A traditional overall distinction between the various versions of retributive theories of punishment, which have dominated the penal theoretical field over the last three or four decades, is that between positive and negative retributivism.Footnote 1 While positive retributivism implies that there exists an obligation for the state to impose deserved punishments on perpetrators, adherents of negative retributivism hold the view that it is wrong for the state to punish a perpetrator more severely than he or she deserves. The latter view is deontically weaker, in the simple sense that it only places constraints on – and thus does not prescribe any obligations to – the imposition of punishment on perpetrators.
Positive retributivist theories – which capture the majority of theories in the modern retributivist field – have been challenged by negative retributivists and other penal theorists in several ways. For instance, it has been questioned whether the basic rationales behind these theories really succeed in establishing an obligation to punish. And even if this is so, it has been questioned whether it follows that the state is the part possessing this obligation.Footnote 2 In the following, I shall not engage in considerations of these questions. In fact, we might well assume that the questions have been satisfactorily answered by the positive retributivists. Rather, what we shall be concerned with in this article is a question which to a wide extent has been ignored in the exposition and discussion of positive retributivism, namely, what such theories, and thus the obligation to punish perpetrators, imply for a society in which the state has many other types of obligation. Two facts prompt a discussion of this question.
First, it seems obvious that a modern state possesses many types of obligation to provide its citizens with goods, services and protection, for instance, obligations to provide some degree of health care, education, protection by an army, decent housing, a political administrative system, elderly care, and so on. Whether theorists would subscribe to the existence of such obligations and, if so, to what extent, obviously depends upon what one regards as the most plausible theory of distributive or social justice, which means that there exists room for major disagreements. However, it seems clear that, like most other theorists, positive retributivists would usually accept the existence of at least some of these types of state obligation. In the following, I shall – and I believe uncontroversially – take this for granted. Second, it is a fact that the fulfilment of state obligations requires economic resources. This is the case with regard to the obligations to punish perpetrators which, and I shall return to this shortly, requires resources to cover what could be referred to as ‘pre-sentencing costs’, i.e. the costs related to tracking down and apprehending perpetrators, and ‘post-sentencing costs’, which include the resources required to impose deserved punishments on perpetrators.Footnote 3 Moreover, it is obvious that many of the other mentioned types of state obligations also require resources. However, and this is the second assumption, resources are usually limited, which means that it cannot always be assumed that all state obligations can be fulfilled.
Given these two facts – that there exist many state obligations and that resources are scarce – it is natural to ask what this implies for a retributive theory prescribing an obligation to punish perpetrators. In the following, I shall start, in section I, by considering whether the obligation to punish perpetrators should be interpreted either as an overriding obligation, that is, an obligation that trumps other state obligations, or as a pro tanto obligation, i.e. an obligation that should be weighed against – and thus may sometimes be sidestepped by – other obligations. It is concluded that the former interpretation is hard to sustain and that it is intuitively unappealing, but also that the latter interpretation of the obligation needs to be theoretically fleshed out in order to make clear what positive retributivism morally implies. Therefore, in section II, this question is considered in some detail. More precisely, several attempts to give a theoretical structure to the content of a retributive theory that interprets the obligation to punish as a pro tanto obligation are outlined. The interpretations are not rejected. Rather, the point of the discussion is to show that they have implications from which retributivists have typically strongly dissociated themselves. Thus, what I argue in this article is that the question concerning how to set priorities of scarce resources constitutes a genuine challenge for positive retributivism. Although a few theorists have directed attention to the resource question in relation to modern retributivism, the significance of this question has, I suggest, not been recognized. For instance, K. L. Avio, who is one of the few to have commented on the question, rightly contends that ‘While retributive explanations of the criminal law have undeniable intuitive power . . . they fail to indicate how many resources should be devoted to criminal justice.’ And, as he adds, this question must be addressed by ‘[a]ny serious theory of punishment’.Footnote 4 However, what I argue is not merely that retributivists have not yet provided an answer to the resource priority question – as if this were simply a matter of adding something to the theories that have been developed so far – but rather that the lack of an answer constitutes a direct challenge to the assumption that seems to underlie much of the discussion in this modern retributively dominated area, namely, that retributive theories have succeeded in providing a justification for the existence of a state-driven punishment practice.Footnote 5
I. PUNISHMENT AS AN OVERRIDING OR PRO TANTO OBLIGATION
A first possible answer to the question of how one should set priorities between a state obligation to punish perpetrators and other types of state obligation is to hold that the state has an overriding obligation to punish perpetrators. What this means is not that the only resource-requiring obligation the state carries is to punish those who break the law – as mentioned, very few, if any, would subscribe to this sort of view – but rather that the obligation to ensure that perpetrators get the punishment they deserve trumps other societal obligations. Thus, on this view it is only when perpetrators have been appropriately punished that the state should start devoting resources to other kinds of goods or services for the citizens. This outlook, at first sight, might seem to have some appeal. Many might feel that abstaining from punishing criminals out of considerations of resource priority would be highly dubious. The punishment of criminals does seem to constitute a core priority for the state. Moreover, in purely theoretical terms, the view possesses – in contrast to one that implies a weighing of obligations – a highly attractive degree of simplicity. Nonetheless, on closer scrutiny it turns out to be pretty hard to sustain.
The first thing worth noticing is that this position has rather radical implications. This is most clearly seen if we turn to what was referred to above as pre-sentencing costs, i.e. the cost related to the work of tracking down and apprehending criminals. If there exists an obligation to ensure that perpetrators are appropriately punished for their misdeeds then the suggested position seems to imply that the state should devote whatever resources it takes to tracking down people who have violated the law. This implication deviates significantly from how states usually operate. It is a well-known fact that many people commit crimes without being apprehended. First, there are many crimes which are never registered by the authorities and, at best, only revealed through some kind of victim survey or anonymous perpetrator survey. This is the case with regard to a plethora of less serious crimes. For instance, the majority of cases of transgression of speed limits, jaywalking or minor theft are not registered. And even some of the most serious crimes are known to have very low reporting rates, probably the most significant example being rape.Footnote 6 Second, even when crimes are registered by the authorities it is a fact that there is a difference – often a significant one – between the number of registered crimes and the number of crimes that are detected and cleared up. If there exists an overriding state obligation to ensure that all perpetrators get the punishment they deserve, it seems to follow that the state should devote much more money to the tracking down of criminals than is the case today. Perhaps thousands of extra policemen, combined with the newest technology and other types of procedures and methods, should be set to work to catch each and every jaywalker, speed-limit transgressor, private-property vandal, tax evader, thief or rapist. Some might perhaps object that this claim is premature; all that follows from a positive state obligation to punish perpetrators is that those who are found guilty in court should be properly punished. That is, positive retributivism has no implications with regard to obligations to track down and apprehend criminals. However, this objection is not convincing. If what matters is that people who break the law get the punishment they deserve, then it sounds rather odd to suggest that this position has no bearing at all on the question of whether the state should detect and apprehend perpetrators. And no one, to my knowledge, has seriously defended this sort of position.Footnote 7 Now, the mere fact that an overriding obligation to punish perpetrators has implications that deviate considerably from the resource allocation patterns we know in relation to existing criminal justice systems clearly does not itself constitute an objection against this sort of position. However, it does, as we shall now see, highlight what constitutes the main objection to such a view.
The second and more important thing worth noticing, in relation to the idea of an overriding obligation to punish perpetrators, is simply that this view is very hard to sustain in any plausible way. What is needed is an argument establishing that the state obligation to ensure that perpetrators are punished as they deserve is more important than any other type of state obligation. However, no one in this retributively dominated area of penal theory has presented an argument to this effect. Moreover, it is pretty clear that, at this point, it will not do merely to refer to intuitions. To suggest that large amounts of resources should be devoted to producing a slight increase in the number of detected jaywalkers if this is done at the cost of depriving some citizens of basic and perhaps even life-saving healthcare treatment, of maintaining a basic political administrative system, or of providing assistance for the poorest members of society, does not seem intuitively appealing. In fact, I believe most would agree that such implications strongly challenge the plausibility of this view. Thus, in the absence of an argument, I suggest that the view that the obligation to punish perpetrators constitutes an overriding state obligation should be rejected.
But if this is so, how then should positive retributivism be interpreted? The obvious alternative to the overriding interpretation is to contend that the state obligation to punish criminals constitutes a pro tanto obligation. On this view, the state has an obligation to ensure that perpetrators receive a deserved punishment. However, this obligation may not be decisive in all cases, that is, it may be outweighed by other types of state obligation.Footnote 8 This interpretation of positive retributivism is much more plausible. Since the view implies that there need not be an everything-considered obligation to punish perpetrators, it can be held that if one has to choose between the obligation to track down a few more jaywalkers and to provide citizens with life-saving healthcare treatment, then the latter obligation simply outweighs the obligation to pursue criminal justice. Thus, a view along these lines is from the outset well suited to dealing with some of the just-outlined counter-intuitive implications of the idea that the obligation to punish should be regarded as an overriding obligation. However, as indicated, this is done at the cost of ending up with a theoretically much more complicated position. In other words, the contention that the obligation to punish should be interpreted as a pro tanto obligation only constitutes an overall answer, which leaves it an absolutely open question how a weighing of state obligations should take place. More precisely, what we need is a more exact explication of the moral implications of the positive retributivist position. In the absence of such an explication, it is left absolutely unclear what positive retributivism implies in real life.Footnote 9 As a radical possibility, the retributivist view might be fully consistent with the claim that no perpetrators should be apprehended and punished, a possibility which positive retributivists are certainly not known for having endorsed. Thus, in so far as it is correct – which is taken for granted in the following – that the duty to punish should be interpreted as a pro tanto obligation, what would the positive retributivist theory look like on closer inspection?
II. RETRIBUTIVIST WAYS OF SETTING PRIORITIES
As mentioned, the question how priorities should be set between resources allocated to criminal justice and resources required to fulfil other important state obligations has not been thoroughly addressed by modern retributivists. It seems as if many retributivists believe that the theories they have developed to explain why perpetrators should receive deserved punishments are sufficient to justify the existence of a punishment institution. For instance, von Hirsch and Ashworth have defended an expressionist theory that combines the idea of penal censure and disincentives against criminal conduct; their theory is explicitly developed as an answer to ‘why the institution of the criminal sanction should exist at all’.Footnote 10 This theory, in addition to considerations on why the authority to punish should be entrusted to the state rather than to other parties, is apparently supposed to justify the conclusion that there is an everything-considered state obligation to punish perpetrators. However, if the above conclusion is correct – that the interpretation of the obligation to punish as an overriding obligation is not plausible – then, as we have seen, this conclusion is premature. Thus, how can the retributivist proceed from here? As far as I can see, there are several tracks to follow. However, as we shall now see, each track seems to lead retributivism to conclusions from which modern retributivists have typically strongly dissociated themselves.
The first possibility is to attempt to develop some sort of deontological theory of the weighing of obligations. It is well known that retributivists – despite various disagreements – usually share the basic view that retributivism is very much a deontological approach to punishment. But, although deontological ethical theories in general are usually characterized by the existence of constraints, it is also a fact that deontologists often subscribe to the existence of various sorts of positive obligations or duties.Footnote 11 Since the obligation to punish perpetrators is itself a positive obligation, what is required is a theory with regard to how positive obligations should be weighed. It is, however, important from the outset to realize what such a theory would have to address. First, it would have to be able to deal with what could roughly be called the inter-institutional obligations, that is, the theory has to be able to clarify how the state obligation to punish should be weighed against other state obligations (e.g. to provide healthcare, etc.). Second, the theory would have to be able to say something about intra-institutional priorities, that is, it would have to specify how the apprehension and punishment of some perpetrators should be weighed against the apprehension and punishment of others. With regard to the first task, it is not unfair to say that no retributivists have provided even a tentative answer to the question how this way of prioritizing should be carried out in a principled manner. As for intra-institutional priorities, this is also a question that has been left almost uncommented upon. Lippke, one of the few to have commented on this question, has suggested that if
we must make choices about which offenses to punish, in order to recognize and secure other deserved goods, then perhaps what we should do is prioritize deserved punishment demands, attempting first and foremost to punish those whose crimes are most harmful.Footnote 12
But obviously this is only a very rough outline of a way of prioritizing. For instance, it is far from clear how numbers should affect this way of prioritizing. Is it always preferable to track down and punish one perpetrator who has committed a serious crime rather than, if possible, five perpetrators of slightly less serious crimes? A question like this easily leads to what constitutes the cardinal challenge, with regard to both inter-institutional and intra-institutional prioritizing, namely, that it is absolutely unclear what the rationale behind the prioritizing should be. What should constitute the procedure on the ground of which priorities should be set? What are the parameters on the basis of which the weight of different obligations is determined? It is noteworthy that, when it comes to positive obligations in general within the framework of deontological ethics, questions like these have been only barely addressed. Thus, it is far from clear which theoretical resources a retributivist could draw on in an attempt to deal with this challenge. For the present we cannot determine precisely what the implications would be of a deontological theory concerning the weighing of obligations. Thus, it is not possible to compare such implications with the way in which positive retributivist theories usually present themselves (i.e. to determine whether the theory would have implications that the retributivist would be reluctant to accept). However, it is clear that no such theory exists and that it is very hard to imagine what it would look like.Footnote 13
Another possibility that might come to mind in the attempt to make up for the lack of a purely deontological theory of the weighing of obligations is to take a minor step in a consequentialist direction. More precisely, it might be suggested that the theoretical resources for dealing with the question of how priorities should be set is provided by a mixed theory of the sort traditionally ascribed to Hart. What Hart proposed was a distinction between the ‘general justifying aim’ of the practice of punishment and the principles of punishment distribution. More precisely, he held it to be ‘perfectly consistent’ that the general justifying aim is determined on the ground of utilitarian considerations, while the distribution of punishment is determined on the ground of retributive principles.Footnote 14 Though Hart himself employed only a negative retributive view on punishment distribution (that one should never be punished if one is not guilty of a crime and that those who are guilty should not be excessively punished), it might be suggested that the scheme could also be adopted by the positive retributivist by holding, on the one hand, that perpetrators should be punished in the way they deserve (i.e. in a way that is proportionate to the gravity of their crime), but, on the other, that the whole system ensuring that retributive justice is done must be defensible on utilitarian grounds. By adding the latter condition, it would seem that the positive retributivist has succeeded in providing an answer to the question of inter-institutional priorities: in so far as it is, in utilitarian terms, preferable to maintain a punishment institution than to devote resources to other societal institutions providing services and protection for citizens, this institution should be upheld. Thus, in contrast to the former proposal which, as we have seen, consisted merely in the claim that state obligations should be weighed against each other without any indications of how this should be done, this Hart-inspired proposal provides a principled guide to how the weighing should be carried out. At least so it might be suggested.
However, on closer inspection it is not as simple as that. Even if it is correct that one can distinguish between the general justifying aim of punishment and the distribution of punishment, it is not clear what this distinction precisely amounts to. In so far as the principle of distribution consists in a positive retributivist outlook, two interpretations of the mixed theory are possible.
On the one hand, the idea is that the state should do whatever it requires to ensure that perpetrators receive the punishment they deserve and if a system which works like this turns out – on utilitarian grounds – to be preferable to its non-existence then the whole institution is justified. However, this view faces the same problems that confronted the overriding interpretation of the state obligation to punish considered above. It implies that the state should do whatever it can to capture every jaywalker, every speed-limit transgressor, every pickpocket, and so on, and if this system turns out to be preferable to no punishment system, then it is justified even if some of the resources it requires to pursue justice in such a zealous manner could have been used for other important or perhaps even vital societal services. But, as mentioned, the view that one should still devote resources to track down the last bicycle thief, even if this implies that other important services cannot be provided, seems highly questionable. If one agrees that the overriding interpretation of the obligation to punish considered above is morally implausible, then I take it that this interpretation of the mixed view should also be rejected.
Alternatively, the mixed view could be interpreted as implying that the criminal justice system should attempt to pursue whatever positive retributivism implies with regard to punishment, but that the amount of resources that should be allotted to this purpose – and thus the extent to which this is made possible – is determined by a utilitarian way of prioritizing between various societal institutions and tasks. Thus, on this view the question is: given the fact that the tasks of other societal institutions require resources and that the criminal justice system attempts to operate in accordance with positive retributivist prescriptions, what then constitutes the – from a utilitarian point of view – optimal way of pursuing the retributive goal? For instance, would a system that ends up imposing deserved punishment on 70 per cent of perpetrators be preferable to one that imposes deserved punishments on 40 per cent? However, this interpretation opens up the possibility that only a very few resources are devoted to the criminal justice system, that is, it might imply the criminal justice system could end up with resources to track down and punish only a minor fraction of perpetrators. In fact, it is possible (even if in reality hardly likely) that this interpretation would imply that there should be no punishment system at all. But this is clearly an implication which theorists in the positive retributivist camp would find hard to accept. Perhaps some theorists with stronger consequentialist inclinations would be fully willing to accept this implication. But I take it that most positive retributivists would regard this as much too large a concession in the direction of a consequentialist approach to punishment. Thus, in sum, the idea of making the positive retributivist view work within the structure of a Hartian mixed theory, and the hope that this would constitute a way of avoiding the question of intra-institutional prioritizing while at the same time providing a principled device for inter-institutional prioritizing, does not hold water. Either the retributivist is led to some of the same rigorist justice-focused implications as the overriding interpretation of the obligation to punish, or she is forced to accept the possibility of implications that would constitute a significant deviation from how positive retributivists have standardly presented their views on punishment.
The final possibility of a way of reacting to the fact that there are various types of state obligation beyond the obligation to punish perpetrators is to accept the even more radical step of interpreting retributivism as a genuine consequentialist theory. As mentioned, retributivism is standardly interpreted as a non-consequentialist approach to punishment. However, it might be suggested that, though desert is still what matters morally, the deontological structure of the theory ought to be abandoned. In fact, this is precisely what has recently been suggested by Cahill. Considering the question whether a retributive theory of justice can explain how priorities should be set between the various tasks and players within the criminal justice system, Cahill ends up suggesting that the only version of a desert-based theory that succeeds in dealing with this question is a consequentialist version of retributivism that sets the goal of ‘maximizing the total amount of desert-based punishment’.Footnote 15 By adopting a consequentialist structure, Cahill believes it is possible to determine priorities within the criminal justice system. But obviously it might further be suggested that a full-blown consequentialist theory – perhaps, most likely, one containing intrinsic values beyond desert – would be able (at least in principle) to determine priorities between resources provided for the punishment institution and other important societal institutions. Thus, in contrast to the former suggestions, we may now have a theory that succeeds in delivering a device for both intra- and inter-institutional prioritizing.
Cahill does not consider more closely the possible implications of the suggested version of consequentialist retributivism. However, it is a fact that such a theory – depending on how precisely it is worked out – may have noteworthy implications. For instance, it might imply that a society with criminals who are punished is preferable to a society without any crimes; the former society constitutes a state of affairs of more intrinsic value. Of course, if the theory employs other intrinsic values than deserved punishment – and, as mentioned, this seems reasonable to assume – then the implications for crime victims would probably count negatively in the overall calculus. However, the theory at least implies that a society with many punished criminals is in one respect preferable to the crimeless society. I take it that some would regard this as a dubious implication. In fact, given the hypothetical assumptions that deserved punishment has positive value, that a person will be punished if she commits a crime and, finally, that there are cases where this positive value outweighs the disvalue caused by the crime, it follows that a person might have an obligation to commit a crime. Committing a crime and being punished for it would produce a state of affairs preferable to one in which no crime is committed.Footnote 16 Furthermore, the theory also has another type of implication. In so far as intrinsic values are maximized by punishing an innocent person, then this is what ought to be done according to the theory. This possible implication is particularly interesting in light of the fact that the punishment of the innocent is traditionally regarded as the main retributivist objection against a utilitarian approach to punishment. But obviously what generates this objection is not the value-theory inherent in utilitarianism – that well-being is the only thing of intrinsic value – but rather the consequentialist nature of the theory: that values should be weighed against each other and that the right action is the one maximizing intrinsic value. But this means – as has been convincingly argued, for instance, by Dolinko – that scenarios can also be imagined in which a consequentialist theory holding deserved punishment as intrinsically valuable would have this sort of implication.Footnote 17
Now, the point of directing attention to the different implications that might follow from a consequentialist retributivist theory is not to provide an overall evaluation of this sort of theory. This would require a much more thorough discussion than the somewhat rough and sketchy comments presented above. Rather the point is to indicate that if the retributivist, in order to reach a theory that is apt to deal with the question of how resources should be prioritized, would opt for a consequentialist version of retributivism, then she would have to accept implications which retributivists have usually regarded as highly unacceptable. Thus, to sum up, what we have seen is that, what a purely deontological theory for the weighing of obligation would amount to – and, thus, what implications it would have – remains unclear. Moreover, attempts to make up for this by letting positive retributivism work within the structure of a mixed theory or a full-fledged consequentialist theory would have implications deviating significantly from what positive retributivists have traditionally regarded as morally acceptable.
III. CONCLUSION
The foregoing discussion has addressed the question of how one should understand the positive retributivist contention that there exists an obligation to punish perpetrators. Given the assumptions that there exist other types of state obligations, that the state obligation to punish perpetrators as well as other positive state obligations requires resources, and that resources are limited, it is reasonable to ask how the state obligation to punish perpetrators should be weighed against other obligations. It has been argued that the idea of interpreting the obligation to punish as an overriding obligation – that is, as an obligation that trumps other state obligations – does not seem plausible and would be hard to sustain. And it has been argued that the interpretation of the obligation as a pro tanto obligation is more promising. However, this answer prompted the question how the obligation to punish should be weighed against other obligations. Some of the possible answers to this question and some of the challenges confronting these answers have been outlined. As mentioned, the point has not been to refute any of the possible replies, but, more modestly, to point out that these answers face difficult theoretical challenges or have implications which retributivists have usually not been willing to accept. The latter point, in my view, confirms that there is a case for bringing the question of resource priorities into focus. The positive retributivist, in sum, is confronted with a genuine but widely ignored challenge which, from a standard retributivist perspective, does not seem to allow for any easy solutions.
Three critical comments may at this point come to mind. First, some might find it worth emphasizing that the challenge that has been raised is, after all, not a problem confronting retributivism in general but only positive retributivism. However, the reply to this is simple. It should be noted, first, that even weaker versions of retributivism are also confronted with the outlined problem. For instance, some retributivists defend the position that the state does not possess an obligation to punish perpetrators but, more modestly, that it is permitted to do so.Footnote 18 However, even this deontically weaker position does not avoid the question whether it should be interpreted as meaning that the state is always permitted, independently of what a punishment requires of resources that could have been used for other purposes, or rather whether it is from the outset permitted but might everything considered (weighed against various types of state obligation) be wrong. More importantly, it should be recalled that prominent modern versions of retributivism are in fact positive retributive theories. Thus, the limitation of the focus in the above discussion clearly does not mean that it is devoid of theoretical significance.
Second, it might perhaps be objected that it is fairly reasonable to believe that, given the wealth of western countries, one should uphold a criminal justice system imposing deserved punishments on perpetrators. The discussion might therefore be accused of being more (and perhaps unnecessarily) theoretical than practically pertinent. However, this answer is premature. It is a fact that even wealthy nations have limited resources. Thus, merely to suggest that within these limits a penal system should be effectuated and should be operated as positive retributive theory requires, is simply to presuppose rather than provide an answer to the question of how state obligations should be weighed.
Finally, some might feel that the discussion has somewhat unfairly addressed the resource question as a challenge to retributivism. It might reasonably be held that this problem is one that confronts any theory of punishment, including that is, traditional rival theories of punishment. However, once again the answer is simple. That the question of how resources should be prioritized confronts all theories of punishment is correct. But it might be said of some theories, such as perhaps the utilitarian one, that they, at least in principle, possess the theoretical resources for dealing with the question of how resource priorities should be set. And, moreover, even if this is not the case for all theories, this sort of tu quoque objection does not make the challenge less of a problem for retributive views – which, after all, dominate the modern penal theoretical field.
