Does rule utilitarianism give an adequate explanation of why we should not punish innocent people?

One can think of rule utilitarianism as a system of unmitigated statism and legal jurisdiction with certain rules and limitations attached to it. If punishment can be shown to promote effectively the interest of society it is justifiable on the utilitarian point of view, otherwise it is not justified.

On the utilitarian point of view, however, the extent to which the state and the courts are willing to prosecute determines the degree of guilt involved. If the punishment of an innocent person can be proved — utilitarianism proved? — to promote greater happiness or less pain in society overall, then it should be permitted. So the concept of what it means to be innocent in this system of instrumental jurisprudence is wholly dependent on the outcomes it would produce.

But the utilitarians are far more clever than that. Only Bentham was so naive as to suggest this kind of unmitigated hedonism. Utilitarians now have incorporated checks and balances into their theories of punishment. As John Rawls points out in his essay Two Concepts of Rules, “utilitarians agree that punishment is to be inflicted only for the violation of law.” The utilitarian, accordingly, seeks to limit the extent of punishment by a system of rules.

Yet aren’t utilitarians themselves creating the rules? The rules themselves ought to also be created on a utilitarian calculus, and this would put us in the same position as before. Seeing as it might be the case that punishment of an innocent person could lead to greater happiness, why create a rule that would inhibit this? Rawls says that utilitarians hold that “no official should have discretionary power to inflict penalties whenever he thinks it for the benefit of society.”

On this account, Rawls believes he has reconciled the retributive view with the utilitarian view. Judges decide cases base on retributive forms of justice. Legislators create rules based on a utilitarian form of justice. And the utilitarian justice has primacy.

The problem Rawls is concerned about is whether utilitarianism in general, as the legislator’s prerogative, justifies too much. It would, for example, justify institutions that by their very nature condone the condemnation of the innocent. It is, as Rawls says, an “engine of justification”. He wants to make sure, however, that there are clear rules set in place, and clear rules of recognition specifically, so that legislators know who has the power to decide such matters and by what authority, for all of this “must be written into the rules of the institution”.

What Rawls calls “telishment” is the unjustified punishment of innocent people for the greater happiness of the rest of the society. Rawls says we should come to doubt our justice system if it had ever been recognized as one that telishes it citizens rather than punishes them for offenses they are guilty of. He says that it is highly unlikely, given that we would feel sorry for innocent people, that we would allow a system of telishment to go on.

Rawls ends on the idea that the system of utilitarian punishment is justified, and that it would be unlikely to end in a form of “telishment”, based on his own ideas of what a proper sensibility in society would be like. He says there would be “less danger” of any of that happening because we would give requirements to the arbitrary forces in government not to act in an unmitigated utilitarian way.

It appears, however, that the ultimate justification for a system like this depends on the credibility and the ‘ascertainability’ of the requirements themselves. The requirements cannot be chosen on a utilitarian calculus, because then we would have justified the telish operations we are trying to avoid. The project here rests on something signified that is outside the justifications that utilitarianism provides, and for this reason it cannot be said to be a utilitarian justification. It relies on something else, something like a set of rules based on a principle that would simply not allow some acts to be permitted by government.

This would not appear to be a consequentalist view at all. And a retributivist view, as Rawls has pointed out himself, is a view that the judges hold, not a view that the policymakers or legislators hold. The view of the legislators and policymakers, then, cannot be utilitarian. It must be some form of Kantian or rights-based point of view, that cannot be persuaded by the utilitarian calculus.