Punishment And Purpose ~ The Theoretical Debate

Justice2.1 Introduction
In one of his essays, John Stuart Mill noted that even if we admit the legitimacy of inflicting punishment, many conflicting conceptions of justice regarding the proper apportionment of punishment to offenders come to light (Mill, 1867). This statement touches the core of what theories and philosophies of punishment are about. This chapter discusses the various ways that the State’s reaction to offending can be legitimised as well as the subsequent goals that could guide this reaction. A number of theoretical and philosophical approaches exist that consider legitimacy and goals of punishment in depth. Each approach has its own theoretical and practical problems. Although the different approaches are often mutually exclusive, there have been attempts to compromise.

The theoretical and philosophical debates on the justification and goals of punishment that have ensued, cover a vast area of social, political and legal thinking. This chapter aims at providing a concise overview of the various approaches.[i] It aims to highlight the key arguments from the most influential approaches, frequently by discussing the works of influential writers in these fields.

In Section 2.2, the relevance of philosophies and theories of punishment is discussed. In Section 2.3, the different approaches are categorised under the headings of retributivism, utilitarianism, restorative justice and mixed approaches. In the subsequent sections, 2.4 through 2.7, each category is discussed in some detail. The ideas of several influential writers are presented for illustrative purposes and different directions within each category, each with their own merits and problems, are briefly touched upon.

2.2 The need for philosophies and theories of punishment
Crime threatens our personal safety, our property and ultimately the social coherence of society. We consider criminality to be a serious and urgent national problem (cf. Sociaal Cultureel Planbureau, 1998). Our fear of crime not only stems from the direct threats it poses to us, but also from the general feelings of insecurity that result from the awareness of its existence. Crime exerts external influences on our lives over which we feel we have little or no control (Steenstra, 1994). In an era of mass communication and extensive media coverage of crime, such an awareness is inescapable.

One way of guarding the rules that keep society together and providing us with (a sense of) security, is through the institution of legal punishment: a means by which suitable and just reactions are meted out to those who infringe the rules. The institution of legal punishment has become such a self-evident and intrinsic part of our lives that we even demand a justification for its absence in cases where we expect it (Tunick, 1992). Punishment is a diverse concept. In principle, however, most people would agree on a description of punishment that incorporates the following seven features formulated by Walker (1991, pp. 1–3):
1. It involves the infliction of something that is assumed to be unwelcome or unpleasant for the recipient.
2. The infliction is intentional and done for a reason.
3. Those who order it are regarded as having a right to do so.
4. The occasion of the infliction is an action or omission which infringes a law, rule or custom.
5. The person punished has played a voluntary part in the infringement.
6. The punisher’s reason for punishing is such as to offer a justification for doing so.
7. It is the belief or intention of the person who orders the infliction, and not the belief or intention of the person undergoing it, that settles the question whether it is punishment.

These features, however, are not necessarily limited to the practice of legal punishment; they might as well characterise a parent’s reaction to a child’s wrongdoing. Kelk defines punishment as a well-considered, intentional and avoidable infliction of suffering on someone, for a culpable act that deserves blame in order to reach (a) certain goal(s) (Kelk, 1994b, p. 16). He subsequently identifies four domains in the context of which punishment is to be considered. The first is within the framework of criminal law. The second domain involves legal areas other than criminal law, such as disciplinary law, administrative law and civil law. The third context within which penal actions can be identified is in public life (i.e., on the streets, in shops, public parks). The fourth and final domain is within the framework of intermediary social groups in society, such as the family, work or school. Within the frame of reference of the present study, the terms punishment and legal punishment are used to refer to penal actions in the context of Kelk’s first domain: actions within the domain of criminal law.[ii]

As mentioned above, we expect legal punishment to be suitable and just. According to Walker’s features of punishment, it is done for a reason and those who order it are supposed to have a right to do so. But what is to be considered as suitable and just punishment? Although the institution of legal punishment is self-evident and a fact of life in the eyes of most people, the answer to such a fundamental question is not so evident and consequently, the practice of punishment needs a moral justification that addresses such questions. A justification is required because punishment itself is morally problematic (Duff & Garland, 1994). It is “a deliberate and avoidable infliction of suffering” (Honderich, 1970, p. 7). It involves actions that are generally considered to be morally wrong or evil were they not described and justified as punishment (such as depriving a person of his or her freedom) (see, Cavadino & Dignan, 1997a; Duff & Garland, 1994; Hart, 1963; Hazewinkel-Suringa & Remmelink, 1994; Sullivan, 1996). Even the very threat of legal punishment requires a justification because “it is itself the infliction of a special form of suffering – often very acute – on those whose desires are frustrated by the fear of punishment” (Hart, 1963, p. 22). Next to a justification of the general practice of punishment, we need to have consistent ideas on whom to punish, and how to punish (Hart, 1968).[iii] So when we do have a moral justification for the practice in general, what exactly do we wish to achieve when meting out punishment in concrete cases? These are the issues that are dealt with by the philosophy and theory of punishment.[iv]

The distinction between the general justification of the practice of punishment and the specific aims of punishment in concrete cases is essential for a good understanding of the different philosophical and theoretical approaches (Jörg & Kelk, 1994). Another way to describe this distinction is to separate purposes of sentencing from purposes at sentencing (Morris & Tonry, 1990). While, in the different approaches, the general justification of the practice of punishment is always a normative matter, the purposes at sentencing can be handled in a descriptive or prescriptive manner. Both types of purposes, however, are continuously subject to debate.

Different philosophical theories of punishment offer different accounts of why we punish, whom to punish and what the objectives of punishment should be. Although we do not expect judges and other officials involved in everyday practice to justify all their decisions in these terms, philosophical theories of punishment provide rationalisations for the practice of punishment in most discussions on the subject. Besides this, we expect normative accounts of punishment to form the basis of a systematic and consistent sanctioning practice. They set a critical standard against which the practice of punishment can be measured and scrutinised on a regular basis (Duff & Garland, 1994). It may be naive, however, to expect an explicit unified philosophical theory of punishment to govern both the justification of punishment and the aims at sentencing for all people involved, in each and every case. In practice, elements of different philosophies may be implicit and combined both at the level of purposes of sentencing (general justification) and at the purposes at sentencing (aims). The exact form of such combinations may be determined by eclectic considerations depending on specific characteristics of the offence, the offender, and the sentencing judge. As a result of such a gap between theory and practice, the descriptive value of any single philosophical theory of punishment for the justice system as a whole may be limited. They can play an important role, however, in the analysis of specific decisions and should continue to play the role of critical standard. Theories can (and should) bind the practice of punishment to a certain order and regularity (Janse de Jonge, 1991).

2.3 Categorisation of philosophical theories
In order to gain more insight in the variety of philosophical theories and yet narrow down the number that needs to be discussed for the purpose of the present study, it would be useful to categorise them. Several possibilities for categorising are available. A first general and useful categorisation is that resulting from the distinction between immanent and radical critics of the practice of punishment (see also Hudson, 1996; Tunick, 1992). The logic of this distinction can be clarified through an argument made by Rawls. Rawls pointed to “the importance of the distinction between justifying a practice and justifying a particular action falling under it” (Rawls, 1955, p. 3). Tunick suggests that a theorist of legal punishment is either an immanent critic or a radical critic of the practice. An immanent critic of punishment accepts the institution of legal punishment, seeks a sound moral justification for it and uses this as a critical standard against which to test the actual practice of punishment. The radical critic, on the other hand, questions the existential foundation of the institution of legal punishment. In Tunick’s words:
The radical critic in effect denies that there can be a sufficient justification for any action that is part of the practice; she concludes that the whole practice, root and branch, serves no good purpose, or perhaps a malign one. In contrast, the immanent critic might reject particular justifications that are given within the practice but accepts that in principle actions within the practice can be justified. (…) The theorist who assumes the role of immanent critic is, then, situated inside the practice (Tunick, 1992, p. 18).

From a theoretical point of view, the starting point of any philosophy concerning a social phenomenon should be radical/existential in nature. The fact that a practice exists does not necessarily mean that it is, or can be, justified in its present form, although this might have been the case in the past. The mere existence of the practice of punishment may not be used to dismiss reflection on its necessity (Hazewinkel-Suringa & Remmelink, 1994, p. 887). A radical critique on immanent theories, therefore, is that they compete with each other for the ‘best’ rationale of punishment without asking these more fundamental questions (Doyle, 1995), thereby overlooking possible alternative ways (other than penal policy) for promoting disciplined conduct and social control (Garland, 1990, p. 292). One of the consequences of competing and changing immanent rationales is confusion as to the ‘true’ rationale and meaning of the concept of punishment. About a century ago, Nietzsche pointed to this very problem when he stated that the continuous adaptation to the most varied uses has caused the concept of punishment to become undefinable. It therefore has become impossible to explain why people are punished. (Nietzsche, 1887/1994, second essay – Section 13).

There are radical critics who are not so ‘pure’ in their radicalism but who cannot be labelled as immanent critics either. The distinction between immanent and radical critics is therefore obviously not always clearcut. Although abolitionist theorists, for instance, have long-term radical goals (e.g. Bianchi & van Swaaningen, 1986; Christie, 1977; Christie, 1981; de Haan, 1990; Hulsman, Bernat de Celis, & Smits, 1986), their short-term engagement since the 1980’s is more that of immanent critics (e.g. Van Swaaningen, 1992). This, presumably, is the best strategy in order to have some chance of long-term radical achievements. Modern abolitionists aim for a transformation of criminal law in what they call reparative law (see, Bianchi, 1986); they aim to break away from the conditioned reflex that affirmation of norms should be effectuated in a punitive way (Boutellier, van Swaaningen, Lippens, van de Bunt, & Huisman, 1996). The point of departure for any transformation is (by definition) an immanent one. Immanent elements of such ‘transitional’ theories could therefore be useful in the analysis of positions within the existing system of criminal justice.

Since the present study deals with attitudes and decisions of magistrates in the Dutch penal system, the discussion of philosophical theories of punishment will be limited to those that could have some practical relevance for the analysis of the attitudes and behaviour of officials within the established system of criminal justice. Theorists who could be labelled as ‘pure radical critics’ according to the description given above, will therefore not be considered extensively in the present context. This leaves us with philosophical theories that either have an immanent or a transitional character. An important instance of such system-transitional approaches, which over the last few years has (re)gained attention among theorists and reformers, is that of restorative justice. The restorative approach tries to break away from ‘punitive thinking’ and emphasizes the importance of conflict-resolution through the restitution of wrongs and losses. The victim of a crime plays a central role in restorative justice. The immanent theories can be divided in two groups/categories of philosophical theories: retributivism and utilitarianism. Although there are several criteria possible for making the distinction between utilitarian and retributivist theories, the most prominent difference between the two groups of theories is in their temporal perspective. Utilitarian theories are forward-looking. The justification for the practice of legal punishment is found in its supposed beneficial effects (utility) for the future. This utility outweighs the suffering inflicted on offenders by the act of punishment. Utilitarian theories are, therefore, often called consequentialist or instrumentalist theories. Some authors prefer the term ‘reductivism’ as a specific form of utilitarianism because the focus is on the reduction of crime (e.g. Cavadino & Dignan, 1997a; Walker, 1985). Retributivist theories, on the other hand, are retrospective and non-consequentialist. The justification for the practice, in many retributivist accounts, is found in a disturbed (moral) balance in society; a balance that was upset by a past criminal act. The act of punishment in itself is just, deserved and morally good since it is supposed to redress that balance. A second way to distinguish between utilitarianism and retributivism is by putting utilitarianism in the class of teleological theories and retributivism in the class of deontological theories. In the teleological view, an action must be justified by its consequences; an action must serve some good in society. In the deontological view, on the other hand, actions are not justified by their consequences, but rather by their intrinsic (i.e., independent from any future consequences) moral value.

Both utilitarianism and retributivism have been called groups or categories of philosophical theories. The reason behind this is that both terms represent a whole gamut of refinements and different directions but still fit under the general header of either label. In the descriptions of retributivism and utilitarianism given below, due attention will be paid to such differentiations, although the focus will remain on the most important premises of these accounts of legal punishment.

As will be shown, elements from utilitarianism and retributivism can be combined or mixed to form so-called hybrid accounts of punishment. Although such hybrid accounts do not offer any essentially new theoretical insights, they are interesting and relevant alternatives for pure retributive or utilitarian reasoning. Hybrid accounts appeal to many because inherently one type of reasoning is moderated by the other.

Table 2.1 Schematic representation of theoretic accounts of punishment

Table 2.1 Schematic representation of theoretic accounts of punishment

Table 2.1 schematically presents the approaches that will be discussed in more detail in the following sections of this chapter. It is important to bear in mind that the following sections do not attempt to offer a complete and exhaustive overview of the various theoretical and philosophical directions. Rather, the objective is to offer a concise account of the core arguments that are considered relevant within the framework of the present study.

2.4 Retributivism
Although there are many versions of retributivist accounts of punishment, the unifying theme is that punishment of wrongdoers is intrinsically good. Justice should be done without pretensions for any future utility. Punishment has an inherent moral value as a reaction to past wrongdoing that needs no justification in terms of future beneficial effects (Duff, 1996). Questions as to where the moral necessity for punishment lies, or rather why the particular action of punishment (deliberately inflicting suffering on wrongdoers) is the appropriate and required response to wrongdoing, are answered differently by various retributivists. Frequently the answers to these particular questions can be interpreted in utilitarian (teleological) terms, thereby rendering some retributivists vulnerable to the ‘accusation’ of being ‘crypto-utilitarians’[v] (Walker, 1991). Indeed, “sometimes the differences among retributivists seem greater than the differences between some utilitarians and some retributivists” (Tunick, 1992, p. 67). For present purposes however, it will neither be necessary to settle that debate, nor to choose sides.

2.4.1 Negative and positive retributivism
An important first distinction is that between negative and positive retributivism. Negative retributivism is defined by two rules:
1. Only the guilty can be punished
2. The guilty can only be punished to the extent of their desert (moral culpability) (Duff & Garland, 1994, p. 7).

The principle laid out by these two rules is what Hart calls ‘retribution in distribution’ (Hart, 1968). Relying on this negative principle of retributivism means that punishment is not a necessary response to crime; it is permissible but only to the extent regulated by the two rules. The principle is a negative principle because its purpose is to restrict (limit) punitive action. The fact that theories of this kind are called retributive lies in their adherence to proportionality in punishment. An offender who has been found guilty, may not be punished more severely (as one might wish, for instance, with instrumental aims in mind) than the seriousness of the offence and his culpability permit. Nor may an innocent person be punished to deter potential offenders.

For retribution in distribution the general justifying aim of punishment need not even be retributive (Hart, 1968, p. 9). It is, therefore, not surprising that the principles of negative retributivism are often found in combination with utilitarian elements, for instance as a limiting (negative) principle in consequentialist accounts of punishment (e.g. Duff, 1996; Morris, 1974; 1992). Such theories are classified as mixed theories or hybrid accounts of punishment; they will be discussed in Section 2.6. Positive retributivism attempts to offer a more complete account of punishment than negative retributivism which can only operate in combination with a general justification (utilitarian or retributivist). The positive retributivist holds that ‘justice’ demands punishment to be meted out; punishment of wrongdoers is required by certain principles of justice. In the view of true positive retributivists, it is not only permissible to punish up to the limit indicated by the negative principle, it is even a duty to do so (Walker, 1985, p. 108).

The classical formulation of positive retributivism was given by Kant. Kant’s most explicit writings on punishment are found in the Doctrine of Right, the first section of his Metaphysics of Morals (Kant, 1797/1991). Kant, however, seemed more concerned with the ‘dangers’ of utilitarianism (in the form postulated by his contemporary Bentham) than with formulating a thorough and complete account of punishment. His retributive theory, therefore, is sketchy (Von Hirsch, 1992a, p. 65) and open to multiple interpretation. Kant, like many positive retributivists after him, insists that humans, rational beings capable of moral understanding, should never be treated as a means to promote some future good, neither for themselves nor for society at large. Punishment, in Kant’s view, is a categorical imperative, a moral necessity without any reference to possible consequences (good or bad). A wrongdoer should be punished because he has done something morally reprehensible, because he has committed a crime and for no other reason. In answer to the question of what kind and what amount of punishment should be inflicted, Kant refers to talionic measures (he equates his law of retribution to lex talionis):
Accordingly, whatever undeserved evil you inflict upon another within the people, that you inflict upon yourself. If you insult him, you insult yourself; if you steal from him, you steal from yourself; if you kill him, you kill yourself (Kant, 1797/1991, p. 141).

The question of why wrongdoers deserve punishment instead of some other (non-punitive) reaction to their actions remains unanswered by Kant. For the positive retributivist, the moral necessity to punish must lie in the retributive general justification for the practice. It is to this general justification, the explanation of why punishment is the intrinsically appropriate and deserved response to crime, that we now turn.

2.4.2 The intuitionist approach
There is one very straightforward, but not very enlightening, retributive general justification for the practice of punishment that relies on intuition. The argument simply is that a guilty person should be punished because he deserves it. Drawing on our emotions of love and hatred, we feel that he deserves it.[vi] Although such an argument appeals to our sense of justice and emotions of revenge, which the intuitionist retributivist holds we all share (e.g. Moore, 1987), it does not provide a clear theoretical argument as to why punishment (the infliction of suffering) is the appropriate and required response to crime. If we are to distinguish retribution from mere revenge[vii], we need objective criteria to justify it. Relying on intuition in order to justify the practice of punishment is quite reductionistic, if not a fallacy, since the question that was supposed to be answered is why a person deserves punishment (Clear, 1996; Honderich, 1970, pp. 4,15). Such a question cannot satisfactorily be answered with: ‘Because we feel that it is deserved’. Over and above, we should pose the question of where we get this intuition. Where lies the origin of this feeling that punishment is deserved? Could it be that it “is really a learned reaction to offending rather than an inborn intuition” (Walker, 1991, p. 72)? The intuitionist should then be able to show that the feeling of deservedness and the inclination to punish is a ‘natural’ feeling with which we are born. The retributivist who appeals to our intuition and collective inclination to punish as justification for the practice, can, however, not be accused of being a ‘utilitarian in disguise’ since there is no reference whatsoever to any future benefits of punishment. Although few philosophers explicitly stick to a purely intuitionist justification, any moral justification of punishment that presupposes the existence of objective moral values implicitly contains intuitionist elements that make it prone to discussion.[viii]

2.4.3 Restoring a balance
Most positive retributivists, in one way or another, refer to a balance in society that can be disturbed by the act of crime. Punishment, in their view, is the required response to offset the disturbed balance. The act of punishment is purely retrospective and has an inherent moral value (the deontological argument). One classical type of balance-restoration stems from Hegel. According to Hegel, punishment should be meted out in order to cancel the ‘negation of right’ brought about by a crime, to return to a previous state of affairs (Honderich, 1970, p. 35). Punishment, in other words, is to annul a crime.

In Hegel’s own words, punishment of an offender:
(…) is to annul the crime, which otherwise would have been held valid, and to restore the right (Hegel, 1821/1967, p. 69). Objectively, this is the reconciliation of the law with itself; by the annulment of the crime, the law is restored and its authority is thereby actualized. Subjectively, it is the reconciliation of the criminal with himself (…) (Hegel, 1821/1967, p. 141).

Although annulment may have a ritual function, as Nigel Walker points out, the fact that punishment has been meted out is, in the eyes of the victims, not equivalent to the crime not having been committed. “Victims can be compensated, but not unraped or unmugged” (Walker, 1991, p. 74). Hegel, however, was more concerned with abstract moral notions of right rather than with concrete compensation to victims in specific cases. An omission in Hegel’s account is that he leaves largely unanswered the question of how, and how much to punish.[ix] However, while pointing to the absurdity of talionic measures, he does indicate that punishment should (in some way) be equivalent to the qualitative and quantitative characteristics of the crime (Hegel, 1821/1967, pp. 71–73).[x] Furthermore, as a true positive retributivist, Hegel insists (like Kant) that humans should never be treated as a means to an (utilitarian) end:
To base a justification of punishment on threat is to liken it to the act of a man who lifts his stick to a dog instead of with the freedom and respect due to him as a man (Hegel, 1821/1967, p. 246).

Another influential positive retributivist approach views punishment as a means to restore the balance of benefits and burdens in society. Our system of rules (i.e., criminal law), the argument goes, supplies us with benefits by protecting us from harmful actions such as violence and deception. It defines a sphere for each person “which is immune from interference by others” (Morris, 1968, p. 477). In order to enjoy these benefits, everyone must exercise the burden of self-restraint over inclinations that would interfere with that sphere of immunity. Failure to exercise self-restraint would result in an unfair advantage. In Morris’s words:
If a person fails to exercise self-restraint (…) he renounces a burden which others have voluntarily assumed and thus gains an advantage which others, who have restrained themselves, do not possess (Morris, 1968, p. 477).

Punishment would be the just and required reaction to those who have acquired such an unfair advantage, because:
Matters are not even until this advantage is in some way erased. Another way of putting it is that he owes something to others, for he has something that does not rightfully belong to him. Justice– that is punishing such individuals– restores the equilibrium of benefits and burdens by taking from the individual what he owes, that is, exacting the debt (Morris, 1968, p. 478).

In 1976, Andrew Von Hirsch endorsed this theory as a (partial[xi]) account of punishment in his book Doing Justice (Von Hirsch, 1976). At the time, however, he saw one pitfall in the approach. Imposing a deprivation on the offender in order to redistribute the benefits and burdens does not explain why an offender deserves to be punished “instead of being made to suffer another kind of deprivation that connotes no special moral stigma” (Von Hirsch, 1976, p. 48). Von Hirsch escaped the pitfall by contending that because the offender has done wrong, he deserves blame for his conduct. While depriving the offender of his unfairly gained advantage, the implicit element of reprobation in punishment expresses the blame that is deserved.[xii] A sophisticated version of this account of punishment was given earlier by the Dutch philosopher of law, Leo Polak. Polak called his account an ‘objectified theory of retribution’ (Polak, 1947, Ch. VIII). Central to Polak’s approach is the contrast between the ‘objectively valid morality’ and the ‘subjectively valid immorality’ manifested by crime.

Punishment is required and considered to be just in order to equalise this contrast. To equalise the contrast, two points of reference are necessary. The first point of reference is the unfairly gained advantage by the offender: no thief should be able to enjoy that which he has stolen (Polak, 1947, p. 320). The second point of reference is the blameworthiness of the offender’s immoral (anti-social) character, which also merits punishment. By only taking away the unfairly gained advantage (reference point #1) an offender is merely returned to the subjective place he held before the crime, his status quo ante (Polak, 1947, p. 321). By also punishing his ‘subjective immorality’ (reference point #2), not only has the unfairly gained advantage been taken away, but the objectively valid morality has also been reaffirmed: the immoral will of the offender may not be held valid (cf. Hegel). Combining these two points of reference results in retribution being described as ‘objectifying harmonisation’. Although the benefits-and-burdens approach is a true positive retributive approach, it suffers some major practical problems. How should unfairly gained benefits be measured for different types of crime? Can the gravity of crimes be assessed in a fair way using this principle? Are the benefits and burdens equally distributed in society to begin with? Is there an objectively valid morality shared by everyone?[xiii]

This discussion of positive retributivism will be concluded with a brief discussion of Robert Nozick’s approach. Nozick literally goes beyond the benefits-and-burdens approach in his retributive account of punishment. Before the infliction of a deserved penalty on an offender, his unfairly gained advantage should be removed or counterbalanced (Nozick, 1981, pp. 363–364). The redistribution of benefits and burdens is therefore not an integrated part of Nozick’s conception of punishment. Instead, for the justification of punishment, he assumes a strictly normative outlook: a value-based approach. Retributive punishment is the communication of a moral message. This message is to be communicated through punishment (i.e., infliction of suffering) in order to make sure that it has a substantial effect (in some way) on the wrongdoer’s life. The objective of the message that is conveyed by the act of punishment is not the moral improvement of the offender (this would be teleological). Rather, such consequences should be seen as “an especially desirable and valuable bonus, not as part of a necessary condition for justly imposed punishment” (Nozick, 1981, p. 374). The objective merely is to connect the offender to the correct values. There is an intrinsic moral value in giving correct values some effect in the life of a person who has become disconnected, even though the person himself will never accept these correct values (Nozick, 1981, p. 377).

According to Nozick, the moral message that is delivered by punishment must be delivered in a way that matches the magnitude of the wrong or harm done. This might be interpreted as a talionic requirement. However, Nozick (explicitly) distinguishes himself from positive retributivist hardliners in that he outlines circumstances in which punishment can be refrained from, even though harm was done and the offender can be held responsible. There is no requirement for punishment in the case of an offender who, before he was captured, sincerely repents his wrongful act, has made amends to the victim(s) and lives his life doing extraordinarily good deeds. The correct values now apparently have a significant effect in his life; the person is already connected to the correct values (Nozick, 1981, pp. 384–385). Unlike Hegel, and in a certain way also unlike Polak, Nozick thus abandons the ‘objective’ component in punishment under certain circumstances. For Nozick, the celebration of objectified moral values is not a sufficient justification. A necessary condition for justified punishment is individual disconnectedness at the time of the act of punishment itself.

2.5 Utilitarianism
Utilitarians have a somewhat easier position to defend than that of retributivists who take an essentially moral stance in order to justify the practice of punishment. True positive retributivists aim to show the moral necessity of punishment while trying to avoid doing this in terms of utility. Utilitarians, on the other hand, ‘simply’ need to point out the supposed future benefits of an action in order to justify it. The future good (i.e., the utility) in the utilitarian approach to legal punishment is served by the reduction and prevention of crime: the general justifying aim of the practice. The methods available through punishment to achieve such future benefits are:

Individual and general deterrence:
When people refrain from certain actions because of their belief in possible negative consequences, we say they are deterred from those actions (Walker, 1991, p. 13). A convicted offender might be deterred from reoffending (i.e., individual or special deterrence) because, through the experience of punishment, he has suffered the unpleasant consequences of his wrongdoing. Other citizens who might be tempted to commit a crime might desist from doing so (i.e., general deterrence) from fear of the penalties which they see inflicted on convicted offenders (Ashworth, 1992a). Deterrence is based on the assumption that individuals are rational, calculating beings (see Section 2.5.1 below). Besides inducing fear of offending in the minds of the public, general deterrent sanctions are also believed to function as (re-)affirmation of norms.

Rehabilitation:
Rehabilitation, resocialisation, treatment and correction are often used interchangeably in penological literature. They refer to improving or reinstating the offender’s position in society and/or changing the offender’s personality in order to make him less prone to criminal behaviour. This is typically attempted through techniques such as counselling, psychological assistance, training of social skills and job training (Von Hirsch, 1992b). Traditionally, the ‘rehabilitative ideal’ is associated with the probation service and alternative sanctions. The contemporary conception of rehabilitation finds its origins in the ‘positive school of criminology’ which held the (deterministic) view that individuals “do not act from their own free will but are impelled to act by forces beyond their control” (Cavadino & Dignan, 1997a, p. 48). This new direction thus breaks away from the reasoning of Beccaria and Bentham for whom crime was the result of free will and hedonism of individual offenders (Lilly, Cullen, & Ball, 1995, p. 18). It is the science of the etiology of crime which seeks to identify these forces that are beyond the individual’s control. These forces can include genetic, environmental, social and psychological forces (Van Dijk, Toornvliet, & Sagel-Grande, 1995, pp. 14–15).

Incapacitation:
Incapacitation is the use of physical restraint (or ultimately death) in order to prevent an offender from reoffending. Although its effectiveness is not disputed, it can only be effective for the duration of the restraint. The choice of whom to incapacitate, and for what period of time, has to depend on predictions of dangerousness: How likely is it that a particular offender will reoffend (and how serious would that offence be)? Usually the offender’s prior criminal record is viewed as the best predictor of his future behaviour. However, if the behavioural prediction is not borne out (e.g., a person who has been predicted to reoffend does not do so), it would imply unnecessarily incapacitating a person and therefore inflict needless suffering on him. The general problem with the strategy of incapacitation is that behavioural predictions about offenders have proven to be unreliable (Gottfredson & Gottfredson, 1994).

Utilitarian accounts of punishment share the general justification of promoting the public good (i.e., prevention and reduction of crime). They differ amongst each other in their focus on the available method(s) to attain that common goal. The terms general and individual (special) deterrence are not to be taken as synonyms of general and individual (special) prevention. The former are means to achieve the latter. General deterrence is a means to achieve general prevention, as individual deterrence is a means to achieve individual prevention. Another means to achieve general prevention is the supposed educational effect of punishment on the general public in the sense that it functions as a (re-)affirmation of norms. Contrary to retribu have an inherent moral denunciatory value, in the instrumental view it is supposed to have utility in the sense of general prevention. Next to individual deterrence, individual prevention is thought to be served by rehabilitation and incapacitation. It may be clear that individual and general prevention can put conflicting demands on the mode and severity of punishment. While an individual offender might best be prevented from reoffending by treatment of his ‘deficient personality’ (i.e., individual prevention), a deterrent strategy aimed at general prevention would ask for a more severe and exemplary form of punishment, despite the offender’s ‘needs’. There is a similar conflict between treatment (i.e., resocialisation) and incapacitation: prisons are generally considered not to be the most suitable environments for resocialisation.

Utilitarianism, like retributivism, faces some theoretical and practical controversies. The most prominent of these controversies focus on the following questions: What is left of the utilitarian justification of punishment if the intended future benefits do not appear to be achieved? Do we accept disproportionally severe punishment or even punishment of the innocent if its net effect is to contribute to the maximisation of good in society? The first question has led to often heated debates on the effectiveness of penal sanctions in terms of their rehabilitative potential. This took the form of the oft referred to ‘What works?’-debate (e.g. Mair, 1991; Martinson, 1974; Palmer, 1975; Palmer, 1983; Van der Werff, 1979). Scepticism about the deterrent effects of penal sanctions also emerged (e.g. Beyleveld, 1980; Blumstein, Cohen, & Nagin, 1978). As a result of discouraging research findings in the seventies, renewed attention for retribution emerged, with an ensuing emphasis on the ‘just deserts’- model (Von Hirsch, 1976). Concerning the problem of disproportionally severe punishment and punishment of the innocent, many (modern) utilitarians have embraced the negative (i.e., limiting) principle of retributivism on humanitarian grounds, while sticking to a utilitarian general justifying aim of the practice.

The following section offers a brief discussion of two influential early utilitarians who based their theories on deterrence: Bentham and Beccaria. Subsequently the potential conflict in the utilitarian approach between individual and general preventive strategies will be highlighted by a short discussion of the writings of the Dutch publicist, Nicolaas Muller.

2.5.1 Bentham and Beccaria
Probably the most influential exponent of utilitarianism was Jeremy Bentham (1748–1832). In An Introduction to the Principles of Morals and Legislation (1789), Bentham offers an account of legal punishment based on the principle of utility; it qualifies as a deterrence-based account. In the first chapter Bentham defines the principle of utility:
By the principle of utility is meant that principle which approves or disapproves of every action whatsoever, according to the tendency which it appears to have to augment or diminish the happiness of the party whose interest is in question: or, what is the same thing in other words, to promote or to oppose that happiness (Bentham, 1789/1982, ch. I, sct. 2).

According to Bentham, all human beings are governed by the principle of utility. The principle of utility is so fundamental to all actions that it is not susceptible to direct proof, “for that which is used to prove every thing else, cannot itself be proved” (Bentham, 1789/1982, ch. I, sct. 11). For every action, a man weighs the expected pleasures against the expected pains. If the expected pains are greater than the expected pleasures, the action will be desisted. The principle of utility not only applies to the actions of individuals, but also to the actions of the government. The aim of government is to promote the greatest happiness for the greatest number in society. An action of government conforms to the principle of utility when its net effect is to augment the happiness of the community (Bentham, 1789/1982, ch. I, sct. 7). One of the tools available to the government for promoting this greater happiness is that of legal punishment. Crime produces pains and reduces pleasures in society. Legal punishment, through its supposed exemplary effects, deters crime because it is shown that the benefits of a particular criminal action are outweighed by the pains induced by punishment. However, punishment is itself an evil because it involves the infliction of pain. It can therefore only be justified “as far as it promises to exclude some greater evil” (Bentham, 1789/1982, ch. XIII, sct. 2), that is, to prevent future crimes. In order to effectively prevent offences, an important rule for the level of punishment is:
The value of the punishment must not be less in any case than what is sufficient to outweigh that of the profit of the offence (Bentham, 1789/1982, ch. XIV, sct. 8).

There is an interesting resemblance with the benefits-and-burdens approach in the retributivist account of punishment. It is important to note, however, that Bentham’s objective, in counterbalancing unfairly gained advantages, is instrumental in nature, whereas the retributive view is a moral one. Another similarity with some positive retributivist accounts lies in Bentham’s preference for punishments that share the quantitative and qualitative characteristics of the crime (Tunick, 1992, p. 73). Bentham expects greater preventive effects from punishments that ‘bear an analogy to the offence’ because they are more easily learned, better remembered and more exemplary (Bentham, 1789/1982, ch. XV, sct. 7–9). This talionic preference, however, is not seen as a necessary requirement by Bentham. It is only to be preferred when it is practical and not too expensive (Bentham, 1789/1982, ch. XV, sct. 8). Furthermore, the amount of punishment applied should never exceed what is necessary to attain its utility. If it does exceed that limit, it would imply needless suffering which does not conform to the principle of utility. The requirement that the amount of punishment given should not be less than what is sufficient to outweigh the profit of an offence, in conjunction with the limitation that punishment should not exceed what is necessary to attain its utility, constitutes a utilitarian version of the proportionality principle.

When there is no utility to be expected from punishment, it is useless and should be refrained from. Bentham defines cases (“cases unmeet for punishment”) in which he foresees no utility from punishment (Bentham, 1789/1982, ch. XIII). These cases involve persons who cannot be deterred, for instance, the insane and the intoxicated. He seems to ignore, however, the general deterrent effect that punishment might have in such cases (Hart, 1968, p. 19–20). Apart from such criticism, it is important to note that contrary to positive retributivist views where justice requires punishment regardless of its utility, in Bentham’s consequentialist view there is no inherent moral value in the act of punishment itself that would be sufficient to justify it.

An important and difficult issue that any utilitarian has to deal with is punishment of the innocent. If crimes are substantially prevented by punishment of the innocent, the benefits would outweigh the pain inflicted on innocent persons. Clearly then, there could be utility in such punishment which would therefore be justified. This point entails one of the fiercest retributive attacks on the utilitarian account of punishment. Bentham repeats several times, however, that punishment can only be in reaction to an offence; the innocent ought not be punished. Why this is so, can hardly be explained from the principle of utility. Moreover, it resembles the negative retributive principle. Bentham does not elaborate on this point. However, there have been radical utilitarians who indeed explicitly defended exemplary punishment of the innocent (Goldwin, 1976).

Two decades before Bentham’s Principles of Morals and Legislation, Cesare Beccaria (1738–1794) based his famous treatise On Crimes and Punishment (Beccaria, 1764/1995) on a different kind of utilitarian reasoning. Although his work served as an important source of inspiration for Bentham’s systematic utilitarianism, Beccaria’s account of punishment is a mixture between contractarian reasoning and utilitarianism. Beccaria held that “pleasure and pain are the motive forces of all sentient beings” (Beccaria, 1764/1995, p. 21) and that men are essentially egocentric beings who would sooner benefit from the efforts of others than contribute to the common good themselves: quite an unequivocal early statement of the ‘free-rider’ problem.

Beccaria adopted the theoretical fiction of a social contract to explain the origin of punishment and the right to punish. Men came together in society (through a social contract of sorts) to end the constant threats to their personal safety and the state of nature (‘the state of unsociability’) they were living in (cf. Hobbes’s Leviathan). According to Beccaria:
(I)t was necessity which compelled men to give up a part of their freedom; and it is therefore certain that none wished to surrender to the public repository more than the smallest possible portion consistent with persuading others to defend him (Beccaria, 1764/1995, p. 11). (italics added)

The right of the sovereign is comprised of no more than the sum of those smallest possible portions. The sacrifice of the individuals’ portions of freedom alone, however, does not suffice to protect against private usurpations and to promote public happiness because every man, in essence, possesses a despotic spirit. In Beccaria’s view ‘tangible motives’ are needed to prevent the egoistic inclinations of every man from resubmerging society into a state of ancient chaos. These tangible motives are punishments against offences. They act “as a counterbalance to the strong impressions of those self-interested passions which are ranged against the universal good” (Beccaria, 1764/1995, p. 9). In order to effectively act as a counterbalance, there must be a proportion between the crimes and the punishments (Beccaria, 1764/1995, p. 19). As Bentham did later, Beccaria views punishment as an ill which should never exceed what is necessary to attain its utility. The purpose of punishment has nothing to do with undoing a crime already committed, nor to give offenders their deserts, but rather to
(…) prevent the offender from doing fresh harm to his fellows and to deter others from doing likewise. Therefore punishments and the means adopted for inflicting them should, consistent with proportionality, be so selected as to make the most efficacious and lasting impression on the minds of men with the least torment to the body of the condemned (Beccaria, 1764/1995, p. 31).

Like Bentham’s account of punishment, Beccaria’s account is centred around deterrence. As was discussed above, Bentham did not give a clear utilitarian reason why we should not punish the innocent in order to deter potential offenders. Beccaria, however, draws on his notion of the social contract as a negative, limiting principle regarding the power of the sovereign. Individuals had given up only the smallest possible portion of their freedom to the sovereign “for promoting the public happiness by giving the greatest possible protection to the vital interests of each and every citizen rather than by pursuing the greatest possible aggregate utility” (editor’s introduction to Beccaria, 1995, pp. xx–xxi). Therefore, their contribution to the sovereign could never merit punishment for a crime not committed. Punishment of an innocent person simply does not belong to the right originally invested in the sovereign through the social contract. One of the false ideas of utility, therefore, according to Beccaria is separation of the public good from the good of each individual (Beccaria, 1764/1995, p. 102).

2.5.2 Individual or general prevention? Muller’s utilitarianism pur sang
Many preventive strategies can be perfectly reconciled. For instance, a general deterrent sanction could well serve as an individual deterrent (and vice versa), incapacitate and have an educative effect on the general public all at the same time. However, the case of rehabilitation, a method of individual prevention, is in potential conflict with general preventive strategies. The sentencer who chooses a severe prison sentence in order to deter potential offenders, could do this at the expense of the particular offender who might be helped and prevented from reoffending if he were given some form of treatment.

This conflict is inherent in the writings of Nicolaas Muller, a Dutch lawyer and judge[xiv] from the first half of the 20th century. Muller’s dissertation in 1908 (Muller, 1908) was a typical product of the deterministic ‘new direction’ in penological thinking at that time with its emphasis not on the offence and guilt, but rather on the person of the offender and his deficiencies (Janse de Jonge, 1991). Like all utilitarians, Muller found the general justifying aim of punishment in its contribution to the common good in society (Janse de Jonge, 1991, p. 46).

Muller’s dissertation consists of case-studies of recidivists of property crimes. Through these case-studies, Muller systematically tries to show the causes of criminality in terms of individual inclinations and deficiencies. Muller’s main conclusion on the causes of crime is that individual faults (inclinations) like emotional instability, irritability and restlessness appear to be the most criminogenic factors (Muller, 1908, pp. 498–516). Although Muller recognises an (individual) deterrent effect in punishment, in the long run it does not ensure that a convicted criminal will live an honest life after punishment (Muller, 1908, pp. 523–524). One of the most important means to combat (the repetition of) crime therefore, Muller suggests, is an individual patronage. The patron’s job is to guide and educate his ‘pupil’ into an honest life and guard against outbursts of his pupil’s faulty inclinations. This patronage should span a substantial period of time (Muller does not provide any further guidance as to the amount of time requested) because offenders’ deficiencies in personality are longlived (Muller, 1908, pp. 524–528). What Muller suggests in his dissertation is intensive and continuous rehabilitation of convicted criminals quite similar to (but perhaps more extreme than) what is known in the United States as ‘Intensive Probation Supervision’ (see, Byrne, 1990; Lurigio & Petersilia, 1992; Petersilia & Turner, 1993; Tonry & Hamilton, 1995, part III, ch. 1).

While Muller’s dissertation was primarily focused on individual prevention through rehabilitation,[xv] in the 1930s he developed a liking for the general preventive functions of punishment.[xvi] In his essay of 1934, Muller explicitly argues in favour of the general preventive functions of punishment. In the first pages of the essay, Muller points to the potential conflict between individual and general prevention. While general and individual deterrence are perfectly reconcilable, general deterrence and special prevention through rehabilitation can, in specific cases, be in conflict. If general prevention is aimed at, this might well be at the expense of the individual offender in need of rehabilitation (Muller, 1934, p. 16, p. 37).

Muller’s argument for the general preventive functions of punishment is twofold. At the time there appeared to be a certain disappointment with the potential of the rehabilitative ideal: it was less effective than had been hoped for (not in the least part because of its defective implementation in penal policy). But more importantly, Muller found the massive (epidemiological) occurrence of certain types of crime indicative of an increasing social disorder calling for penal strategies other than strict individual prevention (Janse de Jonge, 1991, pp. 44, 54). If the perceived causes of crime are massive (i.e., shared by large groups in society) or if the general norms in society regarding certain crimes seem to be defective, punishment should be designed for its general preventive effect. General prevention is supposed to be achieved through general deterrence and the general educative effect of punishment (affirmation of norms). Muller’s observations, however, did not lead him to completely abandon the rehabilitative ideal which he had defended twenty-five years earlier: when the cause of a crime is primarily found in the individual’s personality, special prevention (i.e., rehabilitation) is preferred. If explanatory individual factors such as poverty or deviant inclinations cannot be found, faulty general norms about the particular crime must be suspected, indicating a general preventive mode of punishment (Muller, 1934, p. 42). Muller thus proposes eclecticism concerning the utilitarian functions of punishment with the attribution of the causes of crime as the primary criterion in each case. But what should happen if certain crimes (at a certain time) appear to take on epidemiological proportions, indicating defective general norms while, at the same time, individual causes can also be demonstrated? Muller’s solution to this conflict is straightforwardly utilitarian: since in such cases the greatest utility is to be expected from punishment with a general preventive working, general preventive punishment is to be preferred (even at the expense of the individual offender[xvii] over punishment with an individual preventive working (Muller, 1934, p. 49).

While Muller does not permit punishment of the innocent (only those who have done wrong are available for punishment), his ideas on limitations on general preventive punishment are quite sketchy. He leaves it up to others in general, and the sentencing judge in particular, to determine those limits (Janse de Jonge, 1991, p. 49). However, what is clear is that the instrumentalist Muller allowed for the sacrifice of a (guilty) individual to the common good. Willem Pompe (see below) was convinced that Muller would intuitively keep the limits of justice in mind, although such limits are not found in Muller’s theory (Pompe, 1934, p. 252). In relation to this particular point, as well as the fact that his theory did not provide any guarantees for equal treatment of offenders (on the contrary!), Muller was subjected to severe criticism.[xviii]

2.6 Mixed theories
Regarding the justification for punishment and the aims of sentencing, one strategy, although neither truly a mix nor a true theory, is eclecticism. For instance, in cases where the sentencing judge has confidence in achieving prevention through deterrence or rehabilitation he chooses a utilitarian mode of reasoning. When there is little confidence in achieving prevention or when the offence is particularly shocking, the sentencer falls back on retributive reasoning (Walker, 1991, pp. 123–126). This is more of a pragmatic, multi-stage rocket approach to sentencing than a theoretically integrated account of punishment.

Mixed theories differ from such eclectic approaches towards punishment. They do not supply any essentially new insights concerning the general justification of punishment or the aims at sentencing. Rather, they draw upon elements both from retributivist and utilitarian approaches to form ‘hybrid’ accounts of punishment. In such hybrid accounts, through the combination and integration of retributive and utilitarian principles, one type of reasoning is moderated or limited by the other type of reasoning. This makes hybrid accounts of punishment theoretical and practical alternatives for strict retributive or utilitarian reasoning.

The shape of a hybrid account of punishment depends on the theoretical point of departure. Two general shapes are possible:

1. Utility (i.e., the common good) as the general justification for the practice. The negative retributive principle is superimposed to limit punitive action aimed at prevention: Only the guilty may be punished and only to the extent of their desert.
2. Retribution as the general justification for the practice. Retributive demands on punishment are toned down by utilitarian considerations. Although retribution provides the general justification for the practice, ‘justice’ no longer dictates punishment to be meted out to the extent of the offender’s desert. Rather, utilitarian considerations allow for punishing less than would be indicated by desert, and may even allow for refraining from punishment altogether.

The first type of mixed theory has been discussed in Section 2.4.1 in the context of negative retributivism. There are utilitarians who have recognised the necessity of “independent side constraints of justice that forbid the deliberate punishment of the innocent, and perhaps the excessive punishment of the guilty” (Duff, 1996, p. 3). These utilitarians have embraced the negative retributive principle as a limiting principle (see, Hart, 1968; Morris, 1992). The negative principle provides protection to individuals against disproportionate and unfair use of punishment for the sake of utility. Pure utilitarian reasoning, as was discussed in the previous section, cannot provide such a limiting principle (cf. Bentham). An external principle is needed to guard against the potential excessiveness of utilitarianism. Beccaria also saw the need for a negative principle. However, instead of the negative retributive principle, he makes reference to the social contract as a safeguard against the potential excessiveness of utilitarianism.

The second type of mixed theory takes the opposite view. Retribution, in this hybrid account, constitutes the essence of punishment, its general justification. Below the (upper) limits, defined by retribution, notions of utility determine the choice concerning mode and severity of punishment. Although retribution constitutes the justification for punitive action, punishment is not seen as a moral necessity (cf. Kant’s categorical imperative). Rather punishment is permitted on grounds of retribution and only (at most) to the extent of the offender’s desert. Retribution provides a threshold (lower boundary) in the sense that only the guilty may be punished. It is generally believed that in the Netherlands, this second type of mixed theory comprises the dominant point of view. The spirit of our penal code reflects such a view (see, Hazewinkel-Suringa & Remmelink, 1994; Jörg & Kelk, 1994; Kelk, 1994a; Kelk & Silvis, 1992). In the Dutch penal code, there are no minimum sentencing requirements specified for separate crimes.[xix] Furthermore, there are upper boundaries specified for each separate crime regarding the portion of punishment. Within the limits of the criminal code, Dutch judges thus have large discretionary powers in sentencing.[xx] Below the upper boundary specified for a particular crime, utility (i.e., crime prevention) is believed to provide the guiding principles at sentencing. However, the aims of sanctions are not codified as explicit principles in the Dutch penal code (Denkers, 1975, p. 125).

Willem Pompe, one of the founders of the once (roughly from 1945 to 1965) influential Dutch ‘School of Utrecht’, defended a mixed theory of the second type. Pompe went to great lengths to argue for retribution of guilt as the general justification of punishment. His conception of retribution is based on notions of ‘collective’ intuition completed with the disturbance of the objective moral order in conjunction with the personal guilt of the offender (Pompe, 1930; Pompe, 1943). Pompe defines the moral order as the desired order of reasonable, free beings structured by moral laws (Pompe, 1943, p. 33). Central to Pompe’s approach is his view of man as a responsible moral being in possession of free will. Punishment should appeal to the offender’s sense of responsibility and address him in such a way that honours him as a free being, capable of moral understanding (Pompe, 1957).

It is therefore not surprising that Pompe had some fundamental objections to Muller’s notion of general prevention. General prevention (as advocated by Muller) degrades an offender as a means of attaining something for others; it erodes the offender’s human dignity and treats human beings as objects instead of persons (Pompe, 1934, p. 251). However, the common good does play a significant role in Pompe’s approach. Because the goal of the moral laws is the common well-being, the aims at sentencing should be directed towards that goal. The mode and severity of punishment should be guided by utility. However, since the essence of punishment is retribution, punishment of the innocent is unjust as is punishment that exceeds the measure of guilt. Pompe recognises a problem in positive retributive reasoning in that it would require punishment to be meted out to the guilty to the extent of their desert. The common well-being is therefore applied in decisions relating to the amount of punishment and indeed whether it should be meted out at all (Pompe, 1930, p. 119; Pompe, 1959, p. 9).

The two types of hybrid accounts discussed here are each other’s theoretical mirror image. Interestingly, although the normative foundations (i.e., the general justifications) are completely different, implications for the practice of sentencing are quite similar. Eventually, in both hybrid accounts, utility rules within the limits of desert. Hybrid accounts may provide useful guidelines for the practice of sentencing. They also seem to circumvent some of the ethical objections made to pure utilitarian or retributive reasoning. However, their practical and theoretical attractiveness tends to hide some potential dangers. If a theory should bind the practice of punishment to a certain order and regularity (see Section 2.2), do hybrid accounts provide acceptable and stable points of reference?

After all, hybrid accounts are pre-eminently suitable for ‘criminal politics’ (Enschedé, 1990, pp. 14–19). In accordance with temporal and local circumstances or trends, a mixed theory provides ample room for shifting emphasis on functions and goals of punishment within the framework of the dominant mixed theory. The pendulum, however, does not swing from one extreme to the other (Kelk, 1994a). Such shifting emphasis will be believed legitimate because this occurs within the hybrid framework.[xxi]

However, can we determine whether penal action that occurs within the limits of our criminal code is still in concordance with a hybrid view on the justification and goals of punishment? Perhaps the practice of punishment has deteriorated from one founded on truly hybrid principles into a disguised form of eclecticism. As Duff and Garland noted:
One question about any mixed theory is whether the ‘mixture’ is a stable one that can be consistently applied, rather than a shifting patchwork of compromises and arbitrary decisions (Duff & Garland, 1994, p. 5).

Another issue concerning a hybrid account of punishment involves incorporating a multitude of potentially conflicting principles. It has already been discussed that individual and general prevention may put conflicting demands on the mode and severity of punishment. Such a conflict is immanent in the utilitarian approach (but may be resolved by the ‘highest expected utility criterion’). A mixed theory retains this conflict while adding yet another difficulty. If punishment is essentially retribution of guilt, it would be desirable for punishment to be proportionate to the seriousness of the offence and the culpability of the offender. From the retributive point of view, equality and consistency in sentencing is essential.

However, from an (individual and general) preventive point of view proportionality and equality may render sentencing practices ineffective (Denkers, 1975, pp. 124–125; Levin, 1972). These utilitarian principles would prefer highly individualised and differentiated sentences dependent on the expected utility regarding a particular offender or group of potential offenders. Both views have their repercussions within the hybrid framework. These conflicting pulls and pushes produce a permanent field of tension within any hybrid account of punishment (cf. Kelk, 1994a, pp. 5–9).

2.7 Restorative Justice
The final approach of relevance to the present study is that of restorative justice. In Section 2.3 restorative justice has been called a transitional approach since, in the long run, it aims to replace the traditional penal system (based on retributive and rehabilitative ideas) with the restorative paradigm. In that respect, restorative justice embraces system-competitive and therefore radical views. However, the short-term engagement of those who advocate restorative justice is essentially immanent.[xxii]

An important drive behind the development of the restorative justice approach is dissatisfaction with the existing retributive and utilitarian rationales. According to the advocates of restorative justice
(…) the main objective of judicial intervention against an offence should not be to punish, not even to (re-)educate, but to repair or to compensate for the harm caused by the offence (Walgrave, 1994, p. 63).

This entails a perspective on (the reaction to) crime quite different from the approaches of retributivism and utilitarianism. Thus within the framework of restorative justice, one no longer speaks of punishment (i.e., intentional and avoidable infliction of suffering). Rather, the term ‘intervention’ is favoured. The intervention, in turn, is justified by the damage done, and is aimed at compensation (reparation) of that damage. The concept of crime is redefined as a social conflict involving the victim, the offender and the community. Restorative justice promotes maximum involvement of these three parties (Bazemore & Maloney, 1994). In many of these respects, restorative justice is closely related to Braithwaite’s reintegrative shaming approach to crime and punishment (cf. Braithwaite, 1989; Braithwaite, 1999; Walgrave, 1996).

Christie, defining crime as a conflict, has argued that in the penal system as we know it, the state has taken away this conflict from the parties whose interests are at stake. The victim of a crime now becomes a double loser: after his victimisation by the offender, the victim is denied the rights to full participation in the process. The state’s appropriation of this particular social conflict is a lost opportunity “for involving citizens in tasks that are of immediate importance to them” (Christie, 1977, p. 7; see also Christie, 1981, ch. 11). In restorative justice, the conflict is ‘given back’ to the parties involved. The role of the state is pushed back to that of a (pro-active) mediator between victim, offender and the community. In this respect, restorative justice functions more like civil law than criminal law. However, the pro-active and coercive roles of the state are maintained in order to initiate and guard restorative processes.

In restorative justice, crime is viewed as injury to the victim and to the community. The conflict must be resolved by restitution of wrongs and losses.[xxiii] The offender is held responsible and accountable for his actions and should play an active role vis-à-vis victim and community to repair for the damage done. Although the objective is not to punish, nor to rehabilitate, these may be spin-offs from the restorative intervention. The fact that an offender is coerced into participating in a mediation process, or, alternatively, to perform community services, may entail some retributive element. Furthermore, the restorative process may have beneficial effects on the offender’s personality. He is confronted with the consequences of his actions and is expected to contribute to the resolution of the conflict in a positive and constructive way (Walgrave & Geudens, 1996). The central tools in the restorative justice approach for the restitution of wrongs and losses are types of mediation and community service. Originally interventions like mediation and community service were designed as add-on components of sentences, or as alternatives to incarceration in times of prison overcrowding (Jackson, De Keijser, & Michon, 1995; Van Ness, 1990, p. 8). Some reformers however, saw these interventions as opportunities for developing an alternative paradigm of justice: restorative justice. Although damage done is the primary point of reference in restorative justice, the disturbed social order (i.e., disturbed relation between offender and community) is used to complement the justification and to back-up cases where there is no clearly identifiable victim or damage is difficult to define. In those cases the vindication of injustice done to society constitutes the justification for a restorative intervention (Thorvaldson, 1990; Walgrave & Geudens, 1996).

The development of the restorative approach has been simultaneous with the internationally increasing recognition of the rights and needs of victims of crimes since the 1980’s (Ashworth, 1992b, p. 68–69). In 1985, the Council of Europe adopted a recommendation called ‘The position of the victim in the framework of criminal law and procedure’. In 1986 the United Nations adopted the ‘Declaration of basic principles of justice for victims of crime and abuse of powers’. In the Netherlands, increasing attention for victims of crimes is shown by experiments with (pre-trial) restitution and settlement programs[xxiv] as well as by the (so-called) Terwee Act that came into force in 1995. The Terwee Act increases the obligations of prosecution and police to keep victims informed about proceedings and progress in their case. Furthermore, the Act introduced a new type of sanction: the compensation order which entails restitution by the offender to the State. The State, in turn, refunds the sum of money directly to the victim. Finally, the Terwee Act improved the legal possibilities for victims to join a criminal procedure in order to claim restitution (Malsch & Kleijne, 1995). However, although victims of crimes receive more attention nowadays than they did a few decades ago, restorative justice still maintains an uneasy relationship with the existing penal system. Dutch criminal law is non-reparatory in essence and offendercentered (Corstens, 1995, p.2). Some have argued that increasing attention for victims within the criminal law threatens its very essence (e.g. Buruma, 1994). However, that is exactly the objective of those who advocate restorative justice. It is not at all surprising that the intended transition of criminal law into reparative law will have to contend with a fair amount of opposition.

Two related issues are central to the critique of the restorative justice approach: equal treatment of offenders and proportionality between offence and intervention. Restorative justice requires looking at how the particular victim has experienced the offence and considering what the victim regards as being necessary to repair the damage (Davis, 1992, p. 8–9). It should therefore be obvious that taking the (subjective experience of) damage done as the point of reference could result in quite different reparative actions being required from offenders for similar offences. When there is no identifiable victim, as in victimless crimes or when the victim refuses to co-operate, the reparation would take the form of a community service with the ‘social harm’ caused by the offence as the point of reference. This introduces the problem of quantifying reparation to the community for the public aspect of the offence (Wright, 1991, p. 124). How is this symbolic reparation to the community to be calculated and applied consistently if not on the grounds of desert
(Ashworth, 1992b, p. 69)?

Furthermore, taking damage done as the point of reference and reparation as the goal, may very well lead to interventions that turn out to be even more punitive than those required from the desert point of view. Viewing the punitive effects of the restorative intervention as mere unintentional side effects means losing sight of the proportionality between harm done and culpability on one side and the infliction of suffering on the other side. Focussing on the victim and his damage and the obligation for the offender to repair, implies a certain indifference to the (unintentional) suffering inflicted on the offender. In most cases the restorative intervention would be less punitive than a retributive intervention. However, in some cases, entailing extreme levels of damage, a negative (retributive) principle would seem desirable from a humanitarian point of view. This uneasy relationship between reparation and (retributive) proportionality has indeed led some to embrace negative retributivism as a limiting principle defining the upper limits of the intervention (e.g. Cavadino & Dignan, 1997b). The result of superimposing the retributive negative principle over restorative justice actually creates a new hybrid model.

NOTES
i. In structuring this overview the discussions provided by Walker (1991), Duff and Garland (1994) and Tunick (1992) have been very helpful.
ii. The Restorative Justice paradigm, discussed in Section 2.7, could be argued to advocate a shift from the first to the second and fourth of Kelk’s domains.
iii. For a critique on this distinction, see Dolinko (1991). In spite of the critique, I find the distinction useful for analytical purposes.
iv. The terms theory and philosophy of punishment are frequently used interchangeably in the literature. There is, however, a formal difference between the two: penal philosophy is said to be primarily concerned with the general moral justification of punishment, while theories of punishment are more concerned with the formulation of the practical ends of penal action (Duff & Garland, 1994). I do not find this distinction helpful since both kinds of discourse are logically inextricably linked and the important distinction to be made is that between general justification and practical aims of punishment. I therefore prefer the term ‘philosophical theory of punishment’ as used by Duff and Garland (Duff & Garland, 1994, p. 17). When I use either the term theory, or the term philosophy, I refer to an integrated account of the general justification and the practical aims of
punishment (i.e., a philosophical theory of punishment).
v. Hart uses the more subtle phrase “disguised forms of Utilitarianism” (Hart, 1968, p. 9). Braithwaite and Pettit make a similar point in saying that “retributivist attempts to provide a rationale for punishment drift into consequentialist doctrines” (Braithwaite & Pettit, 1990, p. 206).
vi. Explaining retributive punishment by referral to our emotions of love and hatred was explicitly proposed by the Dutch philosopher and psychologist, Heymans. See Hazewinkel-Suringa (1994, pp. 896-897), Polak (1947, Ch. VII).
vii. Nozick names five ways to distinguish retribution from revenge (Nozick, 1981, pp. 366-368):
1. Retribution is done for a wrong while revenge may be done for a harm or injury.
2. Retribution sets an internal limit to the amount of punishment whereas revenge need not.
3. Revenge is personal whereas the agent of retribution need not have a personal tie with the victim.
4. Revenge involves the pleasure in the suffering of another, while for as far as retribution involves emotions it is the pleasure in justice being done.
5. The agent of retribution is committed to general principles mandating punishment in certain circumstance, while the agent of revenge seeks vengeance depending on how he feels at the time about the harm or injury.
viii. For an extensive elaboration on intuitionism and its role in our conception of justice, see Rawls (1971, pp. 34–45). See also Nozick (1981, pp. 482–485), on judgement in ethics.
ix. See Polak (1947, Ch. V) for an extensive discussion.
x. In § 214 Hegel comes back to the problem of determining how and how much to punish. He states that “the only interest present is that something be actually done, that the matter be settled and decided somehow, no matter how (within a certain limit)” (Hegel, 1967, p. 137). Hegel leaves it up to the judge’s discretion to decide (within the limits of the law) in each particular case.
xi. Partial, because Von Hirsch argued that punishment also serves the consequential end of deterrence.
xii. For an extensive and oft cited discussion on the expressive function of punishment, see Feinberg (1970).
xiii. For discussions on these practical problems related to this approach see Clear (1996), Dolinko (1991), Duff (1996), Von Hirsch (1992a; 1993). Because of these problems some of the authors who originally endorsed this approach (including Herbert Morris and Andrew Von Hirsch) have later abandoned it.
xiv. For an extensive discussion of Muller’s life and work, see Janse de Jonge (1991).
xv. Muller uses the term ‘education’. His definition of education (Muller, 1934, p. 19) conforms to rehabilitation as defined in Section 2.5 above.
xvi. Two other Dutch authors, who will not be further discussed, who explicitly endorsed general prevention as a primary function of punishment are Th. W. van Veen (see Van Veen, 1949) and G.E. Langemeijer (see Langemeijer, 1975). They both emphasised the general educative functions of punishment (Berghuis, 1992).
xvii. And even if it would increase the chance of the individual’s reoffending (Muller, 1934, pp. 48-49).
xviii. For a full discussion of these criticisms, see Janse de Jonge (1991, pp. 53-61).
xix. For instance, even though an offender has been found guilty of a particular crime, the sentencing judge may decide not to punish at all. He may make such a decision because of the minor seriousness of the offence, the personality of the offender, circumstances at the time of the crime or circumstances that emerged at a later point in time (such as a terminal disease): See art. 9a Dutch Penal Code (DPC).
xx. The wide discretionary powers of Dutch judges are further discussed in Chapter 5.
xxi. For a concise discussion of the shifting emphases in criminal politics in the Netherlands, see Kelk (1994a, pp. 5-21).
xxii. The fact that the short-term engagement of restorative justice is immanent in character has potential dangers for the approach itself. Some have recognised the risk of instrumentalisation of restorative interventions by the conventional penal system (e.g. Messmer & Otto, 1991).
xxiii. Long before the development of the restorative justice paradigm, Hulsman, Dutch professor of law and abolitionist, endorsed conflict resolution as the primary goal of the justice system (see Hulsman et al., 1986).
xxiv. For discussions on the Dutch settlement and mediation programmes, see Malsch & Kleijne (1995).