Handelingsutilisme versus regelutilisme

Een aantal utilisten meent dat we niet zo zeer het maatschappelijk nut van handelingen als wel van handelingsregels moeten vergroten. Hoewel het idee al ouder is is Brandt de eerste die het idee van regelutilisme meer systematisch heeft uitgewerkt. Een voordeel van regelutilisme is dat het beter recht kan doen aan bepaalde morele verplichtingen en aan morele rechten (zie ook utilisme en rechten van mensen). Harsanyi meent daarom dat regelutilisme beter dan handelingsutilisme in staat is de wenselijkheid van bepaalde sociale instituties en regels te beoordelen. Deze benadering is volgens hem ook in staat aan te geven wanneer uitzonderingen op de regel toelaatbaar zijn
Hoewel veel moderne utilisten regelutilisten zijn, zijn er ook utilisten die zich verzetten tegen die vorm van utilisme. Een voorbeeld is Smart die meent dat regelutilisme, net als vormen van deontologische ethiek, leidt tot regel fetisjisme: regels die gevolg worden om de regels.
Hare is van mening dat de tegenstelling tussen regel- en handelingsutilisme verdwijnt als we ons realiseren dat er twee niveaus of situaties van morele deliberatie zijn. In de meeste dagelijkse situaties is er onvoldoende tijd of gelegenheid om na te gaan welke handeling het meeste maatschappelijke nut oplevert; we kunnen dan beter regels volgen; in meer uitzonderlijke situaties waar de gewone regels niet meer voldoen is een vorm van handelingsutilisme op zijn plaats.

R.B. Brandt

R.B. Brandt

R.B. Brandt, "Some Merits of One Form of Rule Utilitarianism"

For convenience I shall refer to [my] theory as the "ideal moral code" theory. The essence of it is as follows: Let us first say that a moral code is "ideal" if its currency in a particular society would produce at least as much good per person (the total divided by the number of persons) as the currency of any other moral code. (Two different codes might meet this condition, but, in order to avoid complicated formulations, the following discussion will ignore this possibility.) Given this stipulation for the meaning of "ideal," the Ideal Moral Code theory consists in the assertion of the following thesis: An act is right if and only if it would not be prohibited by the moral code ideal for the society; and an agent is morally blameworthy (praiseworthy) for an act if, and to the degree that, the moral code ideal in that society would condemn (praise ) him for it. ...

For a moral code to have currency in a society, two things must be true. First, a high proportion of the adults in the society must subscribe to the moral principles, or have the ,moral opinions, constitutive of the code. Exactly how high the proportion should be, we can hardly decide on the basis of the ordinary meaning of "the moral code"; but probably it would not be wrong to require at least 90 percent agreement. ...

Second, we want to say that certain principles A, B, etc., belong to the moral code of a society only if they are recognized as such. That is, it must be that a large proportion of the adults of the society would respond correctly if asked, with respect to A and B, whether most members of the society subscribed to them. ...

We must now give more attention to the conception of an ideal moral code, and how it may be decided when a given moral code will produce as much good per person as any other. We may, however, reasonably bypass the familiar problems of judgments of comparative utilities, especially when different persons are involved, since these problems are faced by all moral theories that have any plausibility. We shall simply assume that rough judgments of this sort are made and can be justified.

(a) We should first notice that, as "currency" has been explained above, a moral code could not be current in a society if it were too complex to be learned or applied. We may therefore confine our consideration to codes simple enough to be absorbed by human beings, roughly in the way in which people learn actual moral codes. ...
(b) In deciding how much good the currency of a specific moral system would do, we consider the institutional setting as it is, as part of the situation. We are asking which moral code would produce the most good in the long run in this setting. One good to be reckoned, of course, might be that the currency of a given moral code would tend to change the institutional system.
(c) In deciding which moral code will produce the most per person good, we must take into account the probability that certain types of situation will arise in the society. For instance, we must take for granted that people will make promises and subsequently want to break them, that people will sometimes assault other persons in order to achieve their own ends, that people will be in distress and need the assistance of others, and so on.. ..
(d) It would be a great oversimplification if, in assessing the comparative utility of various codes, we confined ourselves merely to counting the benefits of people doing (refraining from doing) certain things, as a result of subscribing to a certain code. To consider only this would be as absurd as estimating the utility of some feature of a legal system by attending only to the utility of people behaving in the way the law aims to make them behave-and overlooking the fact that the law only reduces and does not eliminate misbehavior, as well as the disutility of punishment to the convicted, and the cost of the administration of criminal law. In the case of morals, we must weigh the benefit of the improvement in behavior as a result of the restriction built into conscience, against the cost of the restriction--the burden of guilt feelings, the effects of the training process, etc. ...

It has been thought that the implications of rule-utilitarianisms for two types of situation are especially significant: (a) for situations in which persons are generally violating the recognized moral code, or some feature of it: and (b) for situations in which, because the moral code is generally respected, maximum utility would be produced by violation of the code by the agent. An example of the former situation (sometimes called a “state of nature” situation) would be widespread perjury in making out income tax declarations. An example of the latter situation would be widespread conformity to the rule forbidding walking on the grass in the park.

What are the implications of the suggested form of rule-utilitarianism for these types of situation? Will it prescribe conduct which is not utility maximizing these situations? If it does, it will clearly have implications discrepant with those of act-utilitarianism—but perhaps unpalatable to some people.

It is easy to see how to go about determining what is right or wrong in such situations, on the above described form of rule-utilitarianism—it is a question of what an “ideal” moral code would prescribe. But it is by no means easy to see where a reasonable person would come out, after going through such an investigation….

Far from “collapsing” into act-utilitarianism, the Ideal Moral Code theory appears to avoid the serious objections which have been levelled at direct [act] utilitarianism. One objection to the latter view is that it implies that various immoral actions (murdering one’s elderly father, breaking solemn promises) are right or even obligatory if only they can be kept secret. The Ideal Moral Code theory has no such implication. For it obviously would not maximize utility to have a moral code which condoned secret murders or breaches of promise. W.D. Ross criticized act-utilitarianism on the ground that it ignored the personal relations important in ordinary morality, and he listed a half-dozen types of moral rule which he thought captured the main themes of thoughtful morality: obligations of fidelity, obligations of gratitude, obligations to make restitution for injuries, obligations to help other persons, to avoid injuring them, to improve one's self, and to bring about a just distribution of good things in life. An ideal moral code, however, would presumably contain substantially such rules in any society, doubtless not precisely as Ross stated them. So the rule-utilitarian need not fail to recognize the personal character of morality. ...

The Ideal Moral Code theory has the advantage of implying that the moral rules recognized in a given society are not necessarily morally binding. They are binding only in so far as they maximize welfare, as contrasted with other possible moral rules. Thus if, in a given society, it is thought wrong to work on the Sabbath, to perform socially desirable abortions, or to commit suicide, it does not follow, on the Ideal Moral Code theory, that these things are necessarily wrong. The question is whether a code containing such prohibitions would maximize welfare. Similarly, according to this theory, a person may act wrongly in doing certain things which are condoned by his society. ...

Let us examine the implications of the Ideal Moral Code theory by considering a typical example. Among the Hopi Indians, a child is not expected to care for his father (he is always in a different clan), whereas he is expected to care for his mother, maternal aunt, and maternal uncle, and so on up the female line (all in the same clan). It would be agreed by observers that this system does not work very well. The trouble with it is that the lines of institutional obligation and the lines of natural affection do not coincide, and, as a result, an elderly male is apt not to be cared for by anyone.
Can we show that an "ideal moral code" would call on a young person to take care of his maternal uncle, in a sys- tem of this sort? (It might also imply he should try to change the system, but that is another point.) One important feature of the situation of the young man considering whether he should care for his maternal uncle is that, the situation including the expectations of others being what it is, if he does nothing to relieve the distress of his maternal uncle, it is probable that it will not be relieved. His situation is very like that of the sole observer of an automobile accident; he is a mere innocent bystander, but the fact is that if he does nothing, the injured persons will die. So the question for us is whether an ideal moral code will contain a rule that, if someone is in a position where he can relieve serious distress, and where it is known that in all probability it will not be relieved if he does not do so, he should relieve the distress. The answer seems to be that it will contain such a rule: we might call it an "obligation of humanity." But there is a second, and more important point. Failure of the young person to provide for his maternal uncle would be a case of unfairness or free riding. For the family system operates like a system of insurance; it provides one with various sorts of privileges or protections, in return for which one is expected to make certain payments, or accept the risk of making certain payments. Our young man has already benefited by the system, and stands to benefit further; he has received care and education as a child, and later on his own problems of illness and old age will be provided for. On the other hand, the old man, who has (we assume) paid such premiums as the system calls on him to pay in life, is now properly expecting, in accordance with the system, certain services from a particular person whom the system designates as the one to take care of him. Will the ideal moral code require such a person to pay the premium in such a system? I suggest that it will, and we can call the rule in question an “obligation of fairness.” So, we may infer that our young man will have a moral obligation to care for his maternal uncle, on grounds both of humanity and fairness.

John C. Harsanyi

John C. Harsanyi

J.C. Harsanyi, "Morality and the theory of rational behaviour" in Sen, A. and Williams, B. (eds) (1982) Utilitarianism and Beyond, Cambridge: Cambridge University Press, pp. 58-59.

For example, consider the problem of keeping promises. Traditional morality says that promises should be kept, with a possible exception of cases where keeping a promise would impose excessive hardship on the promise maker (or perhaps third persons). In contrast, act utilitarianism would make the breaking of a promise morally permissible whenever this would yield a slightly higher social utility – perhaps because of unexpected changes in the circumstances – than keeping of the promise would yield. But this would greatly reduce the social benefits associated with the making of promises as an institution. It would make it rather uncertain in most cases whether any given promise would be kept. People would be less able to form definite expectations about each other’s future behaviour and would have a general feeling of insecurity about the future. Moreover, this uncertainty would greatly reduce their incentives to engage in various socially very useful activities on the expectation that promises given to them would be kept. (For instance, they would become much less willing to perform useful services for other people for promised future rewards.)

As compared with act utilitarianism, rule utilitarianism will be much closer to traditional morality in maintaining that promises should be kept, subject only to rather rare exceptions. An act utilitarian always asks the question, 'Would this one act of possible promise breaking increase or decrease social utility?' In contrast, a rule utilitarian has to ask, 'What particular moral rule for promise keeping would maximise social utility?'
As a result, an act utilitarian can consider the socially unfavourable effects of promise breaking only to the extent that these have the nature of causal consequences of individual acts of promise breaking. No doubt, one act of promise breaking already will somewhat reduce people's trust in promises, but normally this effect will be quite small. In contrast, a rule utilitarian can also consider the causal consequences of a general practice of repeated promise breaking. But, more importantly, he can also consider the noncausal logical implications of adopting a moral rule permitting many easy exceptions to promise keeping.
More particularly, he will always have to ask the question, 'What would be the social implications of adopting a moral rule permitting that promises should be broken under conditions A, B, C, etc. -assuming that all members of the society would know that promise breaking would be permitted under these conditions?' Thus he will always have to balance the possible direct benefits of promise breaking in some specific situations against the unfavourable expectation and incentive effects that would arise if people knew in advance that in these situations promises would not be kept. In other words, rule utilitarianism not only enables us to make a rational choice among alternative possible general rules for defining morally desirable behaviour. Rather, it also provides a rational test for determining the exceptions to be permitted from these rules.


J.J.C. Smart

J.J.C. Smart


J.J.C. Smart, "An outline of a system of utilitarian ethics" in Smart, J.J. and Williams, B. (1973) Utilitarianism For and Against, Cambridge: Cambridge University Press, pp. 10-12

I have argued elsewhere the objections to rule-utilitarianism as compared with act-utilitarianism. Briefly they boil down to the accusation of rule worship: the rule-utilitarian presumably advocates his principle because he is ultimately concerned with human happiness: why then should he advocate abiding by a rule when he knows that it will not in the present case be most beneficial to abide by it? The reply that in most cases it is most beneficial to abide by the rule seems irrelevant. And so is the reply that it would be better that everybody should abide by the rule than that nobody should. This is to suppose that the only alternative to 'everybody does A' is 'no one does A'. But clearly we have the possibility 'some people do A and some don't'. Hence to refuse to break a generally beneficial rule in those cases in which it is not most beneficial to obey it seems irrational and to be a case of rule worship.

The type of utilitarianism which I shall advocate will, then, be act-utilitarianism, not rule-utilitarianism.
David Lyons has recently argued that rule-utilitarianism (by which, I think, he means the sort of rule-utilitarianism which I have called the Kantian one) collapses into act-utilitarianism. His reasons are briefly as follows. Suppose that an exception to a rule R produces the best possible consequences. Then this is evidence that the rule R should be modified so as to allow this exception. Thus we get a new rule of the form ‘do R except in circumstances of the sort C’. That is, whatever would lead the act-utilitarian to break a rule would lead the Kantian rule-utilitarian to modify the rule. Thus an adequate rule-utilitarianism would be extensionally equivalent to act-utilitarianism.


I am inclined to think that an adequate rule-utilitarianism would not only be extensionally equivalent to the act-utilitarian principle (i.e. would enjoin the same set of actions as it) but would in fact consist of one rule only, the act-utilitarian one: ’maximize probable benefit’. This is because any rule, which can be formulated must be able to deal with an indefinite number of unforeseen types of contingency. No rule, short of the act-utilitarian one, can therefore be safely regarded as extensionally equivalent to the act-utilitarian principles unless it is that very principle itself.

R.M. Hare

R.M. Hare

R.M. Hare, Moral Thinking, Its Level, Method and Point, Oxford: Clarendon Press, 1981, p. 43

Much of the controversy about act-utilitarianism and rule-utilitarianism has been conducted in terms which ignore the difference between the critical and intuitive levels of moral thinking. Once the levels are distinguished, a form of utilitarianism becomes available which combines the merits of both varieties (H 1972a:I3 ff.; 1976a). The conformity (for the most part) to received opinion which rule-utilitarianism is designed to provide is provided by the prima facie principles used at the intuitive level; but critical moral thinking, which selects these principles and adjudicates between them in cases of conflict, is act- utilitarian in that, in considering cases, actual or hypothetical, it can be completely specific, leaving out no feature of an act that could be alleged to be relevant. But since, although quite specific, it takes no cognizance of individual identities, it is also rule-utilitarian in that version of the rule-utilitarian doctrine which allows its rules to be of unlimited specificity, and which therefore is in effect not distinguishable from act-utilitarianism (FR 7.6; H 1972C:I70; Lyons, 1965:ch.iii). The two kinds of utilitarianism, therefore, can coexist at their respective levels; the critical thinker considers cases in an act- utilitarian or specific rule-utilitarian way, and on the basis of these he selects, as I shall shortly be explaining, general prima facie principles for use, in a general rule-utilitarian way, at the intuitive level.