with perspective: to provide dispassionate, evidence-based analyses, placing unique events in the context of more general facts and grounding subjective hopes in a more objective and reliable knowledge base.
  Walgrave talks of an ‘action-reflection cycle’ – a dialectical process in which activists pose questions for academics and academic research, in its turn, informs and furthers action. Academic research provides activists and practitioners with empirical evidence about the results of their activities, and with conceptual elaboration and refinement of their ideas. This information then helps steer the course of practice and enables activists to pursue their normative vision more effectively. Harnessing science in this way can also improve restorative justice’s credibility, presenting it as an evidence-based discipline rather than an idealist pursuit. In his contribution to this special issue, Walgrave is making a conceptual argument that, he hopes, will advance the work of restorative justice activists by improving the quality of knowledge on which they base their practice.
  Walgrave’s aim is to clarify restorative justice as an object of analysis; to make the concept of restorative justice more precise and more determinate – and thus better able to guide empirical research. Of course, this is not really a ‘clarification’: it is an argument. Nietzsche taught us long ago that any concept that has a history, that has been the subject of struggles and disputes, is incapable of definition. And Lode Walgrave realises this. Nobody, he says, can claim ownership of the concept: there can be no ‘final peremptory definition’. Nevertheless, he appeals to the community of restorative justice scholars – the ‘we’ occasionally invoked in his article – to reflect together on what they have been doing, to exchange ideas and opinions, to clear up misunderstandings and to clarify the range of approaches.
  Restorative justice has increasingly become a contested concept because the term’s usage has shifted over time, broadening out with successive waves of activism and scholarship, encompassing different visions of what restorative justice can be. As noted previously, ‘restorative justice’ is variously understood as a specific kind of procedure; a mode of criminal justice and criminal sanctioning; a form of dispute resolution used in contexts such as education, family and the workplace; the normative basis of transitional justice; and as a core element of transformational justice, more generally. Walgrave’s aim is to cut back this outgrowth of meanings and reassert a conception closer to the term’s original sense. In particular, he rejects the ‘all-encompassing notion’ of restorative justice that views it as a project for the total transformation of social relations, arguing instead for a more ‘clearly delimited concept focused on doing justice after the occurrence of [individual] offences’.
  Walgrave acknowledges some possible drawbacks of his preferred approach. He notes that his definition of restorative justice (i) risks being co-opted into existing criminal justice (he does not explain why: perhaps because this version of restorative justice is formatted in the same way as traditional criminal justice processes?); (ii) does not aim to bring about large structural change (it is content to repair the local damage caused by offending); and (iii) may involve coercion (only idealists imagine that society can ever do without coercion in the face of human deviance).
  What does he mean when he talks about ‘co-optation’? Borrowing the terminology of the Norwegian sociologist Thomas Mathiesen (1974), one might say that Lode Walgrave aims to make restorative justice a ‘competing contradiction’ – an alternative to conventional criminal justice that is radical in character but that nevertheless competes on the same plane and could conceivably be adopted without abolishing the structural features of the current system. This makes restorative justice reforms more liable to be adopted but also more liable to be neutered in effect.
  Against a conception of restorative justice as a generally applicable form of dispute resolution, Walgrave argues that disputes occurring in schools, workplaces and neighbourhoods are importantly different from those that occur after a crime has taken place. (So, for example, in conflicts occurring in schools, the parties have usually been in face-to-face relationships prior to the dispute; the offence is generally of a minor nature; the institution values respect and inclusion; the dispute is not usually a matter of public concern; and the mission of the institution presses towards constructive outcomes, devaluing punishment for its own sake.) According to Walgrave, these differences mean that restorative justice processes in criminal justice are not equivalent to similar processes occurring in other settings, and he argues that the restorative justice concept ought properly to refer only to the former.¹x Walgrave does raise the possibility that one might develop separate but related concepts, such as ‘extensions of restorative justice’ or ‘restorative practices’ or ‘restorative approaches’ to capture these derivative cases. But he leaves the issue unresolved. Against a scaled-up view of restorative justice as transitional justice or as radical structural reform, Walgrave insists that different institutions have distinct functions and that we should not look to criminal justice processes – even restorative ones – to bring about structural transformations. Restorative justice is a form of correctional justice, not a formula for structural change. And other institutions are better able to bring about thoroughgoing change of this kind.
  This more modest remit for restorative justice is something that Lode Walgrave feels he needs to justify and perhaps even apologise for – and several pages of the article are devoted to this task. But he is surely right to think that criminal wrongdoing usually has a situational, interpersonal, human-scale impact – and that the urge to address root causes and unjust social structures is, in effect, changing the topic, often in a way that neglects the needs of the parties to the crime event. Dealing with root causes is different from dealing with crime events, even if both ought to be dealt with. To focus restorative justice on the harms and disrupted relations caused by a crime is not, one might argue, to diminish the movement’s ambition: it is to specify the scope and potential of a distinctive process.
  As for the necessity of coercion, Walgrave notes that some proponents of restorative justice dream of a world where the use of force and compulsion will not be necessary. But against this, he argues that even a good society will have to reserve a place for coercion in human affairs because criminal justice processes begin where deliberation and freely negotiated relations break down, making compulsion a necessary (if regrettable) part of the process. Here, Walgrave might have made the Durkheimian point that normative breaches will necessarily occur in group life – because groups make burdensome normative demands of individuals, and the latter, in the nature of things, do not always comply. In the face of such deviance, formal sanctions will generally be required – to enforce compliance but also to affirm the sanctity and imperative character of the group’s norms. Criminal laws are not merely suggestions; they are, for the most part, important norms, vital to the well-being of the community, carrying the imprimatur and authority of the group. In such circumstances, some coercion will be inevitable. Restorative justice aims to minimise such coercion and bend it to constructive social outcomes: but it cannot eliminate it altogether.
  Walgrave’s concept of restorative justice, with its focus on dealing restoratively with criminal harms, has a narrower, more delimited meaning than many competing conceptions. However, as I discuss later, Walgrave intends this ‘reduced’ conception as a strategic element of a ‘maximalist’ project that aims to make over all of criminal justice in a restorative mode.
  So what is the definition that Walgrave presents us with? Restorative justice, he says, is ‘an option for doing justice after the occurrence of an offence that is primarily oriented towards repairing the individual, relational and social harm caused by that offence’ (Walgrave, this issue; my emphasis). There are several comments to make about this.
  First of all, note that Walgrave defines restorative justice as ‘an option for doing justice’. Restorative justice is one option among others, an alternative to established ‘options’ such as retribution or rehabilitation that might be chosen in preference to them. For Walgrave, a core meaning of restorative justice is that it offers an alternative to existing modes of doing criminal justice.
  Secondly, the defining feature that makes this option restorative is an ‘orientation’. And by ‘orientation’ he appears to mean the option’s aim, its direction, the outcome to which it tends. Indeed, he characterises his definition as an outcome-oriented one.
  Thirdly, in comments on his preferred definition, he specifically rejects a process- or procedure-based definition.
  On reading this definition, I had two immediate reactions. First, I was puzzled when he labelled this as a ‘consequentialist’ definition as this seems to confuse orientation (or aim or intent) with outcome (result or consequence). Orientation and outcome are clearly different things. A practice may be orientated towards producing a specific outcome but will not necessarily succeed in producing it. So, for instance, a court sentence may be oriented towards deterrence (or rehabilitation or retribution or restoration) but may fail to bring about that outcome. To describe a definition as ‘consequentialist’ suggests that it is about consequences (outcomes) rather than aims and orientations.
  On reflection, I think the confusion may result from Walgrave having borrowed from the lexicon of the philosophy of punishment where ‘consequentialist’ approaches use punishment as an instrumental means to an end (such as deterrence, rehabilitation or incapacitation) and are distinguished from ‘non-consequentialist’ (or ‘deontological’) approaches that impose punishment for its own sake, as an end in itself. This opposition between consequentialist and non-consequentialist is a familiar one that counterposes Bentham to Kant and contrasts what Weber would call ‘purpose-rational’ action with ‘value-rational’ action. But since these contrasting ideas are not at issue in the restorative justice context, Walgrave’s use of the term ‘consequentialist’ is confusing. Perhaps if he were to term his conception of restorative justice ‘purposive’ – as opposed to ‘procedural’ – his meaning would be more effectively conveyed. In any case, I think the important point is that Walgrave’s definition highlights orientation, so if a criminal justice practice is ‘primarily oriented’ towards restoration, it should count as an instance of restorative justice even if it lacks the specific procedural arrangements that are often associated with restorative justice. And even, one assumes, if the practice fails to produce restorative outcomes.
  My second reaction concerns Walgrave’s refusal to include ‘process’ as a defining feature of restorative justice. Why does his definition foreground ‘orientation’ and downplay ‘process’ when both might seem to form key components of any restorative justice practice?
  I think he does for reasons that relate to his larger agenda. Walgrave’s overarching project is to inject restorative justice principles into the mainstream criminal justice system, modifying policing, prosecution, courts and penal institutions in ways that embrace restorative principles. The processes and procedures of these long-established institutions tend to be adversarial, or managerial or penal-welfarist: not at all designed on restorative principles. Consequently, it is highly unlikely that these arrangements – many of them guaranteed by constitutional law and foundational ideas of due process, and all of them inscribed in organisational routines – will be altogether displaced by restorative processes. By contrast, it is rather more likely that restorative values, aims and principles could be overlaid onto these existing processes, not by modifying procedure but by influencing the discretionary decisions that occur at various points in the process: when prosecutors decide whether to indict; when courts decide whether to divert; when judges consider which sentence to impose; or when prison and probation staff decide on the kinds of programmes to make available. My impression is that Walgrave prefers a definition that will recognise partial reforms of this kind and allow them to be defined as instances of restorative justice.
  In other words – and here we might recall Nietzsche’s observation that every definition is a disguised argument – Walgrave adopts a definition of restorative justice that will fit with his pragmatic understanding of criminal justice transformation. If restorative justice is to revolutionise criminal justice, Walgrave believes that it will do so gradually, grafting new values and aims onto existing institutions and procedures.²x This is indeed how structural change in criminal justice usually occurs (see Garland, 2001). Walgrave wants to make restorative justice a prominent part of any criminal justice proceeding and sanction: to use restorative justice values and aims as a means ‘to curb the punitive premise in traditional criminal justice’. His aim is to move the institutional response to offending in a restorative direction; to orient conventional criminal justice towards restorative outcomes; and to make restorative justice an intended consequence of all criminal process. (Hence his use of the term ‘maximalist’ to characterise his approach and his definition.) When criminal justice processes are reoriented in this way, they ought, he suggests, to be deemed to be instances of restorative justice – even if they lack the procedural arrangements that a full-fledged restorative justice system would exhibit.
  I should point out that Walgrave does acknowledge that the best way to pursue restorative justice is through procedures designed with that end in mind. Restorative justice process is, he says, ‘the … most appropriate … tool to achieve restoration as fully as possible’ (Walgrave, this issue). But he appears to assume, as a matter of political realism, that a purist commitment to full-fledged restorative justice is unlikely to succeed in transforming criminal justice: hence his narrower definition and his claim that this is nevertheless a maximalist strategy.
  Walgrave acknowledges the value of the full-fledged (usually experimental) restorative justice projects that currently operate in or alongside conventional criminal justice. But he thinks of them not as ‘end stations’ (i.e. the goal to which the restorative justice movement is striving and that will one day replace punitive criminal justice) but as so many means to improve victim-offender mediation and restorative conferencing, to demonstrate the reach of restorative justice, and to point up its limits.
  Walgrave’s aim in all of this seems to me to be a pragmatic one. He seeks to introduce restorative justice principles into conventional criminal justice on whatever basis is possible and to recognise practices that have been modified by these principles as being genuine examples of restorative justice so long as restoration is their ‘primary orientation’. Accordingly, he rejects the approach of restorative justice purists who aim for a revolutionary transformation and reserve the term exclusively for practices that are restorative in origin, in design, in process and in practice. Walgrave’s rejection of process-based definitions and his embrace of an orientation-based one can be understood in this light.
  Walgrave’s definition is, like all definitions, an attempt to resolve arguments about competing possibilities. It embodies his preferred answer to three debated issues: (i) Should the meaning of restorative justice be limited to doing justice in a narrow corrective sense, rather than in a broader distributive sense? (ii) Should that meaning be confined to doing justice in a criminal process, rather than with regard to conflicts and disputes more generally? (iii) Is orientation more important than process? His definition gives an affirmative answer to each of these questions.

• 3 Critical commentary

  Having traced what I take to be Walgrave’s arguments and intentions, my aim in this section is to offer a critical commentary and suggest some ways in which Walgrave’s project might be advanced. I will outline two points of disagreement regarding the purpose and nature of ‘definition’ in social science research and then offer some thoughts about how best to research restorative justice practices as they begin to infiltrate conventional criminal justice.
  To put the point more sharply than I intend, Walgrave’s definition of restorative justice seems unnecessarily exclusionary and may strike some readers as a kind of criminological imperialism. People who practice restorative justice in educational institutions, for example, might be surprised to learn that they are not, after all, doing ‘restorative justice’ – at least not in the sense that Walgrave’s definition sets out. Why make exclusionary distinctions of this kind?
  It seems clear that Walgrave wants to focus research attention on criminal justice practices – hence his definition. But why do so exclusively? Even if ‘our’ main concern (as criminologists?) is to understand how restorative justice works in the criminal justice context, surely we can learn a great deal from observing the operation of restorative justice in other settings. Walgrave’s definition is a stipulative one, designed to exclude. It says, in effect, that unless a practice is restoring relations following a criminal offence, it is not restorative justice. This rather restrictive definition aims to guide observation, moving it away from broader forms of social repair towards the specific business of restoring relations in the wake of a criminal offence. It excludes certain phenomena from consideration by defining them out of the analysis. But why exclude all this from consideration?
  Walgrave’s argument is that the characteristics of non-criminal conflicts and disputes are simply not the same as in criminal justice, so restorative efforts in these contexts cannot be brought under the same concept as genuine restorative justice. But against this, one might suggest that the fact that relations, stakeholders and contexts differ in these other settings is a reason to differentiate, compare and contrast them with restorative justice in criminal justice – not a reason to put them in a different category and exclude them from our research.
  To borrow another lesson from Durkheimian sociology, we should recall that Durkheim’s research on discipline and punishment in a classroom setting – reported in the posthumous book Moral education – taught him a great deal about the nature of punishment as a communitive undertaking, about the efficacy of punishment, and about the conditions under which punishment can reinforce authority and solidarity – lessons that could then be applied in the criminal justice system (Durkheim, 1973; see Garland, 1990 for a discussion). A theory of restorative justice must aim to explain the mechanisms by which restorative justice works; to identify the conditions that enable its success and those that obstruct it; to establish how it correlates with other circumstances, and so on. Studying the operation of restorative justice in non-criminal as well as criminal justice settings would surely further that agenda – so long as differences of context are borne in mind.
  We can acknowledge distinctions of type while recognising the shared characteristics that make all of these practices in some way ‘restorative’ if we extend the conceptual lexicon associated with the work of social restoration. We might, for example, distinguish restorative practices, restorative justice practices and restorative criminal justice practices (or restorative justice in criminal justice) – which would allow us to view the varieties of restorative justice practices as different species of the same genus. Such an approach would permit comparisons and contrasts – a major rationale of research – while respecting the different contexts and aims of the various genres of restoration.
  My second criticism concerns the nature of definition in theoretical work and how best to define a phenomenon in order to enable it to be properly observed, analysed and theorised. My reflections on the fundamentals of definition were prompted by Walgrave’s argument and some questions that it raised for me. Why, for example, should we focus the definition of restorative justice on ‘primary orientation’, as Walgrave insists, and not around ‘process’ or ‘procedure’ or some other essential attribute? And why do we have to choose? Could we not accept that restorative justice is a complex undertaking that involves multiple components or dimensions and that each of these is variably present in real-world practices that might be described as restorative justice? Indeed, if one of our research aims is to identify the operation of restorative justice principles in criminal justice, and to gauge their extent and their success, ought we not to work with a more comprehensive account of these principles?
  One can think of situations where Walgrave’s restrictive definition might blind us to important information. It may be that in a specific criminal justice practice, restoration is not the ‘primary orientation’, but the practice nevertheless embodies some recognisable element of restorative justice – an encounter between offender and victim, for instance; or an apology by the offender; or even just a victim impact statement that is read out to the offender. Should we not regard these elements as potentially restorative in nature and inquire as to their efficacy and significance in that practice? Walgrave’s definition would stipulate, in advance of inquiry, that such practices ought not to be regarded as restorative justice. In doing so, it might prevent us from learning about the restorative effects of blended processes.
  If we want to enable empirical research to observe, measure and evaluate the impact of restorative justice elements as they begin to appear in the conventional criminal justice system, ought we not to work with a more comprehensive, multidimensional definition than the one that Walgrave suggests? In order to consider this question, we need to reflect a little further on the nature and purpose of ‘definition’ in theoretical work and its relationship to concepts, theories and ideal types.
  Walgrave acknowledges the existence of several visions of restorative justice and multiple conceptions of its essential features. But he aims to persuade others to accept his definition because he believes that an agreed definition is necessary to enable empirical research and thereby generate shared knowledge and understandings. Against this, I would raise two objections. First, it seems highly likely that disagreement about the key feature, or essential characteristic of the phenomenon, will continue, despite Walgrave’s best efforts. And, secondly, it seems clear to me that such disagreement need not stand in the way of effective empirical inquiry.
  There is no need to build a consensus around a single definition that everyone adopts. Nor is there any advantage in doing so. A theory of what restorative justice is, how it relates to various institutions, how it functions, in what contexts it succeeds, and so on ought to be constructed on the basis of that research, not in advance of inquiry. Moreover, one’s definition and conceptualisation of a phenomenon can vary pragmatically, depending on the context and character of the inquiry.³x Consider how Michel Foucault’s conceptions of ‘power’ changed depending on the context in which he was studying the phenomenon. In successive works one reads power understood as exclusion or a ‘dividing practice’; power as positive, as ‘productive’; power as an element of ‘power-knowledge’; power as incitement or excitation involving ‘spirals of power-pleasure’; power as ‘action upon action’ and the ‘conduct of conduct’; power as productive of subjects and productive of truth. See Garland (2019).
  Of course, for empirical inquiry to get started, we have to identify the distinctive elements of which restorative justice is composed. But we do not have to insist on a narrow definition that essentialises the phenomenon or reduces it to one signature characteristic. A definition is not a theory. It is a guiding orientation, a first-order determination of the kind of thing being referred to. Consequently, a starting definition of a phenomenon (such as restorative justice) should aim to be capacious and inclusive; general enough to cover all the elements of the phenomenon, however partial or incomplete, thereby allowing them to be identified wherever and however they exist.
  To clarify this point, let me say a few words about ‘definitions’ as distinct from ‘concepts’ and ‘theories’. A definition limits how a word may be used. It carves out its specific reference, setting out criteria for the deployment and application of the term. Defining a term, and conceiving of a phenomenon, are two different (albeit related) undertakings. A definition picks out the phenomenon in question, distinguishing it from surrounding entities, demarcating its boundaries. It points to the distinctive characteristics of the thing, marking it off from adjacent facts and entities, separating it out from similar things with which it might be confused. So, for example, in Walgrave’s definition restorative justice is not retributive justice, and not welfarist justice; nor is it restorative remediation in a non-criminal context; nor is it transitional justice or the general pursuit of social justice. To define restorative justice is to refer to it as being in the same worlds as these neighbouring, contiguous frameworks (the world of doing justice; the movement for social justice) while nevertheless being distinct from them.
  A definition is therefore a theorisation of a sort. But it is only a first-order theorisation. It does not take us very far. It does not explicate or explain the phenomenon. A definition should serve as a means of enabling focused empirical inquiry, the results of which will, over time, change our understanding of the thing being studied. Definitions are theoretical points of departure and ought, therefore, to enable inquiry rather than artificially curtail it. To define a phenomenon is to provide a signpost that points clearly and precisely to the phenomenon in question, enabling the analyst to pick it out from the myriad other phenomena that surround it. Walgrave’s definition says, in effect, that restorative justice exists when (i) a process of justice is undertaken; (ii) following on from an offence; (iii) in a manner that is oriented towards the repair of the harms caused by that offence. It tells us not to confuse ‘restorative justice’ with seemingly restorative practices in other contexts.
  To conceptualise a phenomenon is a deeper undertaking. It involves a theorisation that goes beyond the identification of the thing’s boundaries or distinctiveness and begins to describe its characteristics, causes, condition, relationships, functioning and significance. Once these characteristics are known, theory development can occur, and general theories can take shape.
  Accordingly, a definition of restorative justice ought not to be confused with a theory of restorative justice. A definition is a signpost, a weak indication, a first-level analytical claim. A conceptualisation of restorative justice will go further, drawing on empirical data and on the analyst’s imagination and theoretical insights. That conceptualisation will, in turn, enable more targeted empirical observations out of which we generate theoretical claims of a higher order – explanatory mechanisms, conditional statements, causal claims and, eventually, a full-fledged theory of the phenomenon.
  Walgrave likens his preferred definition to ‘an espresso’, describing it as ‘short and strong’. Interestingly, the great law and society scholar Philip Selznick suggested the opposite approach when he observed that ‘social science is best served when definitions are weak and concepts are strong’ (Nonet & Selznick, 1978: 12). The point of making a definition weak was, for Selznick, to make it inclusive, enabling its conditions to be easily met. We do not want to define instances of the phenomenon out of existence, or to blind ourselves to interesting but non-standard varieties, or to miss hybrid or quasi-instances – all of which might occur if we begin with an overly restrictive definition.
  If we follow Selznick’s example for a moment, what might a weaker, more inclusive definition of restorative justice look like? And how might it differ from the one that Walgrave develops? One way of making the definition more inclusive – and better able to capture the complex field of restorative justice and adjacent practices – would be to make the definition affirmatively multidimensional. As with any institutionalised effort to do justice while upholding specific values and producing specific outcomes, restorative justice has several interacting components. And given this multidimensional character, there is no prima facie reason to single out any one of these elements as essential and ignore the others. Instead, we should aim for a multidimensional definition – or an ‘ideal type’, as such definitions are usually called – specifying each of the variable elements of which restorative justice practices and institutions are composed.⁴x Bear in mind that each of restorative justice’s components is liable to appear to a greater or lesser extent in any specific practice. Each dimension is a variable, rather than an absolute. So, for example, it might be stipulated that one component of restorative justice is the co-presence of the parties to the offence – victim, offender, affected community members, etc. But co-presence is a variable, so in studying any real-world effort to do justice, we have to ask: are all of the parties present or just the victim and the offender? Are they present in person or online? Are they co-present for extended periods on multiple occasions or just briefly once? Similarly, when we contrast ‘restorative justice’ with ‘not restorative justice’ or else to retributive justice or rehabilitative justice, these comparisons should be seen not as sharp, categorical distinctions but instead as continuums of variation, with most real-world institutions being a mix of some kind and pure instances existing only in our conceptual constructs. Beginning with such a definition will enable us to discover, in the course of inquiry, what instantiations of restorative justice look like in the world.
  A definition of restorative justice might reflect this complexity and variation by identifying all the various dimensions of the phenomenon and including them in a multidimensional definition. Such a definition might best be conceived of as an ideal type – an analytical construct used to pick out the distinguishing features of a phenomenon; present these features in a logically perfected or pure form; and then utilise that conceptual construct as a basis for identifying, measuring and analysing empirical instances of the phenomenon in question.
  What are the dimensions of restorative justice? Its components? Its constitutive elements? Its necessary ingredients? Surprisingly, Walgrave’s definition does not tell us what restorative justice involves, what it is, what its components are: it simply insists that restorative justice is a restorative or reparative way of doing justice following an offence. But if one aim of empirical restorative justice research is to ascertain where restorative justice principles (or practical elements) have been adopted and how they are operating, it would surely help to have a more informative guide to what exactly we are looking for.
  I am no scholar of restorative justice, but it seems to me that the following elements are among those we might regard as component features of a full-fledged restorative justice process:

  • ORIENTATION: Restorative justice aims to do justice following a wrongful act, not by punishing or treating the wrongdoer but by repairing and restoring the social relations and communities harmed by the wrongdoing. Its values include ‘voluntariness, safety, inclusion, dignity, respect, responsibility, accountability, truth-telling and honesty’ (Kirkwood, 2022).

  • STRUCTURE: Restorative justice usually involves a structured environment designed to enable safe, meaningful encounters between the parties and to support restorative outcomes. That environment typically includes facilitators or mediators whose role is to guide the proceedings.

  • CO-PRESENCE: A restorative justice procedure will generally involve the presence and active, voluntary participation of all affected parties and stakeholders.

  • ENGAGEMENT: The wronged person and the wrongdoer are encouraged (and enabled) to engage and communicate with one another in a personalised, respectful process of speaking and listening.

  • COMMUNICATING EXPERIENCE AND EXPRESSING FEELINGS: The wronged party should be enabled to describe their subjective experience, to express their feelings and to grieve their loss. Wrongdoers are encouraged to acknowledge their wrongful behaviour, to explain their actions and to take responsibility for the harm they have done.

  • MAKING AMENDS: Wrongdoers are encouraged to express remorse, to apologise and to undertake to make amends. In response, the wronged party may choose to accept the apology and reparation, and perhaps even forgive and reconcile with the wrongdoer, though any such forgiveness or reconciliation must remain the uncoerced choice of the person who has been wronged and not an expected part of the process.

  • DISPOSITIONS: The sanctions, dispositions and measures available to the restorative justice process are oriented towards repairing harms, making restitution and restoring relationships. Discussion of the appropriate outcomes of the process ought to include all participants and stakeholders. All parties are encouraged to express a renewed commitment to the social norms at stake in the process.

  • AFTERMATH: Wrongdoers are to be stigmatised as little as possible and reintegrated into the community to the extent possible (including provision of social services). The process also aims to learn wider lessons from the encounter, including identifying the conditions and causes that led to the wrongful act – and take steps to prevent similar wrongdoing in the future.

  These, or some similar list of elements, would form the basis for the development of an ideal type of restorative justice, which, in turn, could be contrasted with the other main ‘options for doing justice’ (such as retributive and rehabilitative justice) in a developed typology. These typological contrasts – comparative types that more sharply define restorative justice phenomena – enhance our ability to distinguish restorative justice from other forms of justice and to make sense of mixed-type practices (which are always what we are dealing with in the real world.) Once an ideal type of this kind has been drawn up, it can be used to analyse actually existing practices and establish the number and extent of the restorative justice elements that are present, their interactions with each other, and their relationship to any non-restorative justice elements that form part of the mix.
  Walgrave’s argument for a recalibrated definition in restorative justice studies emphasises the importance of such a definition as a tool of empirical research. But he stops short of spelling out exactly which research questions his definition enables. I have suggested previously that his primary research interest has to do with discovering the extent to which restorative justice principles have become established in the state’s criminal justice system and judging the effects these are having on individual case outcomes. But there are other research questions that are important for the development of an evidence-based theory of restorative justice. Let me conclude by naming some of the empirical inquiries that a restorative justice ideal type might enable.
  One pressing issue for research is ‘to what extent is this particular practice of criminal justice (or this particular criminal sanction) restorative in character and not altogether retributive or rehabilitative?’ (Walgrave, this issue) – a question that relates to Walgrave’s fundamental aim, which is to infuse every aspect of criminal justice with restorative justice values and principles. But there are other important questions: does restorative justice work? Under what circumstances? In what contexts? What elements are indispensable for success? Does it function better when dealing with smaller rather than larger transgressions or with interpersonal wrongs as opposed to crimes against society? Do participants perceive it as more legitimate than conventional criminal justice? Do victims feel coerced? Do offenders? How are communities affected? Is it less stigmatising? Is it more costly? More time consuming? What are the abuses to which restorative justice processes are vulnerable? Is restorative justice supplemental or substitutive? Does restorative justice increase or decrease the inequities of criminal justice? Can this time-consuming, resource-intensive, interactively complex process be reproduced on the mass scale on which criminal justice currently operates?
  Walgrave is surely correct in arguing that the value of our research will be affected by the definitional and theoretical starting points of our inquiries. In prompting the discussions and analyses that make up this special issue, he has already gone some way towards achieving this goal. I am grateful to him and to the journal’s editors for including me in this discussion.

References

• Durkheim, E. (1973). Moral education. New York: Free Press.

• Garland, D. (1990). Punishment in modern society: a study in social theory. Oxford: Clarendon Press.

• Garland, D. (2001). The culture of control: crime and social order in contemporary society. Oxford: Oxford University Press.

• Garland, D. (2019). Reading Foucault: an ongoing engagement. Journal of Law and Society, 6(4), 640-661. doi: 10.1111/jols.12191.

• Kirkwood, S. (2022). A practice framework for restorative justice. Aggression and Violent Behaviour, 63. doi: 10.1016/j.avb.2021.101688.

• Mathiesen, T. (1974). The politics of abolition. London: Wiley.

• Nonet, P. & Selznick, P. (1978). Law and society in transition: toward responsive law. New York: Harper Torch.

• Walgrave, L. (this issue, 2023). Concerns about the meaning of ‘restorative justice’. Reflections of a veteran. The International Journal of Restorative Justice, 6(3). doi: 10.5553/TIJRJ.000182.

Noten

• 1 Walgrave does raise the possibility that one might develop separate but related concepts, such as ‘extensions of restorative justice’ or ‘restorative practices’ or ‘restorative approaches’ to capture these derivative cases. But he leaves the issue unresolved.

• 2 This is indeed how structural change in criminal justice usually occurs (see Garland, 2001).

• 3 Consider how Michel Foucault’s conceptions of ‘power’ changed depending on the context in which he was studying the phenomenon. In successive works one reads power understood as exclusion or a ‘dividing practice’; power as positive, as ‘productive’; power as an element of ‘power-knowledge’; power as incitement or excitation involving ‘spirals of power-pleasure’; power as ‘action upon action’ and the ‘conduct of conduct’; power as productive of subjects and productive of truth. See Garland (2019).

• 4 Bear in mind that each of restorative justice’s components is liable to appear to a greater or lesser extent in any specific practice. Each dimension is a variable, rather than an absolute. So, for example, it might be stipulated that one component of restorative justice is the co-presence of the parties to the offence – victim, offender, affected community members, etc. But co-presence is a variable, so in studying any real-world effort to do justice, we have to ask: are all of the parties present or just the victim and the offender? Are they present in person or online? Are they co-present for extended periods on multiple occasions or just briefly once? Similarly, when we contrast ‘restorative justice’ with ‘not restorative justice’ or else to retributive justice or rehabilitative justice, these comparisons should be seen not as sharp, categorical distinctions but instead as continuums of variation, with most real-world institutions being a mix of some kind and pure instances existing only in our conceptual constructs.