In an ideal legal system, the costs of determining guilt or innocence would be held close to the minimum costs of gathering information and determining its veracity to some acceptable level of probability — “beyond a reasonable doubt” in the case of guilt, and to whatever level of probability would socially justify dismissing charges or discontinuing the investigation if the defendant or suspect appeared to be innocent. Since these costs are positive — indeed, substantial — even an ideally functioning legal system would not wholly eliminate crime, but there would be some optimal quantity of crime191 based on costs of knowledge, costs of precautionary measures, and the inconveniences imposed on innocent parties as a result of rules, arrangements, investigations, and suspicions incident to crime-prevention or crime-detection. While the concept of an “optimal” quantity of crime may be uncomfortable, it is also clear that no one is prepared to devote half the Gross National Product to stamping out every residual trace of gambling. Nor are we even prepared to reduce the murder rate at all cost — when that would mean such stringent administration of homicide laws and such low levels of proof required for conviction as to cause some physicians to avoid accepting some or all patients who might die while under their care. There would be no social gain from allowing thousands to perish needlessly for lack of timely medical care, in order to reduce murders by one hundred. Obviously, no one would advocate going to such extremes regarding gambling, murder, or any other crimes, but the point here is to indicate the reasons why — reasons that apply, to some degree, across a much wider range of situations.
In crime control, as in other social processes, decisions and evaluations must be incremental rather than categorical. It is pointless to argue that this or that action will or will not
Changes in the criminal law change the effectiveness with which knowledge can be transmitted to those deciding innocence or guilt, to criminals contemplating crime, and to the voting public assessing their experience and assessing the protection offered — or not offered — by the criminal justice system.
There are many sources of knowledge, and the behavior of legal authorities puts a higher or lower cost on its transmission or effectiveness. The simple knowledge that a crime has been committed can vary in its availability to the criminal justice system according to the costs imposed on victims, witnesses, or informants. The costs of reporting rape can obviously be increased or decreased substantially by the way police respond to rape victims, by the way opposing attorneys are permitted to cross-examine the victim in court, and by the likelihood that a convicted rapist will be either turned loose soon (perhaps to retaliate against the plaintiff or witnesses) or given a retrial on a technicality. In the landmark