Retributivism and the (Lack of) Justification of Proportionality

Abstract

The principle of proportionality has gained widespread adherence in the modern retributively-dominated era of penal theory. It has often been held that, if one subscribes to a retributivist theory, then one is also committed to proportionality in punishment (or at least to proportionately-determined upper punishment constraints). In the present article, this assumption is challenged. It is shown that the inference from the fact that one offender has committed a more serious crime than another offender, to the conclusion that this offender should be punished more severely than the other, presupposes the satisfaction of two conditions: The Necessity Requirement and the Sufficiency Requirement. It is argued that modern expressionist accounts of retributivism fail to satisfy these requirements. Moreover, it is suggested that the satisfaction of these requirements constitutes a challenge for other retributivist theories. In so far as this is the case, the inference from retributivism to proportionality will be blocked.

The principle of proportionality has gained widespread adherence and plays a significant role in both penal theory and penal practice. However, over the last couple of years, increasing attention has been directed to a plethora of theoretical challenges facing this principle. It has been shown that the idea of comparing and ranking crimes in gravity gives rise to several serious problems (see e.g. Ryberg 2004, 2020). Moreover, the prima facie simpler task of comparing punishments in severity also turns out to be encumbered with challenges (Husak 2020; Kolber 2009; Ryberg 2004, 2010; Tonry 2020). Furthermore, even if scales of crime and punishment were to be properly constructed, the question as to how these scales should be combined, in order for the principle to provide genuine guidance, in itself seems to prompt various problems (Duus-Otterström 2020; Ryberg 2004). Even though all these theoretical challenges are important and, in my view, have not been properly answered in the recent debate, they will not be subjected to further scrutiny in the following. Rather, what I will focus on in this article is a question that in a sense is logically prior to any of these ethical technicalities; namely, whether one should at all accept the idea of proportionality. Even if scales of crime and punishment can be constructed and combined in ways that overcome all theoretical problems, one would still be confronted with the question as to whether punishments should be imposed in a proportionate manner; that is, whether the principle of proportionality is justified in the first place. It is this issue that will be the subject for consideration in the following.

The question of how proportionality is justified depends upon more basic theories of punishment. Some theorists have followed in the footsteps of Bentham by contending that a consequentialist theory of punishment sustains proportionality in penal distribution (see Bagaric 2001). Whether this is the case is ultimately an empirical question and not one that will be pursued in the following. The majority of modern adherents to proportionality subscribe to some version of retributivism. The idea is that, if one accepts a retributivist theory of punishment, then one will also be committed to the principle of proportionality. However, in the following it will be argued that this inference is premature. In order to sustain this contention, the article will proceed as follows. In Sect. 1, the general structure of a retributivist argument in favour of proportionality will be outlined. An obvious problem concerning the retributivist approach to punishment is that “retributivism” does not denote a single theory. Rather the term covers a range of theories that—though they all pay particular attention to the moral significance of desert—may vary significantly. However, despite differences, retributivist theories usually share a certain feature, namely that what an offender deserves from having broken the law is given by the imposition of a punishment. Given this structure of retributive theory, it will be argued that an inference from retributivism to a proportionalist conclusion presupposes the satisfaction of two requirements. However, it is suggested that these requirements are often not fulfilled. In Sect. 3, this general point is exemplified by considering more narrowly one of the versions of retributivism that has gained widespread adherence, namely expressionism. Theorists subscribing to expressionism have often been explicit in showing how this account of retributivism sustains the principle of proportionality. However, it will be argued that expressionism fails to satisfy the two requirements outlined in Sect. 1 and, consequently, that the theory fails to justify proportionality. Section 3 follows up on the previous arguments by considering possible objections. It will be considered whether retributivism can be construed in a way that satisfies the outlined requirements or whether to simply bite the bullet and accept that the requirements are not satisfied. It is suggested that both answers have serious drawbacks. Finally, Sect. 4 sums up and concludes.

However, before embarking upon this, there are a few conceptual points that need clarification. First, the idea of proportionality can be given somewhat different interpretations. One aspect of proportionality concerns the relative distribution of punishment. If one crime is more serious than another, then it should be punished more severely. Correspondingly, if two crimes are equally serious, they should be responded to with equally severe punishments. This is usually referred to as “ordinal” proportionality (Duus-Otterström 2020; von Hirsch 1993, 2017; von Hirsch and Ashworth 2005). Another aspect concerns the way punishment non-relatively comports to a specific crime (e.g. it would be regarded as disproportionate to punish pickpocketing with ten years in prison even if ordinal proportionality requirements are strictly observed). This is often referred to as “cardinal” proportionality (Duus-Otterström 2020; von Hirsch 1993, 2017; Ryberg 2004). In the following, I will not engage in the more complicated considerations of cardinal proportionality, but only consider proportionality in the ordinal sense.

Second, retributivist theories differ significantly in the extent to which they leave room for considerations of consequences.¹ While a positive retributivist might hold that offenders should be punished in ways that fully observe considerations of proportionality—that is, neither more nor less than the punishment that is proportionate to the gravity of crime committed—this is not the case for theories that leave more room for consequences. For instance, so-called limiting retributivists would hold that proportionality sets upper and lower constraints on punishment, which leaves room within which the punishment can be fine-tuned on the grounds of consequentialist considerations. Somewhat along the same lines, a negative retributivist would hold that proportionality only places upper constraints on punishment or what is sometimes referred to as “punishment ceilings”.² Both theories, therefore, open up the possibility that two offenders who have committed equally serious crimes could be punished differently (i.e. that strict ordinal proportionality is not observed). However, as will become clear, since both theories imply that the upper limits of deserved punishment should follow an ordinal ranking—that is, for instance, that the punishment ceiling is higher for rape than for burglary, and higher for burglary than for theft—the considerations in the following will also be relevant for these theories. Therefore, even though—for reasons of ease in exposition—I will in the following talk in terms of proportionate punishment (and not upper proportionality constraints), it should be kept in mind that the subsequent arguments are not only relevant in relation to a strict positive interpretation of retributivism.

1 The Retributivist Justification of Proportionality

It is a widespread belief that proportionality is closely tied to the retributivist approach to punishment. This is not surprising. If an offender’s desert debt increases with the seriousness of the crime, and if the severity of a punishment should reflect the magnitude of the desert debt, then it seems to follow that a more serious crime should be responded to with a more severe punishment and that equally serious crimes should be met with equally severe punishments. However, on closer inspection, it turns out that the picture is more complicated.

Modern theories in the retributivist tradition have often been careful to specify precisely what it is an offender deserves from having committed a crime. For reasons of simplicity we will refer to this as the “desert object”.³ The claim then is that the desert object can be given to the offender by subjecting him or her to a punishment. The desert object varies between different retributivist theories. However, if, for the present, we leave this out of consideration, then we can depict the relation between crime and punishment in the following general terms. An offender has committed an offence of a certain gravity; the offender, therefore, deserves a desert object of a certain magnitude; this is given to the offender by imposing a punishment of a certain degree of severity. However, in order to reach the relation between the seriousness of the crime and the severity of the punishment that is expressed in the principle of proportionality, two requirements will have to be fulfilled. The first requirement is:

The Necessity Requirement: The desert object can be given to the offender only by the imposition of a punishment the severity of which reflects the magnitude of the desert object.

It is easy to see why this requirement must be satisfied in order for the retributivist argument for proportionality to be conclusive. If the requirement is not satisfied, that is, if the desert object can be inflicted on or befall the offender in other ways than through a punishment, then it will no longer follow that one offender should be punished more severely than another, even if the former has committed a more serious crime than the latter. For instance, suppose—as some retributivists have suggested—that the desert object is suffering (without further qualification). It is obvious that an offender can suffer for various reasons. He or she might suffer as a result of a disease, from the loss of a relative, or for numerous other causes. Therefore, if a person has committed a crime but subsequently suffers for some non-punitive reasons then the offender may, thereby, have paid his or her desert debt. If the amount of suffering corresponds to (or surpasses) that which the offender deserves from having committed a crime of a certain degree of seriousness, then there is no longer any desert-based ground for punishing him or her. Or, if we imagine that two offenders have committed equally serious crimes, but the first has subsequently suffered for some non-punitive reasons, while this is not the case for the second, then, in order to ensure that both offenders get what they deserve, it will be necessary to punish the second but not the first (or she or he may be punished up to the extent that remains in order to have paid the full desert debt). In short, if the object of desert is suffering simpliciter, then it is clear that the Necessity Requirement has not been fulfilled, which means that this version of retributivism fails to justify proportionality between the seriousness of the crime and the severity of the punishment. There will be cases where a more serious crime should not be responded to with a more severe punishment, and where equally serious crimes should be met with punishments of varying severity. Clearly, there are many retributivists who do not subscribe to a theory according to which the desert object is suffering without further qualification.⁴ But the point here is simply to illustrate that the Necessity Requirement must be satisfied in order for the inference from a retributivist theory to the principle of proportionality to be valid.

However, the Necessity Requirement is only one of the conditions that a retributivist theory will have to satisfy in order to buttress proportionality. Another condition concerns the extent to which a punishment succeeds in ensuring that the offender actually gets what he or she basically deserves. We can refer to this condition as:

The Sufficiency Requirement: The punishment must always succeed in imposing the desert object on the offender.

Once again, it is easy to see why such a requirement is necessary in order for the proportionality principle to be deducible from a retributivist theory of punishment. Let us imagine two offenders who have committed equally serious crimes and who are both put to prison for one year. Then, if the requirement is not satisfied, that is, if the prison time has only succeeded in imposing the desert object on one offender but not on the other, there would still be a moral requirement to keep on punishing the first offender—the one who has not received what he or she basically deserves—while this would not be the case with regard to the second. In other words, if the Sufficiency Requirement is not satisfied then there can be cases where offenders have committed equally serious crimes but should nonetheless receive punishments of varying severity, or where an offender who has committed a more serious crime should be punished more leniently than someone guilty of a less serious crime. Thus, if the requirement is not satisfied, the inference from a retributivist theory to the principle of proportionality will be blocked.

If these considerations are correct, that is, if both requirements are necessary in order for retributivist theory to justify proportionality, then to what extent do modern retributivist theories succeed in satisfying these conditions? Before answering this question, it should be underlined that both requirements must be satisfied; that is, in order to block the inference from the fact that one offender has committed a more serious crime than another, to the conclusion that this offender should be punishment more severely, it is sufficient that one of the conditions is not satisfied. In my view, it is the Necessity Requirement in particular that constitutes a challenge for retributivist theories. However, the answer to the question is troubled by the fact that retributivist approaches to punishment, as initially noted, cover a wide range of very different theories. It is therefore not possible to examine here the extent to which each possible interpretation of retributivism fails to meet the two requirements. More modestly, what I will do is examine to what extent one of the most influential modern versions of retributivism, namely so-called expressionism, succeeds in satisfying the conditions. This is the task to which we will now turn.

2 Expressionism and Proportionality

Inspired by Feinberg’s early contention that punishment has “symbolic significance” (Feinberg 1970), a number of theorists in the modern retributively dominated era have developed and defended expressionist accounts of retributivism (Duff 1986, 2001; Lucas 1980; von Hirsch 1993, 2017; von Hirsch and Ashworth 2005; Wringe 2016). The basic idea is that a punishment can be seen as a sort of language or, more precisely, as a way of sending a condemnatory message to the offender. As has been underlined, the term “expressionism” is not fully apposite in the sense that what matters is not merely that something is expressed by performing a punitive act, but rather that an appropriate message is conveyed to the offender. In this sense, the most influential versions of the theory should be seen as communicative. According to these theories, what an offender deserves—i.e. the desert object—is to be condemned or censured for having performed a reprehensible act. The idea, then, is that this communicative purpose is served by imposing an appropriate punishment on the offender. Several exponents of this theory have also underlined that an expressionist theory implies that punishments should be proportionate. Briefly put, the idea is that if a crime is more serious, then the offender deserves to be more strongly blamed, and that the severity of a punishment expresses the stringency of the blame. Therefore, an offender who has committed a more serious crime should be punished more severely and equally serious crimes should be met with equally severe punishments.⁵ The question that arises then is, does such a theory succeed in satisfying the two previously outlined requirements?

If we start by considering the Necessity Requirement, then at first glance it seems obvious that the answer must be in the negative. It is clearly not the case that the desert object, that is in casu condemnation, can only be communicated to an offender by the use of punishment. The most frequent procedure in our condemnatory interaction with other people is to blame them by simple verbal means, that is, by the use of language. Furthermore, it is clear that it is also possible to use various sorts of symbolic means. But if this is so, then it seems obvious that proportionality does not follow from the expressionist theory. If other communicative means can be used, it no longer follows that an offender should be punished. In fact, it would seem morally absurd to hold that offenders should be punished if the desiderata of the communicative enterprise could be satisfied in ways that do not involve the hard treatment that characterizes a punishment. The imposition of unnecessary hardship on someone—even an offender—cannot be morally right. Advocates of the expressionist theory are, of course, aware of this. Thus, the task they are confronted with is to explain why the communication should take a punitive form. In short: why hard treatment? The answers that have been provided vary between different expositions of the theory. In the following, we will look at three possible answers.⁶

According to the theory presented by Andreas von Hirsch—one of the theorists to have done the most to defend proportionality—condemnatory communication should take a punitive form, for reasons that go beyond mere considerations of desert. In his view, hard treatment is preferable because it serves a crime-preventive function. An offender, censured for a crime, gets the message that the act was wrong and is thereby given a reason for desistence. However, if the message is conveyed through hard treatment, rather than through other means, this will provide the offender with a supplementary reason for resisting the temptation to engage in future criminal activity. Thus, hard treatment is necessary for crime-preventive reasons—by functioning as an additional prudential disincentive—but, importantly, it is suggested as operating within the desert-based framework. But if this is the case, that is, if what basically matters is the communication of censure and imposition of a disincentive via hard treatment—i.e. if we interpret hard treatment censure as the desert object—then we can repeat our initial question concerning the Necessity Requirement. Namely, can this be achieved only by the imposition of a punishment? On closer examination, it turns out that the answer is in the negative.

Suppose that two offenders O₁ and O₂ have committed equally serious crimes and therefore, according to the theory, ought to be censured to the same extent. Suppose, furthermore, that it is the case that if a minor amount of hard treatment is imposed on O₁ then he or she will be completely disincentivized to engage in future criminal conduct; whereas this is not the case for O₂. In this situation, how should the criminal justice system react to the two offenders? As we have seen, what is crucial is that they should be censured to the same extent. But is punishment necessary to achieve this goal? The answer is in the negative. Since von Hirsch, as noted, fully acknowledges that censure can be conveyed in other modes than by hard treatment, what one should do is to inflict hard treatment (i.e. punishment) on O₁ up to the level where he/she is fully disincentivized (i.e. where further hard treatment will not provide an extra disincentive) and then convey the remaining degree of censure in other modes (say, in everyday language or symbolically). In other words, given the assumption that censure can be communicated in other ways than by hard treatment, and that it is wrong to impose unnecessary hardship (i.e. in this case, hardship that is not required for crime-preventive reasons), it follows that O₁ should not be punished as severely as O₂. And this is so, even though the two offenders have committed equally serious crimes.⁷ Thus, the desert object both can and should be given in ways that only partly involve punishment, which means that the Necessity Requirement is not satisfied and the theory does not prescribe proportionate punishment (only proportionate censure). Along the same line of reasoning, cases can also be constructed where a more serious crime should be punishment more leniently than a less serious crime (and where this is fully consistent with the goals of conveying proportionate censure and providing a supplementary prudential disincentive).⁸

It might perhaps be objected that the presented argument presupposes that an appropriate degree of censure can be conveyed by the use of two different modes (punishment and, say, a symbolic mode) and that the censure or condemnation that is conveyed via two modes will be working in an additive manner, so that the censure communicated in one mode will not only constitute a repetition of what is communicated in the other mode. However, it is difficult to see why this assumption should not be accepted. In the same way that a person can be praised by receiving both money and a medal, and will thereby be praised more than another person who receives only the same amount of money, it also seems reasonable to believe that condemnation, if properly communicated, can be conveyed though different media and work in an additive manner (see Ryberg 2004, p. 32). Therefore, in summary, even though von Hirsch is of course aware of the fact that censure as a desert object does not per se satisfy the Necessity Requirement, his attempt to make up for this by his two-pronged theory does not succeed. Thus, his theory fails to justify proportionality.

A second and very different answer to the “why hard treatment” question has been given by another influential exponent of expressionism, namely, Antony Duff. He shares the basic idea that a punishment can communicate censure to the offender, but also underlines that such messages can be conveyed by purely symbolic means. In his view, a punishment should ideally function as a kind of penance.⁹ He believes that a punishment should result in some sort of penitential reform in the offender. More precisely, what he suggests is that a penance serves several interrelated purposes: it focuses the offender’s attention onto his or her wrongdoing; it symbolically portrays the character of the crime; it aims to make the offender recognize and repent the wrongdoing; and it functions as a vehicle of self-reform (see e.g. Duff 1998, pp. 164–165). Thus, while censure itself does not require hard treatment, it is the penitential function that necessitates such treatment being imposed on the offender. The question now is whether this account of expressionism succeeds in satisfying the Necessity Requirement. In order to answer this, it is instructive to see how Duff envisions a punitive process might take place (see Duff 1986, p. 289). As he rightly underlines, people are often unwilling to face up to their misdeeds. An offender may well have a powerful temptation to evade the issue by various sorts of self-deceptive excuses. When hard treatment is imposed upon an offender it may well be the case, therefore, that the hardship or suffering is not readily accepted by this person as a penance. What happens then is that, ideally, the hard treatment functions as a way of persuading the offender to accept the punishment as a penance. Thus, what begins as a coercive attempt to attract the unrepentant offender’s attention will ideally become the penitential suffering that the repentant offender accepts for him/herself. This shows that initiating and going through the kind of penitential reform Duff has in mind may well be a demanding and time-consuming process and, it would seem, one that offenders need not go through in precisely the same way. However, this opens up the following possibility.

Suppose that the offenders O₁ and O₂ have committed precisely the same crimes and, also, that O₁ is quickly persuaded that he/she has done wrong, and that he/she has been through the penitential reform after spending six months in prison, whereas this process takes eight months in prison for O₂. Now, if the penitential reform constitutes the justifying aim, then it would seem that O₁ should spend two months less in prison than O₂. After all, he/she has been fully reformed in a shorter period. However, if the stringency of the censure that reflects the gravity of the crimes that have been committed corresponds to eight months in prison, then how severely should O₁ be punished? That is, what should the criminal justice system do if there is a discrepancy between what the censuring function and the penitential process require in terms of hard treatment? It might be held that, if the seriousness of crime is reflected in eight months of imprisonment, then this is the appropriate punishment also for O₁. However, this answer is premature. As we have seen, the expressionists rightly acknowledge that censure can be conveyed in other ways than via hard treatment. But this means that the right way to punish O₁ would be to keep him/her behind bars for six months—which is what the penitential reform requires—and then impose the remaining censure by other means.¹⁰ In that case, O₁ and O₂ will go through the proper penitential reform and both be censured to the same extent reflecting the gravity of their crimes; but importantly, they will be punished differently. To insist that O₁ should in this situation also receive eight months in prison would be tantamount to holding that one should inflict unnecessary hardship on this offender; a claim that, as noted, seems ethically highly dubious. Therefore, in short, what this shows is that an expressionist theory of the type presented by Duff does not satisfy the Necessity Requirement and, consequently, does not succeed in satisfying proportionality in punishment.

Whereas the first two theories discussed below try to answer the “why hard treatment” question by adding extra theory to the basic expressionist idea that offenders should be censured and blamed for their wrongdoings, there is a third and theoretically much simpler answer that should be mentioned. It might be suggested that, even though it is the case that a condemnatory message can be conveyed to an offender by the use of normal language or by symbolic means, this does not ensure that the offender understands that the message is really meant.¹¹ There are some cases in which a message needs to be followed up by certain actions in order for the addressee to understand that the message is really meant. For instance, Thomas Baldwin has suggested that if a lover is to communicate his or her love to the beloved, then merely formulating the right words may be insufficient (Baldwin 1999). In order to show that the message is really meant the words must be followed up by action; that is, the lover must spend time with the beloved or do whatever else is required to vindicate the claim.¹² Correspondingly, if the goal of a communicative theory of punishment is to ensure that the offender understands that the condemnatory message is really meant, then it needs to be accompanied by action; namely, hard treatment. A punishment, therefore, is a way of showing the offender that the message being conveyed is important and really meant. Would an expressionist theory thus construed satisfy the Necessity Requirement? Once again, I believe the answer to be in the negative.

The first question that arises is whether it is correct that an offender would not believe that a condemnatory message was really meant if it was presented by a judge and by the use of symbolic means of communication. This is an empirical and testable question, but there seems to be no evidence to back up the suggestion. However, even if we set this point aside and accept the basic assumption that a serious condemnatory message needs to be followed up by action in order to leave the impression that it is really meant, it still does not follow that the Necessity Requirement has been satisfied. To see this, consider the following analogy. Suppose that two persons each declare their love to the person they love and follow up this communication by proposing marriage. Suppose further that one of them has expressed an even stronger love than the other person. Would it follow then that the expression of the stronger love would not leave the impression that it was really meant simply because the action that follows it (the proposal) is the same as the action that accompanies the less strong expression of love? I think the answer is clearly in the negative. The point is that the accompanying action is sufficient to leave the addressee of the expression of love with the impression that the message is really meant. Correspondingly, it seems reasonable to believe that once a certain amount of hard treatment is imposed on an offender this will suffice to leave the impression that the condemnatory message is really meant (if the offender did not already believe this in the first place). But this means that the Necessity Requirement is not satisfied: the amount of hard treatment (punishment) need not reflect the magnitude of the desert object. Suppose this time that offender O₁ has committed a more serious crime than O₂, that hard treatment communication is used up to the level where both believe the message to be seriously meant, and that the extra censure that needs to be conveyed to O₁ due the seriousness of his/her crime is then communicated in other ways not involving hard treatment. In this case the communicative goal will be satisfied—O₁ is condemned more stringently than O_2—despite the fact that they are being punished equally severely.¹³ As this shows, this version of expressionism does not manage to sustain proportionality either.

In summary, what we have seen so far is that all three expressionist theories that have been considered fail to satisfy the Necessity Requirement and, consequently, do not succeed in justifying the principle of proportionality. The time has now come to consider whether expressionism fares any better when it comes to the second condition, namely, the Sufficiency Requirement.

This condition, as we have seen, concerns the question of whether a punishment is sufficient to ensure that the desert object is always imposed on the offender. To illustrate the point, suppose that one is under a moral obligation to deliver an important message. Suppose, furthermore, that one has sent the message but subsequently realizes that it has not reached the addressee. What should one do? The answer seems simple. Unless there are reasons to believe that the communication is always doomed to fail—which is usually not the case—one should try to deliver the message again. One might perhaps try to change or adjust the communication method, but the important thing is that the moral reason behind the first communicative attempt seems ceteris paribus to constitute an equally good reason to engage in a new attempt to reach the addressee. If the same thing could happen in an attempt of penal communication, that is, that the appropriate condemnatory message either does not reach the punished offender at all or does not do so to its full extent—that is, the offender is not sufficiently condemned—then there would seem to be a reason to re-punish the offender.¹⁴ However, this would constitute an obvious problem in terms of proportionality. Suppose again that the offenders O₁ and O₂ have committed equally serious crimes and have both been sentenced to one year in prison. Suppose also that O₁ gets the full condemnatory message, while the message fails to reach O₂. In that case, there would be a reason to re-punish O₂ up to the level where he or she has got the same message as the one that reached O₁. But giving O₁ one year in prison, while O₂ perhaps ends up with two years behind bars, would seem to violate proportionality.

It might perhaps be objected that this scenario seems highly unlikely. We all know that normal communication may well fail—in fact, it often does—but if one is using such a tangible communicative instrument as a punishment, then this will not fail to deliver the appropriate message. This answer, however, strikes me as clearly unconvincing. First, why should there not be some offenders who know that they are being punished for having committed a crime, but who do not in any way experience they are being condemned? Second, when it is kept in mind that what matters is not merely that a condemnatory message is conveyed to the offender, but that the message has a precise magnitude, such that he or she is condemned less than another offender who has committed a more serious crime, and more than an offender who has committed a less serious one, then it no longer seems plausible to hold that such communication cannot fail. That punishment should constitute the first kind ever of infallible communication is not a tenable claim. But, if this is correct, then it seems that the Sufficiency Requirement has not been satisfied and that this, as illustrated, means that expressionism has not succeeded in justifying proportionality.

In summary, what has been argued in this section is, firstly, that neither of the accounts of expressionism that have been considered succeeded in satisfying the Necessity Requirement. And, secondly, that neither does expressionism succeed in satisfying the Sufficiency Requirement. What this means, as illustrated, is that expressionist versions of retributivism fail to justify the principle of proportionality.

3 Possible Objections

What we have seen so far is that, in order for a retributivist theory of punishment to justify proportionality between the seriousness of the crime and the severity of the punishment, it will have to satisfy the Necessity and the Sufficiency Requirements. Moreover, I have briefly indicated that a version of retributivism that prescribes that what an offender basically deserves is suffering or hardship without further qualifications cannot deliver the goods (it failed with regard to the Necessity Requirement). Furthermore, I have argued somewhat more thoroughly that expressionist accounts of retributivism fail to satisfy both requirements. Finally, I have suggested—but not argued—that this problem of failing to satisfy one or both of the requirements is one that also confronts other versions of retributivism. For obvious reasons, it is not possible here to sustain the latter contention by going through each and every possible interpretation of retributivism. Space does not permit such an endeavour. However, suppose arguendo that this point is correct. Are there any ways in which one can meet the challenges that have been raised and, thereby, maintain the standard assumption that a retributivist distribution of punishment is proportionalist? Here are three possible answers.

A first objection concerns the assumption behind the two requirements, namely, that there is something—the desert object—that an offender deserves and that is given to him or her by imposing a punishment. This distinction between desert object and punishment may be questioned. Why not simply hold that what an offender basically deserves is the punishment? If this is the case, then it will no longer be possible that the desert object can be inflicted on the offender in other ways than through the punishment, and neither will it be possible that a punishment will fail to impose the desert object on someone who is being punished. In other words, both the Necessity and the Sufficiency Requirements will be satisfied. However, the answer to this approach is simple. Even though it is correct that a retributivist theory need not have a structure that involves a distinction between desert object and punishment, this is nevertheless the structure that some theories possess. In fact, this is the structure that characterizes some of the theories—such as expressionism—that have done the most to defend the principle of proportionality. Therefore, there will at least be some influential theories that are vulnerable to the inferential challenges that follow from the lack of satisfaction of the two requirements.¹⁵

A second objection concerns the simple fact that the principle of proportionality is a moral principle. It might be held that even if a retributivist theory does not satisfy the Necessity Requirement, that is, if the desert object can befall an offender in other ways than by a punishment, this does not show that alternative non-punitive ways of ensuring that the desert object befalls an offender are morally preferable. Therefore, the retributivist argument in favour of proportionality is not threatened by the fact that the desert object can be inflicted on an offender in other ways, as long as the theory is able to establish that the imposition of the desert object through punishment is morally preferable to alternative ways in which this object could be imposed on the offender. However, there are several answers to this objection.

First, the objection presupposes that it is possible for a retributivist theory to explain why the imposition of the desert object through punitive means is in fact morally preferable to other possible means. And, it should be kept in mind that it would have to be shown that punishment is the preferable method to ensure that the desert object in its full magnitude is inflicted on the offender. As we have seen, it was precisely at this point that the expressionist theory—which could be interpreted as possessing the structure underlined in this objection—failed. (It was hard to explain why punishments should be used as the only communicative means, even if one accepted either that a supplementary prudential disincentive should be given to the offender or that hardship is necessary in order to ensure a penitential reform in the offender, to underline that the condemnatory message was really meant.) Second, even if a retributivist theory provides arguments showing that the desert object, though it can befall an offender in other ways than through a punishment, ought to be imposed by the use of punishment, this will not suffice to justify proportionality in practice. To see this, suppose again—to make it as simple as possible—that the desert object is suffering simpliciter. As initially pointed out, suffering can befall a person in various ways. However, suppose that the retributivist was able to show that suffering through punishment is morally preferable to any other way of inflicting this desert object on the offender. Now, even if this is so, it could still be the case that suffering de facto has befallen an offender in other ways (even though these ways would not have been preferable had it been possible to choose how the suffering should be imposed). In such a case, it would still be wrong to punish an offender to the full extent (because he or she had already suffered to some extent). Thus, in practice one might still end up in a situation where two offenders have committed equally serious crimes, but where one should be punished less severely than the other because he or she had already suffered more than the other offender. What this illustrates is that, as stated in the Necessity Requirement, it is crucial that the object can only be inflicted by a punishment. Merely to contend that it should be thus inflicted will not help in ensuring a proportionalist conclusion. Finally, it should be noted that the objection would be of little use for retributivist theories that fail to satisfy the Sufficiency Requirement. The problems that this requirement may cause for the possibility of justifying proportionality cannot be avoided, even if one is able to establish that punishment is morally preferable to the imposition of the desert object in non-punitive ways.

The third and final objection that merits consideration would be to adopt a bite-the-bullet approach to the arguments that have been presented in the previous sections. It might be suggested that what really matters is not that there is proportionality between the seriousness of the crime and the severity of the punishment, but between the seriousness of the crime and the magnitude of the desert object. For instance, if what an offender deserves is to suffer, then the morally important thing is that the amount of suffering that befalls an offender should be in proportion to the gravity of the crime committed. Likewise, an expressionist might hold that what matters is that an offender is censured and condemned to an extent that reflects the seriousness of the crime. In other words, proportionality follows from the retributivist theory and should thus be observed, but as a relation between crime and desert object, not as one between crime and punishment. If this is what one holds, then one might of course acknowledge that the Necessity and Sufficiency Requirements are not always satisfied. But a punishment will function as an instrument that is used to ensure that an offender gets what he or she basically deserves. For instance, if the offender deserves to suffer, but has already suffered to some extent for reasons that have nothing to do with punishment, then the punitive reaction from the criminal justice system should take this into account by imposing only the suffering that remains to be inflicted in order to ensure that the offender gets what he or she basically deserves. And if an offender can get a condemnatory message by non-punitive means, then the punishment will only be needed to ensure that the full amount of censure and condemnation is imposed on the offender. Thus, on this bite-the-bullet approach the severity of punishment is not per se an element in the proportionality relation, but simply an instrument that is (sometimes) required in order to ensure proportionality between the gravity of the crime and the magnitude of the desert object.

If this is the answer to which a retributivist would resort, then it is obvious that there is no longer a challenge concerning the inference that the fact that one offender has committed a more serious crime than another leads to the conclusion that this offender should be punished more severely. The reason simply is that the principle of proportionality no longer expresses a relation between the gravity of a crime and the severity of the punishment. However, it should be noted that this way of biting the bullet comes at a cost. Meting out punishment on individual offenders becomes a rather complicated task. Even if it is assumed that all challenges concerning the comparison of crimes in gravity and the punishment in severity have been solved, one would in each case of sentencing have to consider to what extent part of what an offender deserves has already befallen him or her and, if so, what remains to be imposed in order to ensure that the offender gets all that he or she deserves. Also, it will have to be taken into account whether there are other non-punitive means that, at least partly, ought to be used in the infliction of just deserts. The appealing simplicity usually associated with the traditional interpretation of the principle of proportionality, namely that a more serious crime should be punished more severely and the equally serious crime should be met with the same punitive response, has been lost.

4 Conclusion

“Retributivism” denotes a wide range of different theories of punishment that, even though they usually share the view that an offender’s desert plays a crucial or at least an important role, may nevertheless vary significantly with regard to the more precise exposition of why an offender is deserving, and what it is that the offender basically deserves. However, despite the fact that these questions have been answered very differently among retributivists, it has also often been held that retributivist theories share the same basic view of penal distribution, namely, that they imply proportionality in punishment (or at least that constraints on what constitute acceptable upper penal levels should be proportionately determined). It is this belief that has been questioned in the previous sections. What I have suggested is that if a retributivist theory has identified what it is an offender basically deserves and uses this to explain and sustain the view that punishment is the appropriate response to crime, then, in order for the theory to inferentially establish proportionality in punishment, one will have to presuppose what I have called the Necessity Requirement and the Sufficiency Requirement. If a theory fails to satisfy one of these requirements, then the inference from the retributivist theory to the principle of proportionality will be blocked. As underlined, for obvious reasons it has not been possible to show that no retributivist theory has succeeded in satisfying the requirements. But it has been indicated that a theory holding that what an offender deserves is suffering or hardship does not satisfy both requirements. And it has been argued that neither do influential expressionist accounts of retributivism succeed in this respect. For the versions of retributivism that have not be considered here, the two requirements will constitute a challenge that needs to be addressed if one wishes to maintain the standard view that retributivism is intimately related to the principle of proportionality.