Accueil RED&S

On Crime and Time

Nikolas H.M. Roos
Law Faculty - Maastricht University


Télécharger ce document au format RTF


On Crime and Time a été publié dans une
version moins elargie dans : François Ost & Mark van Hoecke, Temps et Droit. Le
droit a-t-il pour vocation de durer ?, Bruxelles 1998 (Bruylant), pp. 385-424


The question to be explored here is how the criminal law (hence: "CL") operates with time (internal time) and how this has changed in historical time (external time). Internal time is determined by the temporal criteria the CL sets to make human behaviour qualify as crime, to organise the criminal procedure as a process in time, and to relate, as a function of time, sanctions and crimes. The historical dimension, the evolution of the CL and its time structure, is important for the reflection on how to interpret considerable changes in the internal time of contemporary CL.

We will begin our analysis with the transition from the pre-classical to the classical model of the CL, starting in the second half of the eighteenth century. The development of the classical model was part and parcel of Enlightenment. Its spirit was to secularise, rationalise and humanise the CL. It will be shown (section 2) that the classical model followed a functional approach of crime and punishment, which it regarded as radically temporal and empirical phenomena.

The validity of such an account of the classical model can be challenged because there has not been a single justification of the classical CL. Absolutist theories focus on retribution as the CL's primary function, relativist theories on prevention. Absolutist theories certainly do not take a functional approach to crime and punishment. However, my claim about the functional nature of the classical model is both historical and critical. My contention is that the founding fathers of the model (Beccaria, Feuerbach and Bentham) were relativists, whereas retributivism cannot account for a number of very typical features of the classical model (section 3).

Absolutists have always argued that the relativist view is incompatible with one or more of the principles of criminal justice. In that case, relativism would not provide an adequate account of the classical model either. In fact, the most well-known modern legal utilitarian, H.L.A. Hart, frankly admitted that the utilitarian approach to criminal punishment was inherently incompatible with the principles of criminal justice (section 4). I will argue that Hart was mistaken and that the relativist view can be made entirely coherent if cast in a framework of an appropriate political theory. In fact, the theory to be proposed, the so-called "Janus-face view of the CL", can be grounded in Hart's very own ideas concerning a "minimal content of natural law". It integrates retribution and prevention in one coherent theory. Moreover, it is not only a theory of philosophical interest. It can also set a new critical criminological research agenda (section 5).

In section 6 a krypto-religious element within classical relativism will be analysed, which can help to explain both Foucault's famous connection between disciplining and punishing and the fact that classical relativism did not develop a Janus-face view sort of theory of the CL. The fact that they were so preoccupied with the effectiveness and the efficiency of the CL explains why they were much less interested in the principles of criminal justice as political principles. In fact, the interest of the early absolutist in them was quite political, as will be argued in section 7. Moreover, in the struggle with the inherent weaknesses of the absolutist position, the Dutch legal philosopher Leo Polak came very close to a theory of the CL like John Rawls'. Rawls view comes close to the Janus-face view again, although there remains a very important difference with the latter (section 7).

Whatever the virtues of the Janus-face view of the CL, it has definite limits, because it is based upon a classical model or a neo-classical model (analysed in section 8) of the CL. The post-classical CL would seem to lack the inner coherence of the (neo-)classical model. In fact, it even has some pre-classical features. It is as if contemporary CL is falling apart in pieces which can no longer be held together under the umbrella of the general principles of the CL. However, in my view these pieces may find new coherence in the social field they apply to in interaction with other types of law relevant to that field. I will suggest to interpret this process of legal resystematisation as a further step in the process of functional differentiation and dynamisation of the CL. Although there is, from this perspective, a continuity in the development from the classical (and neo-classical) model to post-classical CL, the post-classical transformation is much more dramatic. The difference in time structure of post-classical CL is reflected in a structural reversal of the two components of the Janus-face view, punishment and rehabilitation. This reversal suggests that a normative understanding of post-classical CL requires a justification of criminal risk responsibility (section 9).

2. The epistemology of the classical model

The temporal framework of the pre-classical model still was a christian-religious one. I propose to call it the Augustinian model. In that model the state figures as the temporal representative of a supra-temporal order, the world of eternal happiness (a world without time) from which man has fallen due to the original sin. The state has been established by God as an act of grace to give mortal sinners a chance to escape from eternal damnation in hell, a subtemporal world in which time, pain and suffering never ends. The state fights a war against crime, because crime is an instrument of Evil to destroy the state which, by preventing anarchy, is an absolute condition of redemption. As the state attempts to save us from hell, it is no wonder that its criminal practices might seem to preempt those of hell. Punishment is very often a physical form of punishment organised as an intimidating public spectacle to warn the spectators to keep their sinful inclinations within bounds. The CL foreshadows the eventual separation of the black sheep from the white sheep. The fact that the state is a temporal institution nevertheless, implies that it is bound to make mistakes in doing so. This explains the great importance of confessions in the CL, to be extracted, if necessary, through torture, which kept traits of an ordeal.

By contrast, the classical doctrine tended to treat crime and punishment as purely secular phenomena for the first time. Crime was no longer seen as directed against a religious or political order as such, unless it would be in the form of special political crimes. The aim of the designers of the classical model was to put an end to the abuse of the CL against people accused of things not known to be a crime, people accused of unobservable things and people sentenced without facts or guilt having been proven. The advocates of the classical model wanted the criminal sanction to be useful at the lowest possible cost for both society and the offender. They wanted to put an end to the capriciousness, cruelty and inefficiency of the CL as it existed under the conditions of feudality and of unenlightened absolutism. The classical programme was, in short, a programme of modernisation.

Although the classical view is indebted to the work of Montesquieu and Voltaire, it came to a first influential expression in Cesare Beccaria's Dei delitti et delle pene, published anonymously in 1764. However, one of the reasons of its almost immediate success may have been that it had been written as a pamphlet rather than as a systematical treatise of the foundations of the CL. The classical approach was first developed as a truly theoretical system by Jeremy Bentham in England, notably in his An Introduction to the Principles of Morals and of Legislation (1789) and the Traités de Législation civile et pénale (1802) and Théorie des peines et des récompenses (1811). In Germany, the classical approach was presented as a systematic theory by Paul Johann Anselm von Feuerbach, notably in his Revision der Grundsätze und Grundbegriffe des positiven peinlichen Rechts (1799/1800) and in his Lehrbuch des gemeinen in Deutschland geltenden peinlichen Rechts (1801).

The works of these three authors in the field of the CL have been very influential. Beccaria's work inspired much of the reform in several European countries during the second half of the eighteenth century (Sweden, Prussia, Russia, Austria, Tuscany), but most of all, and in a more systematic fashion, the revolutionary Code Pénal 1791. Feuerbach has been directly influential as the main author of the Bavarian Penal Code (1813), which was to serve as a model for the later Prussian - (1851) and Imperial German Penal Code (1871). Bentham's influence has been greatest in England although it was slow in coming and his proposal for a Penal Code did not materialise. His philosophical ideas reinforced the generally utilitarian English reformist spirit that was responsible for the later acceptance of so many of his concrete proposals. Bentham had a more immediate influence on the continent, notably on the Code Pénal of 1810. He also became directly involved as a reformer when he was invited as a commentator to the draft Penal Code of Spain in 1821.

The secular view on crime and punishment of the Enlightenment implied that they had to be dealt with as purely temporal, that is, as empirical and functional matters. The CL was constructed as a social system that has an end in itself, which is the reduction of the amount of crime as well as of the human and social costs of criminal punishment. The rationality of its use became crucially dependent on temporal factors. The designers of the classical system had to answer the following questions. How should the facts it should operate upon be identified as events in time? How should they be processed through the criminal procedure as an ordered time sequence? How should punishment function as a feed-back mechanism, that is, as a mechanism which relates events in the past to projected events in the future?

A. The object of the CL

We will first analyse the temporal epistemology of the subjective side of the object of the CL, the conceptualisation of the criminal actor. From a functional point of view it simply makes no sense to address norms to people when they cannot help them controlling their behaviour. By consequence, only behaviour which is normally under the control of a person can be the object this person may be criminally accused of. Therefore, one can also not bear criminal responsibility for the acts of another person unless one had both the responsibility and the possibility to prevent that person from committing a crime. Direct control, therefore, presupposes a certain degree of rationality, and it excludes criminal responsibility of animals, the mentally insane and children under a certain age.

A corollary of the individualist-functional conceptualisation of the (criminal) act is the principle of equality. No class privileges were allowed to figure in the CL as was usual under feudal conditions, when the clergy and noblemen had courts, procedures and material norms of their own. This is not because there can be no exceptions to the principle. However, from a functional point of view there is no argument for categorical privileges. An MP or a medical doctor can refuse to make declarations as a witness, but only if and in so far the declaration demanded is related to their functioning as such.

The temporalised-functional conception of the CL also has implications for the conceptualisation of the objective side of a crime. A crime and its sanction must be defined in advance and must be specific enough to be applied to events which can be located precisely in time: 'nulla poena sine lege, nulla poena sine crimine, nullum crimen sine poena legale' (Feuerbach). A person cannot bear criminal responsibility for an act he could not know to be a criminal act. In a functional approach there are no crimes "by nature" any more. Whether or not a particular illegal act can be punished or can only be sanctioned via other legal means, is, in principle, a contingent matter, depending on relative efficiency. Another functional reason supporting the legality principle is that if knowledge of the severity of the sanction one can expect is a functional condition of the effectiveness of punishment as a feed-back system, it is important that the connection is specified by the law. However, that consideration belongs to the analysis of the criminal sanction and its consequences.

To be identifiable in time, it is necessary that an act must have been observable. It cannot be an act that was merely a projected act. However, if an act can be recognised as a particular criminal act such that, as far as it depends on the actor, it is most likely to have the consequences the law tries to prevent, the act is no longer merely a project. Such an act is called a criminal attempt if it does not have the consequences the actor intended it to have. By implication, an act can be a criminal attempt even though an offender might still have interfered with the normally expectable consequences, but could not do so because factors out of his control made it impossible. Therefore, the CL can define an act in a way that surpasses the actual facts, but it does it in a defeasible way. If the offender stops out of his own free will before the criminal act has been completed, it will not speak of a criminal attempt. This defeasible epistemology implies that two cases which might have been the same from a supra-modal point of view, may be treated unequally.

B. The criminal legal procedure

The temporalisation of the CL in the classical model is a radical sort of temporalisation. Not only the object of the CL is seen as a temporal matter, so is the CL itself. It is a homeostatic, secondary and therefore "a fortiori" temporalised system. The secondary nature of the system implies that it must monitor the primary system 'in time' and order its procedures sequentially according to the inner logic of the criminal legal process from the moment of detection up to the conditions of the execution of the sentence. Therefore, it must also monitor its own 'timely' operation. The system does this, among others, via rules on dismissal of the prosecution and through statutes of limitation, after which a prosecution may not even be started any more. The system's secondary nature and the fact that its own malfunctioning is worse than not functioning at all, is reflected, for instance, in the in dubio pro reo-principle, which is one of the subprinciples of the praesumptio innocentiae as expressed in art. 9 of the Déclaration des droits de l'homme et du citoyen of 1789. In fact, all of the major principles of the criminal legal procedure, which are usually understood as having their origin in the humanistic values of the classical system, can be understood as characteristics of the criminal legal system (hence: "CLS") as a temporalised feed-back system.

The presumptio innocentiae is primarily a procedural principle. Not only would it otherwise contradict the presumptio nocentia that is required to arrest someone in connection with a crime, but it would also be groundless in all cases in which an offender has been caught in the act or a clear and convincing confession has been made. The presumptio innocentiae has important implications. One is that the burden of proof is carried by the state, with the right of the suspect to remain silent as a corollary. Another implication is that judges should not only be independent and objective, but should also not express a (final) judgement before the whole procedure has been carried through. An implication of the demand for neutrality is that a judge that has been involved in the inquisitorial phase of the process, cannot participate as a judge when it comes to the trial phase. The right to counsel from, in principle, the moment of arrest, also follows from the presumptio, as an arrested person will need an intermediary to have access to the factual and legal information necessary for his defence. Furthermore, protection against unjust detainment during investigation will increase the longer a suspect is being detained. This reflects the temporalised nature of the process, which, if delayed too long without good reason, may be ordered to stop altogether.

Finally, the ne bis in idem-principle, the fact that, in principle, one cannot be put on trial for the same act after a final verdict, also expresses the temporalised nature of the criminal process. The "same act" refers to a complex of facts in space and time and not to the possible plurality of legal qualifications of those acts. It may be that later facts show a suspect to be guilty far beyond reasonable doubt, but once a crime has been declared unproven in a (final) judgement, the case belongs to the past as far as the CL is concerned. The asymmetry that the principle does not apply if new facts prove a condemned person to be innocent, reflects the secondary nature of the system, which only takes its own malfunctioning into account when, on hindsight, it should not have started to operate to begin with.

C. The criminal verdict, the sanction and its consequences

Let us now take a look at the temporal construction of the verdict, the sanction and its consequences. In a purely functional perspective a sanction makes sense only if it feeds back into the criminal event and its consequences. Otherwise there would be no reason to take the occurence of crime as the point of departure of the system. Therefore, care must be taken that the criminal fact, the identity and the guilt of the offender(s) is proven through means whose causal link to the crime will establish it beyond reasonable doubt.

The pre-classical system was static in its self-reflection. It assumed that the wisdom of the tradition would be able to absorb all possible cases. Changes in society, if occurring at all, were seen as recurrent stages of society which were already reflected in the traditional legal wisdom. Moreover, the pre-classical CL had kept, at the same time, features of the supernatural. Acts of God do not demand to be rationally understandable, on the contrary, they represent superior forms of wisdom. Motivating sentences could be a desacralisation of that wisdom.

By contrast, in the classical model the verdict is the point in time when the system changes from its cognitive phase to its pragmatic phase, the actual process of feed-back. A temporalised system can only gain an identity through change if it can account for its development from moment to moment in time and, quite disrespectfully, declare parts of its past irrelevant if it finds them to be mistaken on hindsight. Unlike supernatural systems, it cannot claim infallibility. Traditionally the duty to motivate is seen from the perspective of the rights of the defence and the legitimation and political control of the system. Epistemologically it can be seen as an act of self-control, an account of how a case fits into the system and checking if a case demands an adaptation of the system.

Given the feed-back functional nature of the system, purely emotional or religious grounds for punishment, in so far as they have no causal effects on the occurrence of crimes and their negative consequences, are irrelevant for punishment. Moreover, even if a feed-back relationship between crime and sanction exists, punishment may not become just an occasion to exploit offenders in ways which have nothing to do with punishment as such. This would be a purely external sort of rationality of the CL, because it does not support punishment as a feed-back mechanism.

Even if the feed-back relationship between the crime and the sanction is an internal one, its rationality still depends on the structural or non-structural nature of that relationship. From a functional-temporal perspective the structural effects the CL should aim at are the prevention of crime or the compensation of victims, if there are any. However, norms of the CL can also have non-structural, ultimate aims which are not automatically realised if the CL's structural effects are. In fact, they can be incompatible. For instance, incarcerating entrepreneurs for tax fraud may have a deterrent effect, but it may also have negative effects upon the activities of some of these entrepreneurs such that the economic and tax losses resulting from punishment are greater than the gains. But that is not something which can be easily taken into account by the CLS. The operationability of the system depends on an abstraction from the usually many, variable and uncertain factors that determine the non-structural effects of criminal sanctions. As functionally differentiated social systems it is typical for both the classical model of the CL and its successors that they take abstraction from the systems non-structural effects. They operate, to use Luhmann's terminology, as operationally closed systems.

Operational closure is sometimes misunderstood as an exclusive input-orientation of a system, whereas it means nothing but a systemic selectivity of purpose. In fact, the evolution from the classical to a post-classical CL can be seen as a process in which more and more conditions of the effectiveness of criminal punishment are taken into account. Because of this increasing output-orientation the complexity of the CL increased, whereas the differentiation between the system's structural goal, the prevention of crime, and the ultimate, political goals of criminal punishment, which depend on the specific nature of a crime, became less clear.

Operational closure is not a sufficient condition for a system to be structurally effective. Given the structural goal that a system sets itself, there still is a need to choose between different possible means to realise that goal and to structure the operation of the system accordingly. In fact, a CLS must be highly selective as to what it wants to aim at (e.g. special deterrence, general deterrence or appeasing victims). Aiming at too many effects at the same time or not putting these aims in an order of priority, may make its internal structure incoherent and unmanageable (see next section). This problem only becomes the more pressing as the CL becomes more output-orientated.

The classical system still was an almost purely reactive system. It took about a century before a neo-classical model established itself. However, it was still a reactive system. It was only more output-orientated in the sense that it was less blind for the ineffectiveness or the counterproductivity of the classical system. It tried to cope with its ineffectiveness by the development of what was later to be called "alternative" sanctions: education, therapy, counseling or, if nothing would help and an offender was thought to be dangerous, incapacitation. The restrictedness of its output-orientation explains why it was perceived as shifting the emphasis of the CLS from general prevention to special prevention. However, in my view it is correct to speak of a neo-classical model precisely because special prevention remained a secondary function within a system that kept its fundamental classical traits (see next section).

As we will see, the post-classical CL is no longer just a reactive system, but it has traits of an interactive system. An interactive system does not take its environment as given, but it tries to influence it in order to prevent crime or to handle it more efficiently. Moreover, it will adapt itself to changes in its environment, changes which it may have induced itself. However, being an interactive system does not imply that there is no operational closure any more, because it is still the incidence of crime that is the primary concern of the system.

3. The coherence of function and structure in the CL

As explained in the last section, the classical model of the CL as a radically temporalised system presupposes the existence of a negative feed-back relationship between the criminal sanction and crime. It was also stated that such a system would have to be highly selective as regards the function of the criminal sanction because its inner structure would depend on it. In this section I would like to show by way of a logical reconstruction that the modern CL has, indeed, made such a selection, at least tendentiously, because the selection has not been equally consistent in all countries.

The way I will operate is to list plausible structural consequences of the CL and then analyse which of these are compatible with classical legal doctrines which make the sanction or its size dependent on time factors. I will look at the following four potential structural functions of the CL:

  1. retribution
  2. appeasement of victims
  3. special deterrence (SD)
  4. general deterrence (GD)

Three of the four functions are obviously connected to rival theories of justification of the CL. Only appeasement of victims has a somewhat unclear relationship to them, because appeasement would seem to be a consequence of the other functions, or at least to be based on the belief in these other functions. The reason to put it on our list, is that appeasement would also seem to be a useful effect of incapacitation. Therefore, if we exclude incapacitation from our list, as we did, appeasement of the victims should be put on it as a possible aim of incapacitation. I have also not included norm confirmation and rehabilitation of the offender in the list. They are obviously possible secondary effects of the functions already listed.

As according to our thesis relativism was the twin brother of a radical temporalisation of the CL, it makes sense to see to what extent the different functions listed can explain doctrines which connect temporal aspects of the criminal act with the applicability or the size of a criminal sanction. The doctrines to be analysed are not universal in western legal systems. However, majorities of these doctrines figure in all of these systems. Moreover, even though a doctrine may not exist as such in a particular CLS, it may nevertheless be followed in practice, or exist 'under cover' in other doctrines. The legal doctrines that will be reviewed are:

  1. impossible crimes
  2. criminal attempt
  3. periods of limitation
  4. plural concurrence
  5. recidivism

A. Impossible Crimes

In all continental CLS's impossible crimes remain unpunished. Impossible crimes are part of the laughing stock of the CL. A person thinks he is committing a crime, but he does not. The reason can be a defect in the object of the crime (e.g. an abortion on a woman that is not pregnant), a defect in the means (e.g. poisoning believing that what is in reality sugar is strychnine) or because of the legal relationship between the would-be offender and the object (e.g. stealing a bicycle that happens to be one's own). Impossible crimes are not punished. Retribution cannot explain this, because there is an actus reus. An impossible crime can create great indignation among the victims that were targeted and, therefore, appeasement cannot explain this either. Nor can SD because the fact that the would-be offender made a fool of himself, is no guarantee that he will not try again. Quite the contrary. However, GD can explain why no sanction follows. The crime has aborted itself and the legal community can laugh about the foolishness of the would-be offender. Because of that foolishness he sets no example of a successful crime. Punishing such crimes might even be counter-productive. Offenders might take more precautions not to make mistakes.

B. Criminal Attempt

In many countries criminal attempts are punished less severely than the completed crime. In the Netherlands, for instance, the maximum penalty will be one-third lower. By contrast, in France and Germany no distinction is made between attempted and completed crime as far as the maximum penalty is concerned, unless the law determines otherwise for a specific crime. Sometimes this takes the form of special crimes for preparatory acts. The characteristic of a criminal attempt is that the crime has not been completed because of factors that were not under the control of the offender. Thus, if a thief ends the job because he has finally decided not to steal, there is, in most CLS's, no attempt in a legal sense. However, if he decided to do so because of an alarm, and is then caught, he will have made a criminal attempt. Retribution cannot explain the differentiation between attempt and completed crime because there is no difference between them as far as 'mens rea' of the offender is concerned. Appeasement is not a very good explanation either. Suppose that the thief is caught by the owner-victim after having stolen. This may be much more satisfying for the victim than if he had caught the thief before the act was completed. SD will also not suffice as an explanation. Why would an unsuccessful offender be less dangerous than one who completed the crime? If frustration would be a recidivism preventing factor, one would also have to take crime-frustration into account after the completion of the crime. It must be much more frustrating for a thief to be caught after getting stuck in the elevator on the way out than being caught before he had his hand on the loot. Again, GD can explain a lesser penalty. A criminal attempt is a failure to upset the legal order. However, differently from impossible crime, not punishing a criminal attempt might stimulate both the frustrated offender and others, who think that they can do better than him, to try it again. GD can also explain why it is wrong to take post factum-frustrations of crime into account. If offenders who have been caught with the loot would be punished less than those who were caught after they had hidden it, the more severe penalty in the latter case would suggest that the thieves have earned themselves a title for keeping what they have stolen.

C. Periods of Limitation

Periods of limitation after which a crime can no longer be prosecuted, would seem to accord with theories of retribution and appeasement. After a certain period the need and desire to keep the offender in check or to teach him a lesson may have subsided. However, both theories cannot explain the all or nothing-nature of periods of limitation. One day before the end of the period a crime can still be punished as severely as when it was just recently committed. One day after it, it cannot be punished at all. If the desire for retribution or appeasement becomes less over time, it would be more logical to have a gradual decline in the maximum penalty rather than an all or nothing-limitation. The same goes for SD. At first sight one might think that periods of limitation make sense in connection with SD. If a person has not committed another crime for a longer period of time, it is plausible that he has become a law-abiding citizen again. Again, however, the all or nothing-nature of the period of limitation is therefore unexplainable, whereas the whole argument fails if the person in question has committed new crimes after the one which could no longer be prosecuted once the period of limitation had elapsed. If we then consider GD, at first sight it may seem as if periods of limitation are undesirable. The more crimes are punished, the stronger will be the deterrent effect. However, the longer it is since a crime has been committed, the less will be its negative effect on the trust in the CL as a deterrent of crime. The longer ago the crime was committed, the more it will disappear from the collective memory. It may be wise to let a crime rest then because a prosecution will bring back those memories back to life. The risk of disappointing the victims whose memories are reactivated when it comes to sentencing is also considerable, as many a war criminal's criminal process has demonstrated. From the point of view of the economics of GD, it is more important that especially those crimes which are solved and prosecuted are those which are still relatively fresh in the collective memory. Moreover, proof of a crime will become more difficult as time goes by and it is important that the police and the public attorneys invest their very scarce and expensive energies in cases which have a good chance of leading to end with the imposition of a criminal sanction. However, all these considerations are stll not conclusive as far as the all or nothing character of periods of limitation is concerned. However, from the point of view of GD it is important that potential criminals will not be deterred less by the prospect that they will be punished less the more time passes before they are caught. In fact, periods of limitation should not be so short in that potential criminals may speculate that the length of that period increases their chances of impunity noticeably.

D. Plural Concurrence

Plural concurrence refers to an offender who has committed two or more pragmatically unconnected crimes before he is charged in a criminal court. For example, someone who breaks a window in order to be able to steal something, will not be charged for two separate crimes, because their is a pragmatical connection between the two acts, be it that stealing and damaging property at the same time may be a separate sort of burglary that is also punished more severely. By contrast, someone who is on his way to a robbery at night using a defective car, commits two, pragmatically unconnected crimes. According to Dutch law, in case of plural concurrence of felonies the maximum-penalty of the most serious crime can be raised by one-third. A case of plural concurrence in which the lacking connection is evident is when driving a defective car takes place a day before the robbery takes place. In many legal systems it is not possible to sanction all cases of plural concurrence separately or if this is allowed, the execution of the sentences may not take place cumulatively.

From the point of view of retribution, non-addition makes no sense. However, plural concurrence as a reason for non-addition is not a universal feature of western legal systems. In the U.S.A., for instance, each individual crime is punished separately, although it may be decided that the penalty will be executed simultaneously. In that case, the American system works out in the same way as the Belgian and French absorption-system in which only the largest penalty is relevant. The latter system is only feasible if the maximum penalty for a single crime is never given. If it were, committing lesser crimes later but before being caught, would be 'gratis'. However, applying maximum penalties in cases of plural concurrence only does not exclude that the price of an extra crime is considerably lower than it would have been had only that crime been committed.

SD does not provide a satisfactory explanation for non-addition. Assuming so, would suppose that the marginal deterrent effect of punishment is increasing, which is highly unlikely. Can GD explain non-addition? I think it can. After all, if a person who has committed, say, five crimes, is punished in the American way, that is cumulatively, it would be as if there is no difference between one person committing five crimes or five persons each committing one crime. Obviously, one's trust in the deterrent effect of the CLS is much more shocked when five persons commit one crime than when one person commits five before he is arrested. To a certain extent he was stimulated to commit more crimes as the criminal system failed to catch him in time. Put in a context of justification in which the CL is not a device to appease the individual victims of each crime, but also has the function to rehabilitate criminals (as will be defended later), GD would seem to explain non-cumulation very well.

E. Recidivism

A higher maximum-penalty in the case of recidivism cannot be explained by retributive considerations. Given the fact that an offender has a certain criminal disposition, there is no reason to punish him more severely other than one's frustration about the ineffectiveness of the earlier sentence. However, that frustration is based upon the SD-function ascribed to the sanction, which is immaterial according to retributivism. Although SD would seem to provide a plausible justification for the raising of maximum-penalties in the case of recidivism, it is not a very convincing explanation. After all, if the maximum penalty was not applied the first time, it would seem that the judge simply has failed to punish sufficiently then, and that the offender has not got the deterrent support he needed. And even if he got the maximum, one might still say that the legal maximum was apparently set wrongly. In fact, if special prevention were the primary function of the CL, there would be no place for maxima. All would have to depend on the strength of the criminal inclinations of an offender. But that again is totally incompatible with the idea of criminal justice because the same crimes might then be punished very differently depending on the judge's opinion of the character of each offender.

Appeasement of the victims can explain the legal doctrine of recidivism to a certain extent. Victims may become more scared when they would know that the offender was not deterred by having been punished before. They might feel safer when he is put away longer. However, from a rational point of view such a fear would depend on a belief in SD. Moreover, appeasement of victims is not a plausible ground in case of victimless crimes. Can GD do better as an explanation? I think so. People convicted of a particular sort of crime will be, on average, more prone to commit such a crime. The reason to believe so is not an unjustified believe in SD. In fact, in case of imprisonment it would be more realistic to include the failure of SD as part of the explanation of why convicted persons are more prone to commit crimes. Potential recidivists represent a class that has proven to need, on average, more deterrence. It would be a form of GD for a special group. Of course, one may question this explanation because one might think of still more punishment after a second act of recidivism, etc. But that is an even more demanding question in terms of SD. From the point of view of GD, a restriction to a one time rise of the maximum-penalty might be defended on the basis of the assumption that extra punishment for two or more time-recidivists would have very small marginal returns in deterrence and would represent a waste in terms of spending the scarce means of the CLS.

The result of this analysis is that all the doctrines analysed are best explained by assuming that GD is the primary function of the CL. The result of our analysis implies that both retribution and SD drop out as possibilities for the logical reconstruction of the CL. However, although GD was certainly the main function the reformers had in mind when introducing the classical model, we still have to consider if GD is logically compatible with the principles of criminal justice. After all, 'absolutists' have always claimed that the relativist view is incompatible with one or more of the principles of criminal justice.

4. The principles of criminal justice

The concept of criminal justice consists of four interrelated principles.

The first principle is that punishment must follow a crime. If an act was not a crime at the time when it was committed, there is no inner relationship between the act and its punishment, as punishment only makes pragmatical sense if it can be known ex ante as a possible consequence of the act (legality principle).

The second principle is that punishment can only be justified if it follows upon a criminal act for which the person to be punished is accountable, and to the extent that he is accountable for that act (the mens rea-requirement or proportionality between guilt and punishment).

The third principle requires equality. Two people who commit the same crime in terms of guilt and seriousness, should be punished with equal severity.

The fourth principle requires proportionality between the seriousness of the crime and the severity of the sanction.

Absolutists argue that relativists have no good reason to cling to the principles of criminal justice, or at least not to all of them. The strength of these absolutist arguments is indicated by the fact that a relativist of the stature of H.L.A. Hart saw no way around them. Hart gave two examples of the incompatibility between criminal justice and the utilitarian demands of deterrence. The first one is that judges will often punish more severely when crime rates rise. Obviously this is, at first sight at least, in conflict with the principle of equality. In fact, it is also in conflict with the proportionality requirement, assuming that the earlier, more lenient sanction was proportional. If judges would be just, they should, one might even argue, punish less severely, as it is understandable that people will tend to commit more crimes as the deterrent force of the CL is obviously declining when crime rates rise. If judges yield to utility in the given example and punish at the expense of justice, Hart's second example is one in which the opposite is the case.

The mens rea-principle demands proportionality between guilt and sanction. That is why we will not punish people who must be held to have been insane at the time when they committed a crime. However, Hart wrote, there may be a utilitarian reason not to accept the insanity defence. People who are not actually insane may speculate that they can get away with their crime by pretending to be insane. Therefore, the possibility of the insanity defence might undermine the deterrent effect of the CL.

Because Hart saw no possibility to reconcile the conflict between utility and justice in the given examples, he proposed a distinction between retribution as the theory of criminal accountability and GD as the utilitarian theory of function, whereby retribution acts as a limit on deterrence. Hart's proposal for a "mixed" theory must be regarded as very unsatisfactory within his general utilitarian approach, as the distinction simply is a reformulation of the problem, not a solution. However, not everybody has accepted this conclusion of Hart's.

Hans Crombag, a prominent Dutch psychologist of law, has suggested that Hart's concession to retribution was unnecessary, because the principles of criminal justice can indeed be explained in a utilitarian way. According to Crombag, criminal justice can be explained in terms of the psychology of learning. Guilt means, psychologically speaking, accountability. Accountability means that the actor normally knows that the act is wrong and may be punished. Proportionality between the seriousness of the crime and the criminal sanction can be explained as a necessary demand for learning how to steer one's behaviour in view of the consequences. If the nature of behaviour can no longer be related to sanctions or the size of sanctions, and the source of the disturbance cannot be identified and be eliminated through aggression, a situation of "learned helplessness" will be created. This can be demonstrated by punishing laboratory rats capriciously. After some time they will retire into a corner of their cage and start squeaking softly because they have become totally disorientated. In other words, according to Crombag, our sense of justice in these matters reflects a cognitive-behavioural demand and taking it into account is necessary to maximise the utility of the criminal sanction.

In my opinion Crombag's answer to Hart can be a part of the right answer, but only a part of it. The psychology of rats is not necessarily the same as of humans. For instance, in the case of humans I would doubt the necessity of a fixed relationship as a prerequisite of successful learning. After all, many offenders only have a vague idea about that relationship when committing their crime and, if hesitating at all, simply gamble that they will not be caught. In fact, a fair amount of uncertainty as to the severity of sanctions, as is usual in practice, might be more deterring on balance, because people in general (in contrast to typical gamblers) prefer to limit their chances of a loss in comparison to an equal chance of a gain. If humans may seem to be less intelligent than rats in so far as they like to gamble, I am still inclined to believe in a somewhat greater complexity and intelligence of humans. If Crombag believes his argument to be a refutation of Hart's argument, he must assume that humans can understand why they are punished differently when crime rates rise. I doubt whether rats are able to relate their own behaviour to the behaviour of their fellows in such a way. I also doubt whether punishing insane people would create problems for the learning process of sane people. It may be disturbing for an insane person to get punished for a crime for which he was not accountable at the moment of acting, although a cynic might argue that an insane person will not think so as long as he is insane. However, after regaining sanity, he might understand the utility of being punished in relation to feigned insanity defences. Therefore, being punished may be very unfair to someone who committed a crime being insane, but that would not exclude an understanding of the rationality of the practice as explained by Hart. If this objection may still be judged to be very speculative, the following is a very fundamental one.

I can agree with Crombag that the insanity defence is -psychologically speaking- based upon an offender's exceptional insensitivity to the threat of punishment. Its cause can be that the offender's had a crazy idea of what he was doing or that he suffers from a structurally absent or highly reduced capacity to get activated by the threat of punishment. It would be understandable that we would not punish people who are totally insane in the sense of the definition, but it would not explain why we punish people less who are not totally insane. On the contrary, it would be rational, it would seem, to punish those people more severely in order to deter them. This would be no more difficult to understand than that judges punish more severely when crime rates surge. Therefore, whereas the psychological theory of learning can explain criminal proportionality to a certain extent, it cannot explain it completely, unless "justice" and learning are related in a more complex, but as yet unexplained way.

5. The Janus-face view of the criminal law

5.1 Narrowness of the traditional concept of criminal justice

In order to find a solution to the problems we are dealing with, we should start by taking the concept of criminal justice seriously. If we are talking "justice" indeed, what kind of justice are we dealing with? Is it "private" or is it "public" justice. Once we ask this question, it may strike us that retributivists would seem to think about criminal justice as private justice. The evil you do to another person, will be done to you in the name of criminal justice. Obviously, relativists aim at a public good to be secured through punishment. Whether that good can be realised with more or less than would be deserved following the retribution-principle, is irrelevant. Obviously also, if a relativist would be in need of a concept of justice at all, he should use a concept of public justice. In other words, he would have to find the basis of his concept of criminal justice in political philosophy and he should look at the place of the CL within the state as a political community.

It is most surprising, to use an understatement, that a political philosophical approach to the problem of the justification of state punishment, has been advocated only relatively recently by Ted Honderich and Nicola Lacey. This is suprising simply because the word "justice" in "criminal justice" would strongly suggest to explore whether the general concept of public justice can explain the principles of criminal justice as an application of the general concept. The simple notion of public justice is the equal treatment principle. In criminal matters this principle would be as follows: each citizen has an equal right to be protected against crime and each citizen should be treated equally when being punished for a crime.

I strongly agree with Honderich and Lacey that a philosophy of the CL which remains detached from political philosophy is a sort of blind man's game. However, I disagree with their Rawlsian and communeautarian background political philosophies, which create new problems for the justification and explanation of criminal justice. Honderich and Lacey operate with too general and too vague concepts of equality and social community. Honderich and Lacey did not get to their conclusions via the embarassingly simple reasoning I have just followed. In fact, their advocacy of political philosophy is based on their conclusion that the CL would be a much too complex phenomenon to reduce it to the simple notion of public justice.

By contrast, that simple-minded direct approach is what will be followed here. The reason why this proposal might seem to be hopeless and why it was never even persued, is that the utilitarian's primary concern of criminal punishment is effectiveness whereas inequality and disproportionality would seem to be required if (equally) effective deterrence would be the aim. Differently from "pur-sang" utilitarians I will not claim that the principles of public justice can be reduced to their effectiveness in furthering some pre-ordained natural good or that justice can be completely reduced to some principle of learning that produces order. Both effectiveness and the possibility of learning are important to realise public goods in an ordered society. However, neither of them is a value in itself that can explain justice. In my view the principles of justice are ultimately based on an irrational preference for peace and rationality. The two problems we will have to consider, then, are the following: 1. can the equal treatment principle be extended to the sphere of criminal justice; 2. is that extension compatible with the four principles of criminal justice as outlined in section 4?

When discussing Hart's case of more severe punishment in reaction to rising crime rates, I pointed out that we might even speak of a double injustice. Not only were people punished more severely than others, they would also be treated unequally because the deterrent effect of punishment would be less when crime rates rise. We can speak of a double injustice on the basis of an intuition that criminal justice has two sides. It would be wrong to look at punishment as just a process in which the criminal is used to deter other members of society. Apart from the fact that he himself may profit from being punished by being deterred from committing crimes again, an at least as important aspect of being punished is rehabilitation.

Criminal justice, seen in this way, has two faces, like Janus, the double-faced Roman god of doors and porches, who looks into the past and the future at the same time. The past-orientated face of the CL is concerned with the reinforcement of the GD-effect of the threat of punishment that has been weakened by the offense. On condition that this tribute to the past is paid, Janus will open the door to the future in which the offender can reappear as a rehabilitated citizen. This "Janus-face view" of the CL would seem to understand criminal punishment of the offender as the price, in the form of a functional retribution, to be paid for his rehabilitation as a citizen.

In fact, a basis for this view can be found in the very work of H.L.A. Hart himself. In his theory of the Minimum Content of Natural Law, in the ninth chapter of The Concept of Law, Hart presented a number of truisms about man, the world and society. One of them is man's limited understanding and strength of will. It is this limitation, according to Hart, that makes the CL a necessity for the survival of society. However, the fact that the truism, by its nature, applies to practically every man, implies that the CL is not only a necessity of collective survival. It is also a necessity for the potential offender that practically everybody is, as it scares off crime and also provides a way to get reconciled with society in the form of criminal punishment when its deterring effect fails.

The idea that the CL serves everybody's interest in both its preventive and its rehabilitative functions is in conformity with modern criminology. It has taught us that most crimes are not so much a matter of psychological character, but predominantly a matter of variable social relations, chances and environment. Under those circumstances permanent exclusion of offenders from society is simply too expensive and too inefficient as a general solution for the problem of crime. However, the Janus-face view is not exclusively modern. It can claim ancient roots in Stoicism, as in Seneca's De Clementia, in which he argued for human solidarity and a reconciliatory attitude in the face of the universal human weakness of character.

The Janus-face view is a functional theory of CL within the broader framework of a political theory. It explains the CL as a device to cope with the problem of the general human proneness to crime in an efficient way, following the general principle of public justice. Each citizen should equally benefit from the CL (by being protected by the state against crime or by being rehabilitated after having been punished) and each citizen should be equally charged for the service of the CL (as a tax payer or as an offender). The four principles of criminal justice that we have discussed so far, only regard the principle of equal charges for offenders. The Janus-face view of CL has a broader concept of criminal justice that will now be shown to enable it to refute the thesis that relativism is incompatible with the traditional criteria of criminal justice.

5.2 A refutation of the incompatibility thesis

The Janus-face view is compatible with the legality-principle. Its argument for the legality principle is simple. It is unjust that citizens are charged for things they could not know in advance they could be charged for. Moreover, as a functional theory it accepts the psychological unwisdom of "punishing" acts that could not be known to be illegal and punishable. Nevertheless, the Janus-face view cannot accept all psychological insights concerning the greater efficiency of punishment. For example, I had argued against Crombag that a certain degree of uncertainty concerning the severity of punishment would make its deterrent effect greater rather than smaller. However, this would be in conflict with the equality principle. This general equality principle demands that the charges and profits of public services are distributed equally. As a public service the CL is subject to this principle.

The Janus-face view can also deal with Hart's problem in connection with the guilt-principle and the insanity-defence. As insane people as defined above are not, or much less, able to profit from the deterrent effect of the CL, the price that can be expected to be paid by them must also be proportionally less.

It may seem as if there is an inconsistency between, on the one hand, punishing insane people less or not at all, because they can profit less from the deterrent effect of the criminal sanction, and, on the other hand, punishing recidivist more for the very same reason. However, an insane person does not know himself to be insane at the moment of acting and would not realise that a more severe punishment can be expected, or, if he does, he may be structurally insensitive to that fact. However, a partially insane person can know quite well that he has been punished for the same act before and should expect a more severe sentence. A partially insane recidivist may also be punished proportionally more severely therefore. However, is punishing recidivists more severely compatible with the equality and the proportionality principles

We have already argued above that punishing recidivists is certainly incompatible with a retributivist view of the CL. Our justification from a GD-functional perspective was that people who have already committed a certain crime will be, on average, more prone to commit such a crime. By raising the maximum-penalty in case of recidivism this class of offenders may also profit more from the sanction preventively. The scales of CL-justice remain in balance, therefore, as far as the equality principle is concerned. However, what about the proportionality principle? Is the damage caused by a recidivist crime also greater than the same crime committed by a first offender? Not as far as the damage to the rights of others is concerned. However, there obviously is a difference in so far as the GD-effect that has been damaged and has to be repaired is a greater one. In that respect, a recidivist crime is more serious indeed. In fact, in principle the size of the increase in the severity of punishment should be such that the crime rates of potential recidivists remain the same as with potential first offenders. If there is no difference between the chances of committing a crime between the two classes of potential offenders, it is wrong to punish recidivists more severely.

The more general problem of the compatibility between GD and the principle of equality is not so easy to solve. Can the Janus-face view justify the fact that judges will start punishing more severely when crime rates rise? The reason that a judge would feel the need to punish more severely in that situation would be that he believes that punishment no longer deters as much as it did before and that extra deterrence may compensate this. Is punishing more severely in this situation doing the same as what a judge does when he punishes recidivists more severely? In fact, it is not. Our justification for punishing recidivists more severely was, as far as the equality principle was concerned, that recidivists would also profit more from a greater threat of punishment preventively. Provided then that, in order to respect the legality principle, the rise in the severity of punishment would be announced beforehand, the problem of the equality principle could be solved. However, a greater severity of the sanction would still be in conflict with the proportionality requirement. Someone who commits a crime in a period when crime rates rise, does not necessarily cause more marginal damage to the GD-effect of the CL than someone who committed the same crime before. The difference with the case of recidivism is that we had stipulated that the rise in the severity of the sanction would depend on the existence of a greater marginal damage to the GD-effect.

Now that the comparison with recidivism has been shown to fail, another, more promising comparison can be made. Punishing the same crime differently depending on the GD-effect that can vary at different moments in time, would seem to touch upon the same problem as that different crimes of equal seriousness will often not be deterred as well if both would be punished with equal severity. This is nothing but the fundamental problem of criminal justice that retributivists have confronted relativists with. Are we not forced to the conclusion that we must give up the proportionality requirement if we would like to have effective deterrence? No, in the framework of the Janus-face theory we are not.

The Janus-face view demands, on the one hand, an effort of the state to equally protect its citizens against crime, and, on the other hand, a just and equal treatment of the citizens which break a CL. For both parts of this task the principle of equal charges and equal profits is valid. Equal protection is a function of the seriousness of crimes and the chance of becoming a victim of crimes. On the other side, the side of the citizens that commit crimes, a right can be claimed to be punished in accordance with the seriousness of the crime that has been committed. Sticking to the proportionality principle implies that punishment can have a different GD-repairing effect with different crimes that differ in kind but are of equal seriousness. However, this difference corresponds to a difference in the marginal GD-damaging effect of a crime. Therefore, damage and repair remain in balance. However, assuming equal chances of becoming a victim of both sorts of crime, citizens would not be equally protected against those two sorts of crime. The implication is that if one wants nevertheless to create a balance in the level of protection, one should look for other means than the threat of punishment to protect citizens from crime. Obvious means for such compensation are police activity to arrest more offenders, subsidising programmes and technical means of crime prevention, reforming criminals were the cost-efficiency of doing so are high, and, finally, compensating victims for costs not covered by normal insurances against crime.

Does our solution for the problem of the proportionality requirement not also provide an answer to Hart's problem concerning the equality principle? Yes, be it a negative answer. Punishing more severely, assuming that the present level of the sanction is proportional, is wrong as a reaction to a rise in a particular crime rate. However, it would seem that Hart implicitly assumed that a judge has no alternative but punishing more severely. Even if this were true, the judge's only alternative is not the CLS's only alternative. The consequence is that we have to reject the practice that was used by Hart as an argument for his incompatibility thesis. This is not because it is in conflict with the equality principle, but because it violates the proportionality principle.

Our conclusion notwithstanding, there is one cause of rising crime rates that can be regarded as making those crimes themselves more serious. In that case, punishing more severely can be compatible with the Janus-face view.

5.3 The paradox of the seriousness of crime

The cause of raising crime rates is not only that for whatever reasons a crime has become more attractive. It can also be that at least a part of the increase is due to a public depreciation of the seriousness of a crime. Thus a paradox is created because a lack of public appreciation of the seriousness of a crime may make it more serious in reality.

An excellent example of such a paradoxical process of depreciation is what I like to call the "black bicycle plan" in the city of Amsterdam. The Provo's, a movement of young anarchists in the mid-sixties, proposed a "white bicycle plan" to increase the efficiency of the use of bicycles and to curb the desire to steal bicycles. The idea was that the community would provide white painted, simple but strong bicycles, in such quantities that one could practically always find one standing nearby when necessary. Instead of this utopian, white plan, a black "plan" was realised without any government interference, or better, thanks to an almost total lack of such interference. Stealing bicycles became so frequent and many people became so annoyed after their fourth or fifth bicycle had been stolen in a relatively short period, that they felt almost "justified" to steal one back themselves. This sense of justification was reinforced by the police who had almost given up doing anything against bicycle thefts. If one was lucky they were willing to register a theft for insurance purposes. The consequence was that the average quality of bicycles decreased because people were too afraid that a good bicycle would be stolen, the more so as it became very difficult to get a bicycle insured. Obviously, the morality of and respect for private property became seriously eroded.

One can imagine that a judge would like to do something about this discrepancy between the real and the perceived seriousness of stealing bicycles. A judge might announce his decision to punish more severely than he used to do formerly in order to signal that the erosion of values can no longer be tolerated. In doing so the judge would hope to create a shock in the public consciousness created by a belief that bicycle-theft will henceforward be taken seriously again. Would such a policy of criminal punishment be reprehensible in view of the equality principle and/or the proportionality principle?

The reason that such a policy can be seen as compatible with the Janus-face view is that when crime rates increase significantly, a crime is more serious indeed when it is caused by a misperception of the seriousness of a crime. In fact, we hit upon a factor of crime causation that can justify a differentiation in the severity of sanctions even in individual cases, that is without a rise in crime-rates. A crime that not only expresses the weakness of character of the offender but also a serious want of appreciation of the interests that are protected by the CL, is more likely to cause damage to its GD-effect as it sets an example of committing crime for frivolous reasons. Therefore, it is also not true that an interest for the moral character of an offender and margins of judicial discretion to determine the severity of sanctions are out of line with a GD-functional approach of the CL.

However, in the case of rising crime rates it may be very difficult to know to what extent surging crime rates are due to a change in the attractiveness of a crime or to a changed perception of the seriousness of a crime. A change in attractiveness may even be the most important cause of such a change in perception. Moreover, the state itself may be responsible for the rise in crime rates when it has been to slow in counteracting increased attractiveness of a crime through other means than increasing the severity of sanctions. Practically speaking then, the possibility of success of the questionable practice of punishing more severely to cause a reversal in the public appreciation of the seriousness of a crime, may crucially depend on increasing arrest rates at the same time. Moreover, if the policy is not successful in correcting a general want of appreciation of the seriousness of a crime and in reducing crime rates again, it can no longer be justified, whereas, both when it has caused a reversal and when it has not, it is very difficult to lower the severity of punishment again once it has been raised. The reason is that judges usually fear that punishing less severely is interpreted as a message that a crime should no longer be taken so seriously. This explains why it has often been observed that severity of punishment is sticky downwards. Therefore, the policy of punishing more severely should only be used under circumstances nearing an emergency situation. It should certainly not simply be accepted as uncritically as Hart did. However, with all these proviso's, it can be defended within a political philosophical context in which public authorities are not only responsible for the protection against crime, but also for what Durkheim would have called the "conscience collective" concerning the values protected by the CL. It would be foolish to deny them that responsibility because, as David Garland wrote: "if punishments (and the threat of punishments) were routinely imposed but were entirely ineffective in containing crime, then this could undermine the claims of the moral order and the political powers which pose as its guarantor".

5.4 The Janus-face view as a critical theory

The Janus-face view can claim to integrate retribution and prevention within a coherent relativist theory of justification. Moreover, that it can is due to the fact that the theory itself has been integrated into the more general political theory of the principles of public distributive justice. These two theoretical values are not all it can claim. It can also be fruitful as a modern, criminologically orientated critical theory. It raises important questions like to what extent the principle of equal protection against crime is actually realised. It offers a new framework to make judgements about the use of discretionary powers of judges, public prosecutors and the police. The theory also integrates the historically speaking recent interest in victimology. In so far as particular crimes cannot be prevented as well as others or particular people cannot be protected against crime as well as others, there is a justification for extra expenditures for the support of victims of crime. Moreover, the issue of class justice is transformed because the theory makes class justice against the victims of crime as important an issue as class justice against offenders.

However, the Janus-face theory also has something to offer to offenders, because rehabilitation is the other side of the Janus-face. Therefore, structural secondary effects of punishment should be taken into account when the severity of sanctions are considered and especially so if these secondary factors are also criminogenic. The compensation of criminogenic factors, including stigmatisation, may justify special benefits, like job training, temporarily guaranteed housing and employment after punishment, etc., provided, of course, that such facilities will not have an important anti-deterrent effect themselves.

6. Discipline and punish: a pre-classical inheritance

6.1 The unexplained nature of rehabilitation in relativism

Is it the complexity of the Janus-face theory of the CL that explains why it never found expression in the period when the classical model established itself? After all, the Janus-face view is very much in line with the humanitarian, reconciliatory spirit of Beccaria's foundational work on the one hand, whereas its GD-functional approach is close to the technical-systematic spirit which pervades the work of both Feuerbach and Bentham. The reason, I think, why the two aspects never became integrated in a satisfying way, is not the complexity of the matter. The reason is, I will argue, of a philosophical kind. Beccaria and Feuerbach developed their ideas within the framework of social contract thinking, whereas Bentham, by contrast, operated in the almost anti-philosophical spirit of naive empiricism, an empiricism that was pretty speculative in reality.

We have seen that retributivist theories were unable to cope with the differentiation between morality and law that was actually introduced with the establishment of the classical model of the CL. I will now argue that the early relativists were equally incapable to cope with that differentiation. They never differentiated consistently between GD and SD. The latter, I will argue, played a hidden role in their theories. In fact, it is SD, or, to be more precise, the assumption of a reforming effect of criminal punishment upon the offender, which explains why they were not much concerned to explain why criminals should have the right to become full citizens again. The reason why they could not include the retributive element into their theories, as the Janus-face theory does, is that they did not have GD on both sides of the account of guilt and punishment, but GD and SD. If the damage that is to be repaired is GD-damage, its reinforcement through punishment can settle the accounts. However, SD cannot do that unless the SD-effect of punishment would always be greater or equal to the GD-reinforcing effect. However, that is not very plausible.

6.2 The absence of imprisonment as a sanction with Beccaria

One can see the influence of metaphysics with Beccaria already, although he was much too confused to allow it to be a more systematic influence. The only natural form of punishment, following Beccaria, is being excluded from the social contract, at least temporarily: "Anyone who disturbs the public peace, who does not obey the laws, that is, the conditions under which men agree to support and defend one another, must be excluded from society - he must be banished from it." It points to the novelty of imprisonment as a criminal sanction in the eighteenth century, that it occurred to Beccaria only as a means to detain arrested people in the phase of the criminal investigation. Nevertheless, if banishment were the natural sanction but, as was often the case in Beccaria's time, was no longer possible in practice or seen as disproportionate as a general sanction on crimes, it would have been natural to see imprisonment as a form of internal banishment.

It is puzzling why Beccaria did not come to this conclusion. This is all the more so, as the other sanctions he considers would seem to be highly dubious in the light of his own views on the criminal sanction. Notwithstanding the great success of his book, they were pretty incoherent. In the case of theft, Beccaria considered fines to be the most appropriate sanction. It shows his incoherence that the argument he gave for this, follows the logic of retribution: "Whoever seeks to enrich himself at the expense of others should be deprived of his own". However, he then continued, theft is usually committed by people who are very poor and taking the little they possess would make them only more criminal and would also mean punishing their innocent dependents. Therefore, forced labour "through total personal dependence" will have to replace fines. However, it is difficult to imagine that this would be possible without imprisonment.

Notwithstanding his principle that the least possible torment on the body should be inflicted and notwithstanding his perceptive observations concerning the negative effects of "brutalisation" in the case of the death penalty, Beccaria stated without any further argument that the penalties for crimes against the person "should always be corporal punishments". The fact that he did not dwell to tell the reader what these corporal punishments should consist of, indicates that he had a sense of his own inconsistency. Again, however, the logic of his argument, had he been more coherent, would have driven him to imprisonment, because imprisonment is a corporal punishment, but is not cruel in a directly physical sense. The reason why Beccaria may have shrunk from making imprisonment central in his theory of the criminal sanction, is that he clearly associated it with the hell holes that existed in his time, albeit as places of criminal investigation. The fact, as he pointed out, that people were punished there before their guilt had been established, must have made him fearful that imprisonment and its secondary consequences could never become a proportionate form of punishment. A view, it must be conceded, which is very realistic even in some modern countries were imprisonment must be feared less for the loss of one's liberty, but rather for the risk of being beaten, raped or murdered by inmates or guardians.

When reading arguments like these one may suspect that philosophical motives determined Beccaria's speculations, but not always explicitly. After all, banishment may hurt dependents much more than a fine, whereas a state might not only have a financial interest in fines (another counter-argument Beccaria brought up against fines), but also have an interest to commit people to forced labour. One wonders then whether Beccaria's unexpressed reason to be against fines was not that it is strange to divest someone of that which he had lawfully acquired already. Forced labour "in total personal dependency" is primarily loss of freedom and an unfree person, a slave, is owned by someone else and does not produce anything that belongs to himself. An explicit philosophical argument can be found in Beccaria's argumentation against the death penalty. Nobody, he wrote, would agree to give the state the right to execute him as part of the social contract.

6.3 The prison as social therapy with Feuerbach and Bentham

With Feuerbach we also find that he thinks of imprisonment as the "foundation of every well-ordered criminal law". Feuerbach did not reject the death penalty, but thought it to be an aequate sanction only in case of offenses against the existence of the state or in cases of murder that he thought would disqualify an offender ever to become a citizen again. Fines, he reserved for crimes committed purely for material gain, although they would have to be replaced by forced labour when an offender had no means from which to pay the fine. They are, subjectively speaking, not crimes to hurt another person, but they have that effect. In most other cases in which a personal-emotional motive was behind the crime, imprisonment would in his opinion be the right sanction, because the essence of the threat of imprisonment is that of dehumanisation, of being without the others whom the offender has shown to be inclined to hurt. The prison is supposed then to mobilise through social isolation feelings of sympathy for others to compensate the feelings of hate which have driven the offender to commit his crime.

There was clearly a link between the central place of imprisonment in Feuerbach's theory of the criminal sanction and his background in social contract thinking. In his "Anti-Hobbes" Feuerbach had argued that the aim of the state is precisely to protect the freedom of its citizens which is in permanent danger in the state of nature, because everybody sees in everybody else a potential aggressor. Therefore, he argued, if the state itself would not be bound to the law, the situation might even get worse. The prison is a situation in which Leviathan does not have the right to kill, but has total control over its citizens otherwise. It is therefore the Hobbesian half-way station between the war of everyone against every one and the rule of law. The prison teaches the citizens the value and love of freedom.

With Bentham, finally, the prison in the form of his Panopticon stands out as a means of disciplining. Disciplining in Bentham's view means psychological conditioning of people to the advantages of a regular life of labour and obedience to rules. However, before a prisoner is ready for reform he must be socially and even sensorially isolated. Although Bentham rejected natural law and social contract thinking, he clearly believed that an artificial experience in prison with Hobbes' "solitary, nasty and brutish" life in the state of nature would be most helpful to let an offender see the blessings of a civilised, law obedient sort of life. Thus Bentham thought very much along the same lines as Feuerbach and this is also true in relation to his ideas about the death penalty.

6.4 The metaphysical relation of punishment and discipline

The emphasis on imprisonment in the classical model is something which has lasted until this century. It is remarkable, for instance, that the enforcement of the payment of fines was only introduced in this century in several countries, whereas imprisonment remained as an alternative if a fine was not paid. The social contract and state of nature background of this focus upon imprisonment was also apparent in the fact that, until long into this century, prisoners used to be deprived of their political rights for at least their term of imprisonment. The way the founding fathers of the classical model dealt with the death penalty and their background ideas in relation to imprisonment, shows us that they were still thinking along metaphysical lines, however modern and empirically orientated they pretended to be.

The central place of imprisonment in the classical view of the CL against the background of its earlier uncommonness, is something which Michel Foucault has interpreted as "based first of all on the simple form of 'deprivation of liberty'". It is, Foucault wrote, "a good that belongs to all in the same way....Its loss has therefore the same value for all; unlike the fine, it is an 'egalitarian' punishment" and, very interestingly for our analysis of the relationship between crime and time, Foucault remarked that "there is an economico-moral self-evidence of a penalty that metes out punishments in days, months and years and draws up quantitative equivalences between offences and durations". Foucault clearly saw the metaphysical factor behind the central role of imprisonment in the classical model. Surprisingly, however, he did not see the metaphysical connection between the calculus of freedom and disciplining in the classical CL. We can understand this, however, now that we have seen why the founding fathers of the classical model thought that punishment would make an offender suitable to enter society again. The idea that the criminal sanction reforms the offender was its implicit premise. Therefore the combination of disciplining and punishment in the classical model was not just a matter of the opportunity the prison offered for disciplining, as Foucault has suggested.

Although the fathers of the classical model manifestly focussed on GD as the primary function of the CL, reform of the offender remained part of their construction of the CL as an explanation of why the offender could return as a full citizen after having been punished. The real 'paradox' is, then, that the founding fathers did not think in a radically temporalised-functional way, but that they kept thinking of criminal punishment as a means to reform the very nature of man. In that approach one recognises the state of nature and/or social contract-theoretical background as well as the secularised motives of fall and redemption which pervaded the Enlightenment in both its deistic and its romantic forms.

Put into that metaphysical and crypto-religious perspective, the focus on imprisonment, on "doing time", can also be understood in terms of our analysis of modernisation as radical temporalisation. Prison is the secular, temporalised form of God's hell, the anti-world of heaven, where instead of eternal happiness there is eternal pain and suffering. A criminal is basically seen as someone who cannot cope with time in life. In prison, he will, on the one hand, lose all sense of time in relation to the real world. However, it will also teach him the sense of time, because the temporal routines of the prison are the only thing which mark the difference between reality and the emptiness created by total isolation and sensorial deprivation. In the imagination of the classics it is in the isolation of his cell, and the "operationally closed system" that is the prison, that the offender must try to find his way back to his better self, his consciousness as a human being who can cope with a temporalised social order and the discipline that it demands from him. Whereas the Enlightenment isolated the temporal world from the eternal world, it also created its functional equivalents within the temporal world. Whereas prison represents hell, society in its ever increasing progression represents heaven.

6.5 The classical model and the death penalty

The problematic nature which the death penalty had in the classical model can also be better understood against this background. The death penalty marks the dividing line between the reformable offender and the "deadly" sinner. However, the death penalty cannot be understood very well any longer in a temporalised world precisely because it puts the offender out of time. As punishment has become something which happens in time, the death penalty can no longer be understood as punishment. Of course, the prospect of that punishment can cause tremendous fear, but that is not what the death penalty is about, because it would make its actual execution unnecessary. However, with its actual execution, there is no offender left to suffer, because in a secularised world, where heaven and hell after life no longer figure, it is assumed that death puts an end to all suffering. The death penalty therefore becomes a criminal paradox in the classical model. However, since the classical model also operated with the state of nature/social contract (civilisation) difference, it could go on thinking about the death penalty as penalty, even though it makes no sense in terms of punishment and reform. Of course, the discussion about the death penalty is not over once the metaphysical scales are really dropped, although it would be better, following the Janus-face view, to simply call it incapacitation through execution, because then it no longer has anything to do with punishment indeed.

Another reason why the death penalty became so problematic in the classical model is that it degrades man to the status of an animal. This is because it submits him to the most bestial, irresistable instinct there is, the urge of survival. Beccaria argued very perceptively that the death penalty, like corporal punishment, is a form of brutalisation that risks provoking corresponding responses and behaviour in society. Keeping those sanctions brought the classical fathers dangerously close to the criminal legal practices of their feudal opponents.

7. The absolutist justification of the criminal law and its hidden virtues

7.1 Purified absolutism

After having argued that the relativist model is fully compatible with the classical model of the criminal law, both reconstructively as well as philosophically, we have still not dealt conclusively with the absolutist model. I have argued that absolutism is incompatible with a number of important features of the classical model. Furthermore, the Janus-face view is a refutation of the absolutists' claim that the principles of criminal justice are incompatible with a relativist view. However, an absolutist who concedes that our criticisms of absolutism were justified, might then make a dramatic move. He might argue that the existing criminal law, in so far as it is incompatible with absolutism, is philosophically unjustifed and should be reformed. Thus, impossible crimes, attempts and recidivist's crimes would have to be punished as severely as ordinary crimes. Statutes of limitations must be abolished and sanctions should accumulate in case of plural concurrence. Finally, in order to respect the equality principle strictly, rising crime rates should no longer be met by more severe punishment, but rather, following our principle of equal protection against crimes, by putting more money in other means which can reduce crime rates.

Against such a purified absolutism I would now like to argue that it is also incoherent for philosophical reasons. The main reason to have a look at (early) absolutism, however, are not its incoherence, but certain qualities that were better represented in it than in relativism, notably, the integration of the theory of the criminal within political philosophy. Together with a hidden metaphysics of classical relativism (next section) this may explain why theories like the Janus-face view were not developed even though this view is quite compatible with the social contract view which people like Beccaria and Feuerbach and so many other relativists (but not Bentham) adhered to.

My analysis will be almost completely restricted to a critical review of the theories of the founding fathers of modern absolutism, Kant and Hegel. With one exception, I will not review the many later variants of absolutism, because I believe they all suffer from the same fundamental defect as the theories of the founding fathers. They all believe that the justification of criminal legal punishment is to be found in morality. However, doing so it becomes impossible to explain why not all immoral or at least all illegal behaviour should be punished by the law. This objection can also be made against the one exception on my restriction to the founding fathers of absolutism, the theory of Leo Polak (1880-1941). However, he improved upon the theories of Kant and Hegel, because he took the problem of the operationalisation of retribution seriously. This might not have been a sufficient reason yet to trouble the reader with his rather complicated theory. However, what makes his operationalisation particularly interesting is that it has structural similarities with both Rawls' theory of justice and the Janus-face view. As Rawls theory of criminal justice resembles the janus face view again, it will also give us an occasion to explain the differential foundations and consequences of Rawls' concept of (criminal) justice and the Janus-face view.

7.2 The political nature of early modern absolutism

The 'quia-peccatum' as a sufficient condition in the absolutist view of the criminal sanction does not imply that an absolutist would have to think that the social functions of criminal punishment are unimportant. Even Kant, who stated that if society would come to an end tomorrow, we should nevertheless still hang the last murderers today, defended his theory within the context of a particular political philosophy. It was essential, according to Kant, that criminal punishment would be a public punishment, an expression of collective reason in a criminal procedure that should demonstrate why a particular offender deserves a particular punishment. In his view criminal punishment should be a public responsibility as a confirmation of what, one hundred year later, Emile Durkheim, a (sociological) Kantian, was to call the 'conscience collective'. In fact, in the Kantian view the state has to take care of this responsibility in order to maintain the social contract that justifies the state's existence. If the state would fail to do so, the victims of crime might claim back their natural right to self-protection. However, at this point the incoherence of the Kantian position appears.

According to Kant, the use of force for the protection of each individual's freedom is what differentiates law from morality. The use of force is justified on the basis of the 'Juridical Imperative' (JI). Following the JI one should act in such a way that each person's freedom is compatible with the freedom of all other persons. The JI is, according to Kant, a logical implication of the 'Categorical Imperative' (CI). According to the CI a moral being is obliged to act according to a maxim that is valid for any moral being. In other words, a morally good act is one in which no distinction is being made between self-interest and the interests of others. It does not imply that one may not act in one's own interest if that is hurtful to the direct interests of another person. Competitive behaviour in the market, for instance, may hurt the direct interest of another person, whose enterprise may go bankrupt. However, such competitive behaviour is in conformity with the CI because it is beneficial for people in general and is expected, in principle, from every competitor in the market, including the unlucky ones who go bankrupt but have voluntarily taken that chance.

The JI is a necessary condition of the CI, because one cannot follow the CI if one is not one's own master and if one can not dispose of one's means. However, the JI is just a necessary condition. According to Kant, immoral behaviour that does not constitute a violation of someone's possibility to act as an autonomous being, cannot be an object of legal (criminal) regulation. Punishment for immoral acts which are not illegal can only consist in moral censorship. However, moral censorship can take the form of civil legal acts, e.g. disinheriting someone.

7.3 Criminal functionality and political morality

Because Kant thought in terms of a social contract, one can approach the question of the justification of the criminal sanction by asking whether the retributive principle is also valid for punishment in the state of nature. In the state of nature, punishment at one's own initiative would seem to be justified only if its aim is to protect one's own rights or the rights of one's dependents. But this raises three questions concerning the relationship between the CI and the JI. What if effective protection of one's rights demands sanctions that are more severe than those which can be defended on grounds of retribution? What if legal measures to morally hurt the offender have already consumed all or much of the allowable space for punishment? Finally, what if criminal sanctions are sufficiently effective which are less severe than the punishments that might seem to be deserved from a moral point of view?

It would seem that the identity between moral beings that is assumed by the CI is incompatible with the functional rationale of the JI if the effective deterrence of offenders would require punishment above the maximum amount of suffering that an offender deserves from a moral point of view. As to the third question, there is, in the state of nature, no duty to punish criminally with less severity than would be allowed from a moral point of view, simply because there is, also according to Kant, only a right to moral punishment and not a duty. The question to be raised then is why this would be different after the social contract has been concluded according to which private persons transfer their powers to punish the state. That the state should always act according to the principle of talio is clearly what is implied in Kant's famous statement that murderers should still be hanged at the eve of a society's existence.

The answer to be given from the Kantian point of view is, I think, the following. The state is based on the social contract which is a contract between autonomous beings. Therefore, the state must assume that a crime committed by a citizen is a principled form of behaviour also. In fact, Kant assumed that an offender is setting a norm indeed. In Kant's view an offender is not simply driven by his immoral desires, but he is also supposed to deny the validity of the norm he violates. Kant used the assumption of the norm-setting offender to deduce the talio-principle as expressing the respect of the state for the criminal as an autonomous being. Thus, as e.g. a thief is supposed to set a norm that stealing is allowed, the state will respect him as an autonomous being by following that very same rule when punishing him, that is "steal" from the thief. However, this view leads to an absurd consequence, because the state, by doing so, could no longer argue that stealing is wrong. This absurdity was taken away in Hegel's dialectical interpretation of criminal retribution.

7.4 Hegel: the dialectics of crime and punishment

Hegel gave a somewhat different interpretation of the relationship between autonomy and criminal retribution. According to Hegel the relationship between the state and the criminal is of a dialectical nature. The state does not take the criminal seriously to the extent that it acknowledges the validity of the norm set by the criminal. It just treats him as an autonomous being as far as the form of the sanction is concerned. According to Hegel, it is just an ironical pedagogical device, a doing as if the thief had acted as an autonomous being. By "stealing" from the thief it is brought home to him what autonomy really means. It demonstrates that autonomy and living at the expense of others are incompatible principles. That is the subjective side of criminal punishment.

The objective side of criminal punishment, the meaning and value of punishment for society, is to express that justice prevails. The effect of punishment is that crime is shown not to pay. Although punishment will deter potential offenders, such deterrence is not pure intimidation. The message is not simply that the law has teeth, but that the law has teeth as the embodiment of justice. Doing (criminal) justice, Hegel wrote, means that the state treats its citizens as equals, as identities as far as the law is concerned, or, as Kant expressed it, that "if you steal from another citizen, you steal from yourself".

With both Kant and Hegel criminal punishment is a matter of public distributive justice. In relation to the offender the state does not act on the same plane as a punishing person does in state of nature. It acts in a vertical relationship, following the principle that it will treat each citizen as an autonomous being, that is, according to the rule this person is supposed to have set himself. Thus, it will respect the property of citizens who respect property and it will steal from citizens as much as they think can be stolen from their fellow citizens. In this way the state fulfills its primary role as a keeper of the peace. The thief can not claim that any wrong has been done to him, whereas his victim, assuming that his civil damage is also restored, cannot complain either, because the liberty the thief took to violate the law has been paid back by an equivalent criminal sanction.

Thus the balance of joys and sacrifices which the law imposes has been maintained. In Kant's and Hegel's theory of the criminal law the state appears as a moral bookkeeper under the aegis of the equality principle that it has to follow in dealing with its citizens. Everybody, including the offender, will always be treated as an autonomous being and nobody will be able to take advantage of his fellow citizens provided that the law operates effectively.

The given interpretation of Kant's and Hegel's retribution theories is, I think, the best possible in the sense that much of its alleged absurdity vanishes when its political nature is made explicit. With Kant and Hegel, retribution has nothing to do with talio in the primitve sense of "an eye for an eye, a tooth for a tooth", but with political equality, as they do not think the principle to be valid in the state of nature, where punishment may be less severe than the talio-principle would demand. The fact that Kant and Hegel thought of criminal punishment as a matter of public distributive justice is noteworthy for two reasons. It shows that modern absolutists were, originally at least, more inclined than relativists to think about criminal punishment within the framework of a political theory of justice. Certainly, the social contract idea also figured prominently in both Beccaria's and Feuerbach's theories of criminal punishment, but it had no role to play for the determination of the size and kind of punishment. Once the right of the state to punish has been established via the social contract, the determination of the size and kind of punishment in their theories is purely a matter of the costs and benefits. This was a fortiori the case with Bentham who rejected the social contract idea. The political nature of criminal punishment is irrelevant for his theory. If an effective power to punish were in the hand of a private person, he would have to follow the prescripts of Bentham's felicific calculus as much as a state.

7.5 The functional limits of retributive criminal justice

We have seen that retribution is a matter of public distributive justice from the point of view of the state. However, the citizens may be happy that the state respects its citizens as autonomous beings, but why can they rest assured that a retributive criminal law is also sufficient as a means to protect the legal order? Remember that, in Kant's development of the matter, they transferred their natural right to punish to the state in order to have an objective, unpartial sort of criminal law. However, that, in itself, is quite compatible with a non-retributive sort of criminal law. So why would they accept a retributive criminal law when its possible inefficiency might endanger the legal order?

It is an almost always implicit assumption of retributivist criminal legal philosophers that a retributive criminal law is effective enough to keep crime under control. The latter concept is vague, of course, but it can be understood in terms of the capacity to arrest probable suspects and put them to trial, such that self-help will be a rare phenomenon. If a criminal legal system dysfunctions severely, society is regressing to the state of nature in the sense that self-help is prompted as the way of dealing with crime.

The view of Kant and Hegel must have been that seeing justice being done is an effective means of crime prevention, as it demonstrates that, as Hegel wrote, "crime does not pay". However, what does that mean in view of the fact that this demonstration cannot be based on a GD-effect? I would like to suggest that they assumed that the primary social function of the criminal law is self-confirmation of the legal order. Such an interpretation is suggested by Kant's and Hegel's, in itself pretty absurd idea, that an offender would set a norm opposed to the norm of the legal order. However, from the point of view of the citizens, crime may indeed be experienced as a challenge to the validity of the norms of the legal order. Self-confirmation means two things. Firstly, that the norm which the offender has set has not prevailed and, secondly, that the offender has been told a lesson in a way that is in conformity with the rules of the legal order itself. Not only has the legal order shown its teeth, but it has done so in a way that justice has been done.

Given this interpretation, however, the question concerning the functional limits of the criminal legal system is acute. In this interpretation a legal system will go down the drain if it can no longer deal properly when suspects, although known, are not arrested, or if, when arrested, they do not get the punishment they deserve although they are evidently guilty. But when does that happen, the reader may ask, unless in revolutionary times? In fact, it happens regularly during times when crime rates rise and the criminal system is simply lacking in capacities. For instance, in the Netherlands the trust in the criminal legal system has suffered considerably during the last ten years because known offenders were not arrested or came off the hook very easily because the capacities of the prison system were too limited. Rapists and violent robbers released for that reason were spotted by their victims shorlty after they had been arrested. In Italy, sentences below two (!) years are rarely executed, because the system simply can not handle them any longer. Obviously, such situations will undermine the trust in the criminal legal system enormously.

The retributive principle, apart from other problems of its application which will be discussed below, is not likely to endanger society because the severity of punishments it allows would not suffice. On the contrary, precisely because the severity it demands may cause great inefficiencies in the criminal legal system it may create such a risk. If, for instance, the Dutch judiciary had not started to punish more severely as a reaction to rising crime rates, the problem of a lack of prison-cells might have been much less dramatic and not so many criminals would have been left unpunished.

The argumentation just developed does not imply a rejection of the ideal of retributive punishment as such. All it leads to is that that a state and its citizens that would like to follow the retributive principle may be faced with a dilemma between effective crime control and the purity of the retributive principle. Punishment can be deserved, without being functional, and it can be functional without being (fully) deserved, that is, less than deserved, or more than deserved. Understandably, this is very awkward. It would be nice if such a deviation from the retributive ideal could be justified.

7.6 Two Hegelian theories of criminal punishment?

Hegel's theory of criminal punishment is usually interpreted in the way we presented it before: as an improvement of Kant's theory of retribution. Such an interpretation is strongly suggested also by Hegel's critique of Feuerbach there. However, a moment of reflection on the structure of his "Rechtsphilosophie" should be enough to realise that it does not really fit into it. The retributive theory of criminal punishment figures in the section named "Das abstrakte Recht". The qualification "abstrakt" refers to the fact that this section deals with the priciples of law as they appear before their institutionalisation in the state. "Das abstrakte Recht" is the equivalent of "natural law", that is a major object of criticism in Hegel's book in so far as the natural law tradtion before him assumed that law would remain the same once it would be part of the state. The impact of this critique for civil law is limited. It remains valid, in principle, as it has been developed in "Das abstrakte Recht". However, it will be supplemented and restricted by the rules which are created by the corporations and the state in the interest of public welfare, as Hegel argues in the last section of his book "Die Sittlichkeit". However, The criminal law is a different matter. In fact, it figured at a very special place at the end of "Das Abstrakte recht", where the step towards the section "Die Moralität" was made. The step is taken when it is pointed out that punishemnt as developed there, has still a subjective form. It is essentially "revenge". The subjective nature of punishment is created by the fact, Hegel wrote, that the offender might also see his punishment as a crime against him. However, this would suppose that the offender beleived to have been justified in doing that for which he was punished. Thus criminal punishment leads to a "discussion" of the "good". The section on morality leads over to the section "Die Sittlichkeit" after the subjectivity of the definition of the moral good has been shown. "Sittlichkeit" is, in contrast to morality, not concerned with ethics, but with a functional analysis of different social institutions under the perspective of their contribution to the realisation of freedom in a well-orderd society and state. It is in this section that one would expect a definition that does not suffer from the functional defects we have discussed in the last section. In fact, one finds the criminal law reconsidered there in a section (218) that is rarely dealt with even by Hegelians. The reason that it would be neglected is that it would seem to irreconcilable with the priciples of crimianl justice that Hegel had developed out in "Das abstrakte Recht".

In that section Hegel argues that crime in civil society is on the one hand much more serious, because it is not only th evictim that is felt as being hurt, but all the members of society who see their norms disrespected and threatened. On the hand, the threat is felt as less great because, in contrast to the state of nature, an organised society can feel much more sure to be able to deal with crime and its consequences. This is why, according to Hegel, the point of view of the dangerousness of crime for society becomes predominant and replaces the point of view of the damage that has been done to the victim. Therefore the sanction will depend on the state of civil society. Depending on the danger that crime represents at a given palce and time, the severity of the sanction may vary such that a theft of a few cents may be punished with the death peanlty at one time and place, whereas a theft of something one hundred or more times more valuable may be punished mildly at another. This, it would seem, is something totally different from retribution! However, Hegel softenend the offensiveness of his view by adding that as civil society had become more consolidated in the course of the historical development punishemnt had become milder. Although this observation was adequate form a historical point of view, Hegel did dwell with the fact that this development had taken place under the influence of the theories of the very same people, Beccaria and Feuerbach, which he had criticised so severely in "das abstrakte Recht". This is understandable, because it would have prompted th ecritique that he had actually declared the principles of "abstract" criminal justice superfluous for civil society.

7.7 Feuerbach: a Kantian deterrence-theory

It is very remarkable that Feuerbach, so heavily criticised by Hegel as Feuerbach rejected retributivism as a matter of principle, and not just, like Hegel, as the result of historical development, regarded himself as a more consistent Kantian than Kant had been himself.

Kant had taught that moral freedom is a presupposition of practical reason, but that it can not be explained in causal terms how free decisions operate upon a temporal-spatial world in which nothing is uncaused. The operation of freedom upon action is, in other words, as mysterious and paradoxical as is the creation of a temporal world out of nothing. However, if we can not know how moral freedom causes action, we can also never know whether an act is purely the effect of a free decision or, wholly or in part, of emotional forces, like the fear to be censured or punished.

Feuerbach also argued that, as morality is a matter of the mind, it is perfectly possible to agree with a moral norm and nevertheless act against it because one's character is to weak to resist the temptation to do so. In fact, Feuerbach regarded Kant's idea that an offender sets a deviant norm not only unnecessary, but even absurd. It would only be correct if an offender were a devil, he wrote. Therefore, retribution in the sense that an offender is treated according to the norm he has set himself, also makes no sense. Moreover, according to Feuerbach, punishment does not retribute anything in a concrete sense. If punishment were just temporally backward-looking, it would only be adding more suffering to the misery already caused by the crime. If criminal punishment should have a justification, it must be because of its beneficial effects with an eye to the future.

To a modern reader the above will suggest that Feuerbach would think that SD or GD provides a justification of punishment. However, he did not and he criticised both forms ro relativism. According to him, the function of the threat of criminal punishment is certainly to deter, but punishing does not take place primarily to deter the offender or others in the future, at least not in the sense that punishment would have to reinforce the threat with the offender or others. In his view punishment is simply a part of making a threat credible. It is true, of course, that the threat has failed in the concrete instance in which an offender is being punished for a crime. However, it does not imply that it has failed before, or that it can no longer be effective in the future. However, to be effective in the future, it will be necessary that it will be implemented every time it fails to deter the offender. How can this use of criminal punishment be justified according to Feuerbach?

Feuerbach's foundation of the criminal law was simply an extension of the function which he, following Kant and the natural law tradition in general, ascribed to punishment in the state of nature. Punishment serves the protection of rights. Its size is determined by its effectiveness. However, in the state of nature there is nobody but the victim who can decide what the required size of punishment must be. Therefore, a violation of another person's right implies accepting the price that the victim has set upon doing so. This is a very strange theory, because it would seem that it would be interesting to become a victim of a crime as long as that is profitable. However, what other standard could there be in the state of nature? Even historically the theory has a certain plausibility in so far as compensation and punishment were hardly differentiated as long as law enforcement was in the hands of the victim and his family. In primitive law, sanctions are usually negotiable, at least within certain limits. Moreover, one of the standard arguments in social contract theory, well-known from Locke but also used by Kant himself, was that the aim of the social contract is precisely to put an end to legal uncertainty and the risk of endless strife about the right measure of legal sanctions. On that basis Feuerbach could explain both why the power to punish and the power to set sanctions should be delegated to the state on condition that the state observes the 'nullum crimen..., etc.'-principle.

Interestingly, this principle for which Feuerbach became the most famous, was therefore not directly inspired by the rule of law-ideal in so far as this ideal is based on the idea that it is unjust to punish someone for something he could not know to be a criminal act. That was a position which Feuerbach could not argue for in so far as the criminal nature of an act is a matter of rationally accessible natural law. The 'nullum crimen sine previa lege'-part of the principle is just a logical implication or a limiting case of the 'nulla poena sine lege poenali'-rule. A particular sanction can only be related to a particular crime if that crime has been determined. The reason why it should be determined, is that Feuerbach saw the sanction as the price the criminal agrees with when he commits the crime. Obviously, one cannot agree with a price unless one knows what the price is. The requirement of the definiteness of the sanction with Feuerbach was therefore not derived from the idea of the rule of law as containing the legality principle. His famous rule was derived, in the first place, from the philosophical necessity of constructing the consent of the offender.

In fact, the "mens rea"-principle and the equality-principle, were also not derived by Feuerbach as consequences fom the rule of law-ideal, but as functional requirements of an effective use of the criminal sanction as he had constructed it. If the function of punishment is not reform or deterring others, but only as a prospective threat for the (potential) offender himself, it makes no sense to punish someone who did not know what he did, or if he did, is structurally undeterrible. Although Feuerbach recognised the equality principle as a principle presupposed by the social contract, he applied it to the criminal law only in so far as inequalities are concerned which are functionally unrelated to deterrence. The equality principle implied, in his view, that everybody should be deterred by the criminal law with equal force. However, that requirement would imply sentences of unequal severity. One of Feuerbach's objections against retribuvism was that it would have to discount social-economic factors determining crime, such that people in unfavourable social-economic circumstances would have to be punished less severely, whereas they are in need of more severe punishment from the point of view of the need to deter them with equal force. As Feuerbach was against uncertainty concerning the severity of sanctions, the implication was that it would have to be as severe as to be effective enough against classes of offenders with a relatively great propensity to crime. Although possibly unneccessary severe for members of other classes, they would only be deterred the more, and if individuals of those classes would neverthelss commit a crime, they would still have given their consent to be punished. Therefore, if it would not have undermined the necessity of a determined sanction, Feuerbach would not have objected against unequal sanctions as long as they could be defended on grounds of effective deterrence.

According to Feuerbach one transfers one's natural right to the state because the state can protect it better. The state's duty is optimal protection of the legal order. Whereas we have just argued that Kantians and Hegelians might be drawn away from pure retribution for functional reasons, Feuerbach can be read as arguing that one should be drawn away from it if doing so implies a better protection of citizens from crime. In fact, this is a very important argument which has mostly been neglected by retributivists. The reason why they did so, is pretty obvious. Their insistence on the respect for the autonomy of the offender may imply accepting that more people become victims of crime in case a sanction that differs from the retributively just one would be more effective as a means of crime prevention. This objection against retributivism has usually been pushed aside by retributivists on the ground that optimal protection might demand the most inhuman sorts of punishment. However, this is obviously a false counter-argument. "Optimal protection" may also and more often imply punishing less severely than required by retribution. Moreover, it does not imply that no constraints on punishment whatsoever might be justified, the more so as inhumanity is not necessarily efficient. Rather the contrary as both Beccaria and Feuerbach have argued.

7.8 Philosophical and practical autonomy

Although Feuerbach's theory may seem to follow rather easily from the function Kant himself had ascribed to the criminal law in the state of nature, Feuerbach remained so much of a Kantian that he felt vulnerable to the argument that offenders only figured as a means and not as autonomous beings in his theory. According to Feuerbach, however, an offender agrees to be punished provided that he knew that he committed a crime and was also familiar with the sanction that he could expect if caught. Following Feuerbach therefore, the offender is not as autonomous as to set a norm of his own, like a revolutionary does. However, differently from someone who is mentally disturbed, he is autonomous enough to decide that he prefers to pay, if convicted, a price for the crime he is committing. He is as autonomous as a buyer and a seller on a market. In fact, if committing a crime would be an autonomous act in the sense of setting a deviant norm, as Kant and Hegel would have it, willingness to pay the price for it must be supposed to be included in that act.

Criticisms of Feuerbach's construction can be found with both Kant and Hegel. Kant did not comment to Feuerbach's theory which was published four years before Kant, who was dementing during his last years, died (1804). However, Kant had argued against Beccaria that the right to punish does not depend on the consent of the offender, a consent which Beccaria had incorporated in the social contract, as the natural right to punish also does not depend on the consent of the criminal. In the social contract only the private right to punish is given up. However, and that is were Feuerbach's theory might have been used to adapt retributivism, Feuerbach's construction makes sense if the given consent would refer to the state's right to deviate from pure retributive punishment for functional reasons. On the other hand, the agreement of the offender with the punishment threatened with by the state does not differ from the offender's assumed agreement with the sanction as set by the victim in the state of nature. As the consent of the offender is "automatic" in Feuerbach's theory, Kant's objection against Beccaria makes no sense against Feuerbach, as consent is not given with the social contract, but by committing a criminal act the "price" of which is known to the offender.

In fact, it is not so easy to understand why Feuerbach thought he was vulnerable to the Kantian criticism of using offenders as a a means instead of as an end. After all, the beneficiary of the criminal punishment, within his theory, is the offender himself as much as it is society, because the threat of punishment helps one against one's criminal inclinations. However, according to Hegel, Feuerbach's theory was an insult to human dignity because he dealt with punishment as if it were a stick to threaten or hit a dog with. However, this objection is paradoxical. It is true that Feuerbach assumed that the criminal threat will often help to scare off people to commit crimes. However, that is an effect which the criminal law may have whatever one's background justification of the criminal law is. Moreover, and that is what makes the objection paradoxical, if anything, the decision to commit a crime becomes more autonomous if it is taken notwithstanding the threat of a criminal sanction. However, Hegel was right, of course, in the sense that Feuerbach did not see the decision to commit or not to commit a crime as an autonomous act in a philosophical sense to begin with.

Feuerbach's criticism of the purely philosophical nature of moral freedom led him to a "methodological" determinism for the criminal law. "Freedom" within determinist theories means freedom in a practical sense, that is absence of special ciricumstances which make a person criminally unaccountable for his behaviour.The reason why Feuerbach was compelled to wrestle with the philosphical concept of freedom was that he was a social contract thinker who had to understand all governmental authority as ultimately based on the consent of the citizens.

In fact, the supposition that offenders consent to be punished in a practical sense is a rather idealistic element in Feuerbach's theory. It is also its weakness. It is not much less unrealistic to think of criminal offenders as deviant norm-setters than it is to think of them as buyers willing to pay a price. There is a class of professional criminals, no doubt, for whom criminal punishment is just "all in the bargain". However, if a driver of a car causes an accident in a moment of inadvertence he is not taking the accident and the punishment it brings as part of a bargain. He never wanted to cause an accident. This does not mean that criminal punishment makes no sense in such a case. A greater criminal threat might have made him a more careful driver. However, he would not have been a more careful driver directly out of fear for punishment, but because the greater criminal risks are a signal to him about the seriousness of the dangers of driving which through this threat he is more aware of. Criminal sanctions, in the case of negligence, are usually not deterrents in a subjective sense, but only in an objective sense, because their primary function is a cognitive one, being a reminder of the seriousness of the risks involved in doing something. Therefore, if negligence cannot be constructed as criminal behaviour in which punishment is part of a bargain, an offender may also very well disagree with the sanction, because he thinks that the punishment threatened is too light or too severe.

7.9 Retribution and rehabilitation

So far for the line of criticism of Kant's and Hegel's absolutism from the point of view of the crime control function of the criminal law. Leaving that aside now, we can return to their theory of criminal punishment as the maintenance of a political-moral balance. The implicit contention of retribution is that an offender will be rehabilitated as a citizen after having been punished. Within the framework of the retribution theories of Kant and Hegel this is understandable if the offender would have learnt a lesson from having been punished. But what if he has not? The critical question to be asked is why society should not simply get rid of chronic offenders, that is, with offenders who either do not want to be reformed or who apparently cannot be reformed.

In fact, Leo Polak's theory of retributivism was a device to cut this question short by founding criminal retribution in a suposedly empirical psychological law, formulated by his doctoral father Heymans, that morally good actions should be applauded and rewarded, whereas evil actions should be prevented or, if that is not possible, be made undone as far as possible and also be punished. Satisfying this psychological desire is a goal in itself then. Obviously such a psychologisation of retributivism implies a shift from the political nature of criminal punishment as it was conceived by Kant and Hegel. The queston to be raised in the case of Polak would be why the political community is the relevant point of reference of criminal punishment and not, say, the village in which a crime took place or, if people abroad feel also affected, a community that is wider than the particular state that claims to have exclusive jurisdiction.

The latter question aside, however, one can doubt if Heymans' moral law is an adequate description of how we actually think about moral sanctions. The analogy to moral retribution that criminal legal retribution is based upon is that making a moral offender suffer often has the aim to restore the relationship that had been harmed by teaching the offender a lesson. However, the restoration of the relationship will depend on the offender's learning a lesson indeed, that is, finding him to regret what he has done. Of course, moral punishment is also practised without an intention to restore a relationship. However, in that case punishing would seem to be an act of pure revenge, a means of the offended person to get rid of his frustration and anger. Revenge may be understandable for psychological reasons, but it is nevertheless an immoral act itself if the aim is only to hurt. Therefore, although it can be understood why the state should act retributively in Kant's and Hegel's theory, it remains unclear why retribution would be a sufficient condition for rehabilitation if an offender apparently has remained unreformed.

7.10 The moral justification of the right to punish

Retributivism is lacking as a moral explanation of the criminal law. However, even if that would not be so, a moral explanation of the criminal punishment is insufficient to explain the right to punish. The question of the relationship between law and morality was critically examined and defended by Feuerbach, at age twenty-one, in his philosophical dissertation "A Critique of Natural Law", published in 1796, one year before Kant had even published his philosophy of law. On the basis of Kant's CI one can deduce the JI as a necessity to defend one's freedom such that one can act in accordance with the CI. But how about a right to do things which are morally indifferent or using one's rights for immoral acts? To the first question Feuerbach answered that morally indifferent acts can nevertheless be conditions of acting morally. Thus, in itself it may be indifferent from a moral point of view whether I buy a certain house or not. Why should I have a right then to enforce a contract of sale, if the vendor does not live up to his promise? Feuerbach's answer was that buying the house might be important indirectly from a moral point of view e.g. because the buyer would like to let his children grow up in surroundings from which he expects positive educational effects. Obviously this is a weak answer in cases were the vendor could prove that the house was not part of any plans with moral significance. However, one might still accept the argument by granting that the law must make abstraction from such subjective and therefore often unprovable matters.

Feuerbach's answer to the second objection was even less convincing. His answer was that one would no longer be morally autonomous if the law would force one not to act immorally. It was, in fact, the same objection as Hegel was to make against Feuerbach's theory of punishment. The objection is unconvincing precisely because it is an unavoidable effect of the law to influence morally relevant decisions. However, influencing is not the same as destroying moral freedom. Moreover, as was argued already, if someone acts immorally the threat of the criminal law notwithstanding, it is more likely that his act was an autonomous one. And even if the argument were acceptable as a principle, it takes quite some moralistic fanaticism to maintain that the freedom to act immorally should prevail in cases when doing so would imply that vital interests of others would be sacrificed. Suppose that someone is drowning and could be saved if another person would reach out his stick. Not helping the drowning person is not a form of interfering with that person's rights. However, it is most immoral to refuse assistance in such a situation. Why would it be wrong then to enforce such a moral duty by the threat of legal punishment? What makes it so important to know if someone who saved someone else would also have done so if he had not been threatened by the criminal law? After all, how can one ever know whether a person acts morally only because he is afraid of the moral censorship which others might submit him too and not just out of an inner respect for morality? In fact, how can one ever know that even about oneself? So why care about just an additional reason to have doubts about the purity of the moral nature of an act?

A very important point of criticism of Kant's theory of the relation between law and morality that Feuerbach did not make is that criminal retribution is not applied to all illegal behaviour. If someone wilfully breaks a contract, I may sue him for damages, but such civil-illegal behaviour is usually not forbidden by the criminal law. An obvious explanation for this difference is that citizen's can protect themselves much better and cheaper against such risks than the state can do via the criminal law. However, that is a functional argument which is irrelevant from the point of view of moral retribution. It would be hard to argue, from the point of view of Kant and Hegel, that a creditor should protect himself against such deviousness of a contract partner, because it would assume that citizens can not, in general, trust each other as morally autonomous beings. However, morally speaking the wilful breaking of a contract can be experienced as a violation of one's rights which is as serious as e.g. an act of embezzlement. In fact, no retributivist I know of has ever provided a satisfying explanation we obviously make between those two acts. The only escape that is left of retributivists would be to insist that all blameworthy illegal behaviour will indeed also be punished. However, it goes without saying that a criminal legal system will then become totally overburdened very quickly.

In fact, what the given example of wilful contract-breaking vs. embezzlement shows, is that criminal responsibility is not based on the identity principle ("when you steal, you steal from yourself") on which moral reasoning is based. Retributivists confuse commutative justice, the justice of private law, with distributive justice, the justice of public law. What is it that makes retributivists so blind for this obvious difference between private and public law?

The reason why retributivists have turned a blind eye upon the question concerning the criteria for (de-)criminalising illegal behaviour is, I venture to speculate, that they can think of no other objectifiable standard for the determination of the criminal sanction.

7.11 Leo Polak's "objectivation"-theory

Primarily, the talio-principle seems to offer a very simple criterium for the nature and severity of the criminal sanction. However, on some further reflection it will appear to be far from unproblematic. The importance of Leo Polak's contribution to retributivism has been precisely to face these problems and to try and find a solution for them.

Interestingly Polak started with working out the GD-theory of his spiritual father, Heymans, who had formulated the moral-psychological "law" we have already referred to above. It took Polak many years to conclude that he had to reject Heymans' view of the criminal law, because he found him wanting in explaining the ineradicable notion of retribution. Retribution is, according to Polak, the essentially economic idea that something negative caused by the crime is compensated through punishment. If punishment were simply about preventing evil, the criminal event might not even be a necessary condition to justify taking measures against someone who is highly likely to commit crimes. In my view Polak's own so-called "objectivation-theory" brought out the essentially moralistic nature of the absolutist theories by concluding that the balancing of moral joy and sorrow is the business of the criminal law. The unsolvable problem with Polak's theory is the same as with Kant's and Hegel's: how to explain that not all blameworthy illegal behaviour is punishable according to the criminal law? Moreover, Kant and Hegel could appeal to a theory of state to explain why only illegal and not all immoral acts might be punished by the state. Polak, however, had no such theory of state. His appeal to a psychological-moral law begs the question where the line between law and morality should be drawn.

The really interesting aspect of Polak's theory is his operationalisation of retributivism, which explains why he called his theory the "objectivation-theory". Polak rejected the talio-principle for a number of reasons. Firstly, if it were taken seriously, retribution should depend on the subjectively determined amount of suffering of the victim. Secondly, it cannot explain the usually crime-unspecific nature of the criminal sanction. Thirdly, talio, if taken seriously, would lead to practices which run totally counter to the humanitarian tendencies of the modern criminal law. Finally, the criterion is meaningless sometimes when taken literally. What would be, for instance, an appropriate sanction for dangerous driving? Let the offender cross a minefield? Instead, Polak, argued, retribution is about equalisation in the sense of letting the offender suffer as much as he made his (potential) victims suffer. However, according to Polak, the equivalent of suffering that is to determine "just" punishment, cannot simply depend on the amount of suffering of the victims.

Instead, Polak proposed what he regarded to be an objective criterion: the "punishment" a law-abiding citizen would be willing to undergo for not having to commit a certain crime.

In Polak's psychological approach a law-abiding citizen is not simply someone who never feels the inclination to commit a crime. It is someone who would experience more pain from doing so than satisfaction because he has a "trans-egoïstic" attitude. The sympathy with his potential victims makes him suffer in anticipation of the pain his victims would feel such that, on balance, he feels better off not to commit a crime. An offender, by contrast, is lacking in trans-egoïstic attitude. He cares so much more about the satisfaction of his own desires than he feels pain in sympathy with his victims, that he commits a crime. However, the reason that he does so is that he also suffers more from the non-satisfaction of his immoral desires than does a law-abiding citizen. As punishment is about equalising offenders with the law abiding-citizen, the latter must ask himself how much suffering should be added to his egoistic, immoral desires to let the balance of forces with his "trans-egoistic" feelings skip to the other, criminal side.

Note that the law-abiding citizen is not an ideal citizen, a citizen who feels no criminal desires whatsoever, but only one who has a balance between criminal and anti-criminal desires such that he will never break a law. However, by giving up the standard of moral innocence that was still used by Kant and Hegel, who demanded full retribution, a paradox is created.

The greater/smaller the difference between the moral and the immoral desires of the law-abiding citizens, the more/less severe will criminal legal sanctions be. Assuming that the deterrent effect of the threat of criminal sanctions increases with its severity, at least to a certain point, more in themselves not-law abiding citizens will not commit crimes. The paradox is therefore, that the closer law-abiding and not law-abiding citizens are from a moral perspective, the greater will their difference be in terms of their criminal legal behaviour, although this is compensated by the fact that offenders will not be punished as severely as when the moral gap between them had been wider.

Therefore, people who are equally immoral may become criminal offenders or not depending on the legal community they are a member of. However, it is precisely through a realisation of this paradox that another and more stable point of objectivity can be reached.

7.12 Polak and Rawls

According to Polak, there is a point where the law-abiding citizen decides that he will rather let someone else suffer than being "punished" himself. If he would totally identify with the victim he might, if possible, make himself the victim of his own crime. In other words, in contrast to Kant and Hegel, Polak would seem to have allowed a certain amount of non-identity or selfishness among citizens. Polak's theory is decidely more psychological and less idealistic and moralistic than Kant's and Hegel's. Moreover, Polak's point would seem to be that one should not only sympathise with victims, but also with offenders, as the law-abiding citizen is not a holy man. He has criminal inclinations, although they never overcome his transegoistic inclinations. However, realising the paradox we have analysed, the law-abiding citizen must admit that by taking his own possibily only marginally moral nature into account in determining the size of punishments, he is a creating a relatively arbitrary division between criminal and non-criminal citizens. To that extent taking himself as the starting point of measurement is not being objective in the sense of being unpartial. Moreover, one's moral or immoral nature is a contingently given. In fact, the concept of moral "character" which Polak used, is based on an outdated psychology which abstracted from social and situational factors which determine whether someone will commit a crime or not or will be arrested or not. Therefore, in a really objective approach one should abstract from one's own contingent nature and place in society.

The "law-abiding citizen" who is willing to suffer in order not to commit a crime seems a highly artifical and idealistic concept. Indeed, Polak's "objectivation"-exercise is familiar with Rawls "veil of ignorance"-device in that respect. But not only in that respect. When the law-abiding citizen performs the thought-experiment of self-punishment, he imagines to stand under the same pressure as another person who is driven to crime by immoral desires. It is, therefore, an exercise in sympathy. However, it is a biased exercise, because Polak reasons from the point of view of someone who does not imagine having a real chance of becoming an offender. A neutral point of view would be one which is in conformity with the "veil of ignorance"-standard, that is, the point of view of someone who does not know wether he will stand on the positive or the negative side of being victimised or being punished by the state. Not knowing that he will have to assume average chances of becoming a victim of crime and of being an offender. Therefore, he has to balance his interest of being protected against crime against his negative interest of being punished after committing a crime. That is, in fact, the point of the Janus-face view.

Although the Janus-face view differs principally from Polak's theory because it is GD-functional and not moral, it nevertheless provides an explanation of the indeed ineradicle retributive element in criminal punishment. Moreover, although it does not ground the criminal law in morality, it places the criminal law as a function within a political-moral community. It adds, in other words, this important element of early absolutism, which had gone lost in Polak's theory, in which the relevance of the state as the relevant community was treated as something obvious.

7.13 Rawls and the Janus-face view

Have I then, without saying it, presented a Rawlsian theory of criminal justice? Although that is strongly suggested by the analysis in the last paragraph, such a claim would be wrong. Firstly, the basis I have chosen myself for the Janus-face theory is H.L.A. Hart's theory of a "Minimum content of natural law". Although Hart borrowed the idea of the seperation of the question of the justification of the CL as an institution from the question of its distribution in individual cases from an early article of Rawls', he did not accept Rawls suggestion in that article that the principles of distribution might also be understood in a utilitarian way. In fact, Rawls produced arguments to that effect which remind us of Crombag's psychological understanding of criminal justice. However, this line of reasoning was later given up for the "veil of ignorance"- construction in A theory of Justice: "In agreeing to penalties that stabilize a scheme of cooperation the parties accept the same kind of constraint on self-interest that they acknowledge in choosing the principles of justice in the first place".

The difference with his earlier approach is that it is no longer social utility which determines the outcome, but individual utility, be it one of a totally abstract kind. It is an individual with average chances of becoming a victim of crime and of becoming an offender. In fact, this is the same individual that appeared in our theory, although we did not assume a fictitious individual. As far as the CL is concerned, it can be maintained that it is in practically everybody's interest because it is indeed impossible to know if and to what extent we ourselves and those we care for might become a victim of crime or an offender.

Nevertheless, there are two important differences between Rawls' approach and the Janus-face view of the CL. The first difference is that Rawls, unlike Hart, did not face the problems concerning the equality and the proportionality principle. The second major problem is that it would be incoherent, within Rawls' theory of justice, to use a concept of justice that is restricted to the principles of criminal justice in a narrow sense. In the case of Rawls, the criminal judge would have to take a host of factors into account to make sure that punishment would not only be just in a criminal legal sense, but also fair in a wider sense. All sorts of personal and social-economic disadvantages that make one person more likely to become an offender than another would have to be equalised in the Rawlsian criminal sanction.

Obviously, this is the consequence of the general nature of the equality principle in Rawls' political philosophy, whereas the Janus-face theory is based on a revised reading of Hart's "Minimum content of natural law". It uses an equality principle that is not general, but function-specific for the CL: an equal right to be protected against crime or to be charged for committing one without any possibility to discount that right against rights of equal treatment in other functional domains. However, although function-specific the Janus-face view is, differently from Polak's and Rawls' theory, not just a theory of criminal punishment, but one of criminal punishment as one of alternative means of crime prevention.

8. The sociology of the classical and the neo-classical model

8.1 The internal and the external perspective

We have analysed the conceptual stucture of the classical model. We have also developed a coherent philosophical framework for its justification. It is now time to step out of a predominantly conceptual analysis into a predominantly external, sociological perspective on the classical model and its development. However, in doing so, the interest will still be guided from an internal point of view, because our leading question is if and to what extent the classical model can be applied to modern CL.

8.2 The progressive nature of the classical model

The sociology of the evolution of the CL warns us against naïvety vis a vis modernist interpretations of this development in terms of the "progress" of humanity. It has been pointed out, firstly, that "brutality" in the CL is relative to the occurrence of brutality in society in general. Cutting the hands of thieves may seem utterly brutal to us, but it was much less so in a society in which physical deformity was much more common anyway and monetary sanctions against the numerous poor were impractical. The galleys or banishment were the only alternatives short of the death penalty, but these sanctions were disproportionate for petty crimes. Secondly, both authority and crime were perceived in a religious context, that is in the context of sin and redemption. Paying a terrible price on earth could be used as a discount when knocking on heaven's door, with the authorities' role as advanced instruments of God's wrath as much as of public order. As the forces of order were most ineffective in combatting crime, punishing offenders was more a show of political power than an attempt to do something serious about the social problem of criminality and its underlying cause, poverty. This was believed to be endemic, as indeed it was, until the second industrial revolution.

The great impact of the Enlightenment consisted of the combination of a secularised vision of the function of the CL and a belief in the possibility of social progress, both materially and spiritually. However, it was not so easy to get rid of all the theological and metaphysical shackles that centuries of Christian religion and philosophy had produced. Precisely for that reason we should not be over-critical of the progressive nature of Enlightenment humanism. For all our historically relativist awareness of the dialectics of humanisation and disciplining, it is more appropriate to take a longer-term perspective in the examination of the process of rationalisation in modernity. In fact, the disillusions concerning the wholesome effects of imprisonment on crime came fairly quickly and great debates on prison reforms started everywhere after the first quarter of the nineteenth century. They have never stopped. Functional rationalisation of the CL did not stop after its first, metaphysical-speculative phase.

8.3 The empirical turn: neo-classicism

Although the movement for prison reform began fairly early in the nineteenth century, the classical model of the CL became an object of a very principled discussion only during the last quarter of the nineteenth century. Very interestingly it took the form of an internationally organised discussion. In 1889 the Internationale Kriminalistische Vereinigung (IKV) was founded by the German Franz von Liszt, the Belgian Adolphe Prins and the Dutchman Gerard Anton van Hamel in 1889. This very influential Association criticised the classical view on account of its speculative, non empirical-scientific nature.

Although the IKV did not reject most of the legal doctrines of the classical view and although it believed itself to stand, at least by and large, in the same tradition of Enlightenment and humanisation of the CL as the classical view, humanisation had become a means rather than an end of criminal policy in the approach of the IKV. The IKV was dominated by what was soon to be called "neo-classicism", that is the still prevailing view of the CL.

The main tenets of the IKV were that, firstly, the classical view's pretention that punishment would serve SD was untrue and that, in fact, rather the contrary was the case. Secondly, that the GD-effect of criminal punishment was very insufficient as a means of crime prevention. The IKV favoured a less abstract view on crime than the classic thinkers had taken. Instead of judging crimes, it advocated judging criminals. In other words, it rejected the concept of criminal justice of the classical view with its proportionality between guilt, seriousness of crime and the severity of punishment. The IKV advocated new means of controlling or reforming convicts and it advocated the study of the social-economic conditions that influence crime rates. In fact, the IKV was stimulated by and stimulated in its turn the programme of modern criminology.

The IKV's two main tenets as such have been borne out by criminological studies again and again. Its most important pragmatic ideas have also been widely accepted. The IKV advocated the replacement of short prison sentences, believing, and rightly so, that it would help avoid the criminogenic effects of imprisonment. As a replacing device it advocated fines and conditional imprisonment. Furthermore, it argued for an approach of juvenile delinquents such that punishment would be secondary to re-education to correct seriously deviant behaviour among the young. In that approach it did not really matter whether a crime had been committed or not (yet).

These two propositions did not radically contradict the classical view. Beccaria had stated that the useful effect of punishment should be reached at the least possible cost for both the offender and society. In so far as the costs for juvenile delinquents might be higher than they deserved as punishment, it could be argued that minors do not count fully as citizens yet and should be educated as citizens first. It was only the IKV's third pragmatic proposition which represented a clear deviation from the classical paradigm and which caused, in fact, a great controversy within the IKV, notably between its German and its French sections. This was the in Germany almost uncontroversial proposition of allowing special measures against the hard core recidivist who were (and are), as statistics show(ed), responsible for most of the more serious sorts of criminality. This special treatment could consist in therapy but also, where therapy was not possible, in simple containment. In fact, there was a tendency within the IKV to think about this group as "social degenerates", as a group which is simply hopeless and therefore to be excluded from civil society.

Evidently such a proposal was irreconcilable with the humanist optimism underlying the classical view, that did not include a reservation for hopeless cases which could be excluded from the social contract. However, it should be noted that the advocacy for the death penalty was not very strong among the members of the IKV. In all fairness to the IKV it must also be admitted that nobody, as far as I know, wants to do away with the existing possibilities to impose periods of treatment on perpetrators of aggressive crimes with a high risk of recidivism. In many cases such treatment is a pretty hopeless affair and is nothing else but confinement in disguise. The "dehumanisation" advocated by the "modern direction" was therefore limited to the worst class of criminals. One may, of course appeal to a rather academic distinction between treatment and punishment and demand guarantees against uncontrolled terms of treatment, a demand which, in fact, was also made by the IKV. How come then that the "modern direction" has been criticised more recently as having promoted an approach to crime that is more and more undermining the classical doctrines of the CL?

8.4 The waning of neo-classicism

Although the allegations against Von Liszt & Co are rather anachronistic or only justified as regards the important influence of the criminologists among them, there are certain similarities between the socio-legal conditions at the turn of the last century and the turn of this century. Crime rates were rapidly rising as a consequence of industrialisation and urbanisation at the end of last century and it was pretty obvious that this rise could not be controlled simply by punishing more severely. On the contrary, the existing system of criminal justice was believed to increase crime rates through the criminogenic effect of imprisonment. In other words, without changing the system it was believed to steer into an ever greater crisis.

In our own time we are going through a similar process. Notably the U.S.A. demonstrate that purely repressive policies are largely counterproductive. However, the reasons for the present crisis are very different. The present rise in crime rates is very much a drugs-related rise, a problem which, according to some observers, should be managed through a policy of decriminalisation. However, part of drugs-related crime has to do with a more general reason for the present impotence of the administration of criminal justice. We are living in a society of organisations and crime becomes crime in organisations or at least an organised sort of crime, which is much harder to control, the more so, as much of this crime has international dimensions. In reaction to these developments the state has the inclination to do away with many of the guarantees of the classical CL. It is in this instrumental approach of the law that there is some affinity between the "modern direction" and contemporary developments. Nevertheless, to make the modern direction responsible for the present developments is rather far-fetched and anachronistic. After all, the propositions by Von Liszt and Co. have all been accepted before the present developments and apart from the marginal infractions it made upon the classical criminal legal doctrine, it did not imply nor lead to a fundamental rejection of the classical view. And in so far as such tendencies existed within the IKV, they have been redressed by the both reformist and rule of law minded Association International de Droit Pénal, its still existing successor, founded in 1924.

The IKV can nevertheless be seen as a logical step between the classical model and the contemporary crisis of the neo-classical model (see below) in so far as it undid the classical model of some of its universalist and idealist metaphysical features. Firstly, it restricted the class of citizens to which the classical model may apply by excluding minors (reeducation) and chronically dangerous offenders (therapy or confinement). Secondly, by advocating the replacement of shorter prison sentences by fines and conditional prison sentences, it undid the classical model from its fascination with imprisonment. As we have argued, this fascination is understandable against the metaphysical, to wit, social contract based understanding of punishment, in which punishment appears as temporary exclusion from society. However, if the classical model is interpreted as the epochal step towards a secularised and functional approach of criminal punishment, the replacement of imprisonment by fines was just a further step in its functional rationalisation. A functional rationalisation both regarding the efficiency of reaching the goal of the system at the least costs for both society and the offender, as well as regarding the differentiation of criminal punishment from morality. Much less than imprisonment fines symbolise exclusion from civil society. On the contrary, they rather symbolise punishment as the price of crime, they commercialise and therefore normalise crime to a certain extent as an inherent but necessary risk of social life. It reflects a shift of attention of the criminal system from the authority of the state (the social contract) and the criminal individual and his immoral nature, to the effects of punishment on the social fields in which criminality is generated or is prevented from being generated. In so far as the reforms of the CL as proposed by the IKV have been a step towards further secularisation and functional rationalisation of the CL as well as towards a further functional differentiation, it can be seen as a bridge towards the "postmodern" developments in CL that we are witnessing today. However, these latter developments can no longer be understood within a (neo)-classical framework, as I will argue.

9. An evolutionary view of post-classical CL

9.1 Post-classical phenomena in the CL

The (neo)-classical view still dominates the CL. However, this is a domination in the legal imagination as created in legal class-rooms and the media. It is the imagination of the CL as dealing with speeding, thieves and thugs as well as more or less solitarily operating white collar criminals. This domination exists less and less in legal practice and it is on the wane in legal class-rooms.

The (neo-)classical view of the CL is tied up with the idea of the rule of law. This idea is obviously at odds with the fact more and more rules were introduced during this century which are based on criminal risk responsibility instead of on criminal intent or negligence. The idea of the rule of law is based on characteristic presumptions concerning the possibility to know and predict the CL, to identify reliably criminal acts and offenders as well as their responsibility. According to many authors, contemporary law and life have produced ever more criminal rules and cases in which these presumptions are no longer met. Some writers have a downright apocalyptic vision of these developments. More and more the CL is about organised crime and about crime in and by organisations, both public and private. These sorts of crimes are committed by employers, administrators, bankers and traders and have to do with safety and quality regulations, working hours, money laundering, industrial espionage, illegal copying, bankruptcy, corruption and blackmailing, manipulation of information and stocks, computer-crime, pollution, taxes and social security charges. It is white collar crime, but white collar crime of an often organised, fairly administrative-technical and of a more or less professional kind. Often these crimes are difficult to define, detect and prove, the more so as they increasingly become transborder crimes as globalisation proceeds.

Some of the more traditional crimes, gambling, drugs, trade in women, children, illegal aliens, protected animals and transplantation-organs, car-theft and terrorism, are nowadays also more and more professionally organised. Sometimes the organisations responsible for them are also well connected to the "upper" world. Moreover, they also have become more international as organisations. These sorts of crime share in the modern nature of crime in so far as the organisers and their connections in the "upper" world are often difficult to trace down. Sometimes "unconventional" methods of investigation and prosecution are necessary to catch them. Although organised crime and crime by organisations are differentiated according to whether organisations have an officially legitimate purpose, such a distinction makes only limited sense from a criminological point of view. In fact, criminal organisations sometimes operate in a shadow area in which they are tolerated more or less by the authorities who are too dependent on them. That dependency is not necessarily created by corruption, although collusive relationships will often lead to it. Criminal organisations may at the same time provide useful services in crime control or otherwise.

In part these new types of crime suffer from problems of definition. The increasingly complex and technical nature of society produces norms which make a relatively high demand on citizens' knowledge of legal rules and the situations in which they apply. The negative effects of the behaviour they proscribe are often difficult to foresee or unlikely, although sometimes severe, when they occur. Although these norms put less of a burden on organisations who specialise in the fields to which they apply, they are sometimes formulated so broadly and vaguely that it remains difficult to foresee when they will apply. In part the burdens of these rules are laid upon organisations without any specific personal responsibility, which implies that the ground for such responsibility is also very remote. It may be that the criminal act is just an outcome of a defective organisational design, the criminal consequences of which may have been almost unpredictable or very difficult to avoid always. Adaptation and change of organisations and training of the personnel in the required skills can be very time-consuming. The other side of this coin is that precisely the complexity of rules and regulations and the special knowledge and organisation that is necessary to deal with them, create many criminal temptations and opportunities.

9.2 Functional aspects of post-classical CL

Although the nature of such CL would be decidedly non-classical, it does not follow that the post-classical use of the CL is dysfunctional. A criminal judgement may be precisely the means to draw public attention to the fact that a certain form of action should be avoided and is punishable. It goes without saying that such criminal norms are suitable instruments of political and administrative manipulation. Precisely because it is sometimes hard to predict how lawsuits will end, the threat of a lawsuit can be used as a means to make people go along with the wishes of authorities even if they are unrelated to the subject matter to which the rules are relevant. In practice they are often just bargaining chips to come to deals to prevent further criminal prosecution. Such deals often remain out of the sight of the public. In fact, many organisations are not afraid of punishment as such, but of the nuisance and the public disrepute that goes along with criminal procedures. The CL used in this way is an extra asset of public administration, as similar effects may be reached by threatening with means of an administrative or civil legal kind. Post-classical CL therefore has a feature in common with pre-classical law, where "setting examples" also was the solution to the blatant incapability of governments to combat crime.

Another feature of post-classical CL which can remind us of pre-classical situations and is therefore a cause of great concern, is the gradual erosion of the procedural guarantees of the CL. There has been a global increase in crime rates due to economic growth, urbanisation, migration and mobility. The amount of criminality reported, crimes detected and offenders apprehended has increased much less, although this increase is often too much already to be handled properly by the authorities. Understandably, efficiency becomes the overriding concern under these circumstances whereas global inefficiency will be compensated by show-cases such that real and television staged cases become almost indistinguishable for the public. In order to be more efficient, there has been a trend towards allowing the use of under-cover agents, agents-provocateurs, criminals who are allowed to testify under the pledge of (partial) immunity, anonymous witnesses and many new technical devices which threaten privacy, but the possible abuse of which is very hard to prove.

Although plea-bargaining is far from being as common in Europe as it is in the U.S.A., pre-trial negotiations ending in agreements to "buy off" a trial are used frequently now even in non-trivial cases. In the perception of the public the distance between cops, who are more easily suspected of corruption, and robbers, who tend to be seen as smart guys, has become much smaller. It is public knowledge that in some areas (drugs, gambling, prostitution, sweat shops, environmental crime) tacit agreements exist between criminal entrepreneurs and the authorities because the latter are dependent on them for controlling the more severe or irregular criminals or because the state has a fiscal or administrative interest in their activities.

A lesser distance between state and crime might be interpreted within a broader framework of a "horizontalisation" of the relationship between state and society. Informalisation of the CL implies that social-economic and financial power become decisive factors in the criminal process, in which expertise, knowledge and access to information have become crucial factors. Moreover, as the world is becoming increasingly organised, crime control is becoming the responsibility of non-state organisations more and more. In fact, its financial proportions surpass most government's budgets in the field of crime control nowadays. Organisations employ their own security forces and they have internal procedures to deal with crime by employees as they have a strong interest not to "show their dirty laundry". At the same time, industrial espionage and (patents-)warfare have become as common between industrial organisations as they used to be between states. In fact, state intelligence services, especially after the end of the cold war, seem to be very much involved in industrial espionage nowadays.

9.3 Social structural aspects

One can be wary of these developments because they fly in the face of the classical image of the CL. However, it is hard to imagine that one could stick to one's classical guns in view of the societal causes of these developments. In this connection it may be useful to realise that the development of the classical CL was also a reaction to the change in the nature of crime which went along with the processes of economic change and urbanisation. The classical CL was the product of the rising bourgeois class which desired to control the government on the one hand, and the growing masses of the urban poor on the other. For the new middle classes it was intolerable that the CL would remain an unpredictable factor which could be manipulated to a large extent in the interest of political power holders. At the same time, they were dissatisfied with the mainly symbolic use the government made of the CL by hanging, torturing and maiming a few offenders. The nature of crime changed from predominantly physical violence and petty theft by vagabonds and drunkards to more professional forms of property crimes, a more calculated and intelligent form of crime, at least as far as victims from the middle and higher classes were concerned. Against this sort of crime a regular, professional police force was believed to be a necessity for the apprehension of offenders, and imprisonment was seen as the most efficient means of control and reform. Rationalising and limiting the use of the criminal sanction was not only a matter of humanisation, but also and maybe more importantly a matter of efficiency.

A recognition of the ideological nature of the classical CL makes us realise that the present transformations of the CL may be as fundamental as was the quick rise of the classical CL itself. The classical CL was part of the process of functional differentiation characteristic of modernity. The neo-classical model represents a further differentiation and rationalisation within the classical model. The present, post-classical process is, I believe, one of an even further, more radical functional differentiation. It does not only take place within the CL, but it is also affected by more general legal changes of this kind in which the disintegration of older models is more apparent than processes of functional resystematisation that are going on at the same time. This process is one in which pieces from the traditional branches of the law, private, administrative and CL, become relatively separated from other pieces within their branches, but become more integrated with pieces from other branches in relation to particular social fields and problems. More and more, we will find specialisations like computer law, environmental law, etc. etc., which include and partially integrate private, administrative and criminal legal aspects more or less coherently. To characterise these processes of functional resystematisation, I will use three interrelated concepts: semi-autonomy, horizontalisation and supercomplexity.

Semi-autonomy is a concept which was first used in legal anthropology. It refers to the capacity of a particular social field to develop and maintain its own norms or to use official norms for its own deviant purposes. In contemporary politics and administration, the semi-autonomous nature of many social fields has been recognised by the state administration which realises that it can accomplish very little unless it uses the mechanisms of semi-autonomous social fields for its own purposes. This explains the horizontalised nature of modern law and administration. The state finds strongly organised interest groups in its way all the time, and it can accomplish little or nothing without getting the support of influential groups within the semi-autonomous fields it would like to influence. The neat dividing line between public and private legal powers that was part and parcel of the enlightened modern state is, to use Marx's terminology with intentional irony, withering away. The consequence of this is that the law, in our case the CL, which had developed its own functional rationality, becomes more exposed to the functional rationality of semi-autonomous social fields. This results in supercomplexity, a phenomenon that we can explain with the concepts of the structural, the non-structural and the external rationality of the CL, as explained in section 2 sub C.

Whereas the structural internal rationality of the classical model of the CL is GD, its non-structural rationality depends on the social field it is operating in. Thus a CL against pollution obviously aims at reducing the amount of pollution. The CL becomes supercomplex if it starts reflecting upon its effectiveness in relation to its non-structural internal rationality. To do so, the rationality of the semi-autonomous social fields it is operating in, which may be external to the CL, must be taken into account. A particular field may be primarily orientated to economic gain and not to minimising pollution. The application of the CL within such a field may be very difficult unless important economic incentives are created and social control mechanisms which already operate within such a semi-autonomous social field, are mobilised to make the law effective. In this way the operation of the CL becomes much more interwoven with the social field it operates in. Understandably, it then becomes difficult to keep thinking in terms of general principles, rules and administrative policies for the CL as a whole. Instead of applying traditional generalisations, the rationality of the law becomes more localised and instrumentalist, that is, its expected effectiveness becomes decisive for the principles, rules and policies that will be designed and followed.

Thus, in the given example of pollution control, three different rationalities have to be brought together: the rationality of controlling deviant behaviour, the rationality of restricting pollution and the economic rationality of doing this as efficiently as possible. Depending on the circumstances, the best solution may be: technical prescriptions, flexible but increasing restrictions on pollution whereby the technical aspects are left to the individual firm, or, finally, the so-called 'bubble-concept', that is, pollution rights which are tradable between firms. The first could be said to represent the traditional solution. The second takes a dynamical approach to anti-pollution techniques, which implies a more negotiating, longer-term orientated style of controlling crime. The third approach represents an even more "output-oriented" style of crime control as it also includes the mechanisms of the economic rationality of pollution-rights within a certain social field. In fact, the latter approach makes the participants in the social field to be controlled interested in it not only on the passive side, but also on the active side, because controlling competitors' environmental performance becomes part of one's efficiency calculations. In order to design and apply efficient legal rules it will be necessary for authorities to gather sufficient information and create sufficient willingness to cooperate with them among the people and organisations who are being controlled. Creating an interest to do so spontaneously, is therefore a smart device.

The three solutions have differential impacts on the technical and legal aspects of control of violation of the norms, whereby civil, administrative and criminal-legal forms of control must also be compared in terms of their efficiency and differential effects. Evidently then, the rationality of criminal control becomes very much interdependent with the rationality of pollution control and its social-economic aspects. In fact, the radicalisation of the process of functional differentiation is not something which is restricted to the CL. Instead of thinking in terms of "the" CL we will now think more and more in terms of the CL of a certain socio-legal function (environmental law, media law) or object (computer law, space law) whereby the supposedly general principles of the CL are transformed, restricted or refined depending on the functional context. Similar processes of disintegration and resystematisation take place in the fields of civil and administrative law, which figure in an interrelated way and sometimes as functional alternatives in the context of such functions or objects. Processes of differentiation and reintegration remain very dynamic, however, as legal problems can be related to different subfields at the same time (e.g. pollution control in space).

Functional differentiation, that is, semi-autonomous social fields, horizontalisation and supercomplexity, can also explain, but not necessarily justify, the way that the CL operates in more traditional areas of criminality which have become organised. Take, for instance, the highly dubious and illusory way that the CL operates against drugs-organisations. The drug-scene is a semi-autonomous social field in which specific norms are maintained for the participants in their different roles. The task of crime control is supercomplex because the structural effects of the CL can be counter-productive in terms of health protection and the frequency and severity of drug-related secondary crimes. In fact, as long as a viable form of legalised drug-consumption has not been made possible, a realistic goal of criminal policies in this area can hardly be to suppress the highly profitable illegal trade of drugs altogether. A policy of relative containment is the maximum achievable. In order to realise this it will be necessary to prevent monopolisation of the field by criminal actors. In order to strike down potential monopolists it will be necessary to get the cooperation of competitors. This can be accomplished only if drug-use and even trade is allowed within certain limits and criminal non-liability and safety of under-cover agents and "pentiti" can be guaranteed. Clearly, as soon as the state can only control crime by making deals with criminals and their organisations, it operates in a horizontal rather than in a vertical relationship to the social field it wants to control. It is debatable, of course, if a radical decriminalisation and state-controlled commerce of drugs would not be a much better policy. However, as long as trading of drugs is an affair of well-organised criminals, it is difficult to see how the police could be at least somewhat effective without using means of operation which are highly dubious from a classical point of view.

9.4 An evolutionary interpretation

Can we see the evolution of the CL making full circle then, as postclassical CL seems to remobilise mechanisms of "primitive" social control in such diverse fields as the control of drug-use and pollution, whereas whole schools of criminology now try to build upon such mechanisms? Is this a general trend of the evolution of law, as Donald Black has suggested? Or would it be more adequate to think in postmodern terms of unsynchronised paths of evolution such that some parts of the CL can be in a more advanced stage and others, at the same time, in a rather primitive one, as it takes time to learn how to cope with new social phenomena and reconfirm the inherent doctrinal limits of the CL? Or is post-classical CL part of a more general change in the way in which society observes itself, just as the classical CL was when it became established? These are very speculative questions and the more so as the three theories are not mutually exclusive.

The process of functional differentiation of modern society can be seen as a development of possibilities of self-observation of increasing complexity. A first step was the differentiation between state and society in which the unity of society is projected in a special institution, the state, which facilitates the self-observation of society. Society duplicated itself in this way in order to act upon itself, as if it were a human individual.

The second step consisted in the legal regulation of this process of self-observation via legal constitutions, fundamental rights and administrative procedures. The CL was a special part of the latter because it is an "extraordinary" sort of law in the literal sense of the word. Unlike ordinary administrative procedures, it deals with the contingency of the legal order itself. This may seem to be true of any CL and not just of its classical model. However, the special nature of the latter, which became more pronounced in the neo-classical model, is precisely its self-observational character. It does not attempt to radically destroy whatever threatens its existence, but it attempts to think about the most efficient way to control crime, conscious of the fact that a capricious and limitless CL might be a much greater threat to the survival of the legal order than crime itself.

The third step was that, as described, administrative, criminal and private law became more directly interconnected in relation to specific social functions or objects rather than to their own doctrinal unity. It is a further step in self-observation because, in our case, "crime" is no longer seen as predominantly caused by individual deficiencies which have to be kept under control, but as something which society and the law produce themselves and which may be better prevented in coherence with non-criminal legal or non-legal means. In the classical model "crime" was put in a functional perspective only in so far as "the offender" was treated as a contingency with which society had to cope reintegratively. In the neo-classical model it was recognised that reintegration can not just be a matter of the CL. Both systems were reactive, homeo-static systems. By contrast, a post-classical CL not only regards the offender as a contingency, but also crime itself. It is an interactive system. Its aim is to influence social organisations, and not, at least not primarily, individuals. The world of organisations responds to the intervention of the CLS. Some organisations are being excluded, others learn and adapt, which may imply that they remove personnel that cannot adapt. Conversely, the CLS will have to adapt again to its ever changing object.

If we try to account for this dynamisation of the classical model in terms of the Janus-face view, one could say that the relationship between deterrence and rehabilitation has been reversed, although it would be better, in connection with the replacement of individuals by organisations, to speak about prevention and reform. Whereas in the (neo-)classical model rehabilitation was conditioned on paying the price of punishment, in the post-classical model adaptation of organisations has become a condition for prevention. Where reform of organisations cannot be expected, destruction instead of rehabilitation through reform will prevail.

9.5 The justification of post-classical CL

Obviously the classical category of "mens rea" makes little sense in the context of a post-classical CL that is primarily about the accountability of organisations, and not, primarily at least, about the responsibility of individuals. However, the negative consequences of criminal responsibility of organisations are always borne, in the last instance at least, by individuals, although these individuals are not necessarily the most responsible ones for the criminal liability of an organisation. This raises the critical question of the justification of this sort of responsibility. However, already in the neo-classical CL we find instances of functional responsibility, that is, risk responsibility, which cannot be accounted for by the classical model with its "mens rea"-requirement and which is usually called "strict liability". "Strict liability" is a very controversial issue a thorough discussion of which would demand a separate article. However, not all forms of strict liabilty are forms of risk liability as here intended.

Criminal risk responsibility, it would seem to me, is a form of accountability based on the supposition that although "mens rea" may have been absent, an accountable person has voluntarily taken a certain risk of causing harm by his participating in certain social activities. In other words, risk responsibility is is a form of negligence that is just not as exceptional as "ordinary" negligence is supposed to be. The exceptionality invloved in risk responsibility is not the behaviour in question, but the nasty consequences of that behaviour. The rationale of such criminal risk responsibility is that the chance that such harm will be caused, is, in general, reduced by the threat of the CL. It is only in this way that we can explain e.g. that drivers who cause a fatal accident due to an inattention that happens to practically all drivers now and then, are punishable, even though its occurrence with fatal consequences is almost a matter of chance. However, the reason that such behaviour must nevertheless be punished more severely than the identical behaviour without such consequences, is that punishment is assumed to make drivers drive more carefully.

Evidently, life in our "risk society" gives much more occasion also for criminal risk responsibility than ever before. This is especially true were risks can better be prevented through organisational reform. The principle of distribution of criminal risk responsibility can be derived from this reflection. Obviously, those who are responsible for taking a particular risk in view of the profits they envisage notwithstanding those risks, should be held accountable to the degree in which they have failed to take the necessary measures they were competent to take in order to prevent the occurrence of that risk or at least to insure those risks.

Although one should be wary of an idealisation of crime as far as the motives of offenders are concerned, it seems to me that what are postclassical CL's deficiencies from a classical point of view, can also be seen as a part of attempts to deal with crime in a more integrated and sociologically prudent way. The inner wisdom that can be seen as hidden in the post-classical CL is that it is much more efficient to prevent crime by changing social conditions than it is to punish just to deter.

© Réseau Européen Droit et Société