The treatment ideology emphasizes the character of the individual offender when choosing the line of sanction. The neo-classicists place the main emphasis on the character of the crime. Both extreme positions of the pendulum lead to the loss of advantages offered by the opposite position. In this situation, it is tempting to try to combine the two approaches, to get the best from both of them. With a little help from computers, that might be accomplished. Computers have an unlimited capacity. They could create order. They could combine all relevant individual attributes and give an accurate prediction of possible recidivism. At the same time they could take into consideration all important characteristics of the criminal act and of the relevant mitigating or accentuating circumstances. The importance of every factor could be given a predetermined weight. Lombroso and Beccaria might be equally satisfied. This is not Utopia. This is the system that has been worked out by Gottfredson, Wilkins and Hoffman (1978) for the federal Parole Boards in the United States in connection with decisions on parole for prisoners who have been given indeterminate sentences. The system is in full use. The system has several great advantages. It has enormous capacity. It can include as many factors as we want. It is reliable. Equal factors carry equal weight in all decisions. When correctly instructed, the computer will always give equal cases equal treatment, quite independently of the number of factors taken into consideration. The system might also be characterized as the most democratic, in the 53 sense that it is the legislators and not the administration who decide the moment of discharge. Wilkins can ask the legislators or the central deciding body to determine exactly what relative weight should be attached to every conceivable factor to be considered, such as type of crime, extent of injury to the victim, bed wetting as a child, level of education, risk of recidivism, or conduct in prison. The law may, for example, impose two months' additional sentence for each year of education beyond the normal level (the prisoner should have known better!) or, if desired, two months' remission for each year of higher education (the more highly educated suffer more from punishment). The system also offers the best possibility for administrative control. One can read off in a minute the increase of prison inmates it would lead to if serious drug crimes were upgraded by x points. The system of the calculating machine is also closely related to the ideas of general prevention. It can preach its gospel not only to judges, but to the entire population. In a few years most industrialized nations will be able to receive details of flight and train timetables, restaurant menus and prices direct on the television screen. We will be able to press the buttons and get the answer in a moment. It will be even simpler, since fairly stable charges will be involved, to connect up standard penalties for every type of crime carried out under every kind of.circumstances by every conceivable type of perpetrator. This will be a truly rational form of prevention. Ask your own home-computer and you will get the exact answer to the question what a contemplated breach of the law will cost. But there are also problems. First and foremost, the computer is perfect, infallible. When it is correctly programmed, its decisions are obvious. After guilt was decided, nobody would need to attend before the judge to listen to his decisions if they themselves had some mini-computers at their disposal. This means that chance is taken away from cour-decisions. In civil cases, 54 this would lead to a situation where nobody used the courts. The outcome would be known; why take the trouble? If ideals demand that the courts should be used, a certain degree of uncertainty - not complete uncertainty, just enough to make it worth while trying - seems to be necessary. Another possibility would be to make an attempt to change the computer-programme. This would be the major strategy in criminal cases where - granted that he was found guilty, and most are - the mini-computer might tell the criminal that the outcome would be highly undesirable. Here a second limitation created by perfection dawns upon us. It has to do with the question of who should have the right to decide on what is put into the computer, and also how much it should count, that is what sort of weight should be given to it. Here we can think of a vast number of alternatives in the decision-making bodies. Decisions could be made by: The United Nations in the General Assembly The United Nations in the Crime Committee Regional bodies such as the European Council or the Union for the Arabic States National Parliaments State Parliaments - such as the Californian legislature Sub-units of politicians - such as the Parole Boards or Law Committees within the legislature A random sample of the population questioned through the telephone or personal interviews A sample from the county, or county representatives A sample from the municipality of the victim or the offender A totality of those close to the victim or the offender Or the decisions could be made by the victim and the offender in cooperation. This list is, as you will see, organized so that the decisions on the content of the-computers - the norms that will decide the outcome - are brought in increasing proximity to those 55 concerned as we go down the list. You will also observe that the moment the proximity is perfect, then the computer is also completely superfluous. In that case, people can talk, directly. It is in the United Nations end of the scale that computers are unavoidable. In other words, whether or not the computer is a good and necessary thing within penal law depends on the character of the decision-making system. At the same time it is clear that the very existence of the computers represents a temptation and probably also a pressure towards giving higher priority towards those types of decision-making systems that can make efficient use of computers. Those who dislike such systems will be negatively inclined towards the use of computers within this area. This leads to a third, and maybe the major, problem with computers within penal law. It is not only the parties that do not need to go to court, since their mini-computers could tell them the outcome in advance. But the judge does not need them there either in cases where guilt is clearly established. Why should he? Every category to be considered for sentencing is strictly defined in advance. If he is given the necessary information that makes it possible for him to fill in the correct number in the category, he has no need to see the criminal. Since the categories are agreed on as relevant, and are known in advance, the judge could simplify his task by asking. the parties to deliver written information on the relevant points and order his secretary to clear up any possible disagreements on relevant information before he started the process of sentencing, that is, before he touched the button on the computer for the final answer. The computer in penal law has through these elements double ability to create distance. Decisions on relevance - the computer rules - can be made very far away from the parties concerned. And when they are applied, the ,parties need not be there. Decisions on pain can thus be made in complete insulation from those who are to receive sentence. 56 Here will be no distraction caused by sorrow and tears, by sweat and swear. It will be more like a bureaucracy. Documents, clean desks and - better than in any bureaucracy - clear answers. It will be those answers needed to allow society to remain stable. The principles in the answer will have been decided by people far above. The concrete answer will be exactly similar to the answer given to all criminals belonging to the same category. And the answer is clearly initiated by the criminal himself. The judge has no other responsibility than to push the button. A fourth point with computers has to do with their hidden curriculum. Their hidden message is that conflic;:ts are there to be solved. Computers are calculating machines, they are designed to give answers. But is it so obvious that answers are what is called for? Is it the final outcome that is of primary importance in criminal proceedings, or is it the process? I will come back to this in the next Chapter. Training in law is training in simplification. It is a trained incapacity to look at all values in a situation, and instead to select only the legally relevant ones, that is, those defined by the high priests within the system to be the relevant ones. Neo-classicism is just a logical extension of that whole process of elimination. So few elements of the totality are considered that complete equality is guaranteed. But it is, through its simplifications, a primitive system·. Computers open the way to new, complex possibilities. But now when a technical tool for perfection has been created, we are able to see more clearly that complete clarity. predictability, and pre-programmed behaviour suited for administrative control can never be the only ideals for any legal system. Neoclassicism is maybe an oversimplified attempt to reach a goal that never was. Maybe law is closer related to art than most of us are aware of. But art and power do often stand in a strained relationship.

1. Can computers be used to determine criminal sentences? If so, who ought to write the computer's code? What would Nils Christie think of this kind of computer use?