Act and Rule Utilitarianism

 

Introduction:

            Adopting utilitarianism to produce determinate good or right actions seems to end in failure.  If an ethical theory is supposed to be rational, utilitarianism may fail if we can show that none of the versions of the theory can be consistently and exclusively held, or if any version of the theory tends to put the principle of utility in the background and uses some other principle(s) to determine one’s actions. It seems, unfortunately for the utilitarian, that this may be what happens when utilitarianism (in any of its forms) is adopted as a theory of action or a theory of the judgment of action.

            J.L. Mackie points out that it is sometimes not clear, given any utilitarian position, whether we should distribute the good (whatever that may be) evenly, or whether we should distribute it on the basis of some kind of merit.[1]  And even after considering this question, it is not clear whether we should distribute the greatest happiness to the greatest number, or if it would be better to distribute a greater balance of happiness to a small number who would then receive the greatest benefit, a benefit that would not accrue if a large number of people were to be given a small amount of good.[2]

            Mackie also mentions that it is not clear, on utilitarian grounds, whether we are to apply the principle of utility to the production of good for all sentient creatures, or only for humans.  Further, it is not clear whether we should act according to the rule of the greatest good for the greatest number and at the same time consider the effect of our actions on future generations, or if we should act only for the present generation.[3] 

 

 

 

Fundamental Disagreements Between Act and Rule Utilitarians:

            J.J.C. Smart[4] distinguishes between act and rule utilitarianism by referring to them as "extreme" and "restricted" utilitarianism, respectively.  He claims that the extreme (act) version of utilitarianism will use the rules of morality as rules of thumb only.  Those rules can be changed or overridden at any time in the interest of the production of the greatest good or greatest happiness.[5]  In general, Smart claims that act utilitarianism is superior to rule utilitarianism at least because the act utilitarian recognizes the principle of utility as the overriding principle.  The act utilitarian recognizes that even though it is not possible in every instance to be sure that the determination of a course of action is not biased in favor of one who acts, and it may not be the case in every instance that we have time to weigh the utility of an action, it would be absurd to claim that if we had time and if we were sure that our determination of action was not biased in favor of ourselves, and we were sure that an action against a rule was the right one, that we should not perform an action that does not conform to a rule.  He claims, in short, that to act on a rule when we have had the chance to weigh the benefits of acting against it is to treat the rule superstitiously, as though it were an object of worship, is a most irrational and un-philosophical thing to do.[6]

            M.A. Kaplan[7], however, claims that it cannot be correct to prefer act utilitarianism over rule utilitarianism because it would amount to the destruction of important social relationships (such as the family).  In order to justify this claim, Kaplan distinguishes between instrumental relationships and diffuse relationships[8] such that there are some human relationships that would collapse if the act-utilitarian's principles were used as principles of action.  He gives the example of a husband and wife who discuss whether it would be to the husband's greatest benefit to leave his wife if was ill and had little chance of regaining her health.  On the basis of such a blatantly insensitive and (seemingly) immoral action, Kaplan concludes that the consequences of weighing the benefits are to destroy the family, which is not a relationship of simple instrumental value. This means, specifically, that a diffuse relationship is one in which we expect to be supported in all matters and activities, so the family cannot be considered an instrumental relationship.  An instrumental relationship exists, for example, in a business or social setting in which certain people are given specific tasks that they must perform, but cannot and ought not to expect support outside that specific framework.[9] 

            In short, Kaplan claims "for the most desirable standards to be observed, men must have confidence in each other's behavior."[10]  Kaplan seems to imply here that rules are logical reasons for the choice of actions, and that they are logical reasons because they will tend to produce the greatest happiness for the greatest number.

            But Smart can provide an answer to this criticism.  Smart makes a distinction between logical rules and guides for action.  He claims that "rule(s) (do) not give us a reason for acting so much as an indication of the probable actions of others, which help us to find out what would be our own most rational course of action. . . . The rule 'keep to the left hand side (of the road),' then, is not a logical reason for the action but an anthropological datum for planning actions."[11]  If Smart is correct, the rule utilitarian is acting irrationally in following rules in every single case when it is the case that not following the rules in some specific case would actually produce the greatest benefit.  For Smart, the only rational course of action is to allow us to override the rules in the interest of the greatest utility.

            I think that Smart would also be able to answer Kaplan's charges by noting that we are not all utilitarians in any case, and since there are times when we cannot act on the notion that all others are thinking like us, there will be times when rules of action are not conducive to the general welfare.  Smart's example is of a case in which one lives in a society of non-conformist anarchist drivers who know that the rule of the road is to drive on the left, but who insistently and consistently drive on the right, contrary to the rule.  If I were to act on the rule to drive on the left, even though I know that everyone else is driving on the right, I am not acting in the interest of the greatest utility by following the rule (and I would therefore not be acting rationally).  This shows that rules allow us to predict the behavior of others, but they are not necessarily logical reasons for action.

 

Other Criticisms of Rule Utilitarianism:

            The disagreement to be found between Smart and Kaplan is not all there is to the problems inherent in utilitarianism.  Extreme (act) utilitarianism can be criticized on the ground that it lacks a principle of justice, that it cannot provide an acceptable answer to the problem of punishment.[12]  Tziporah Kasachkoff claims that the problems with act utilitarianism in this case are that 1) it can justify punishment of the innocent, 2) it can justify severe or unfair punishment for a crime, and 3) it can justify deception about cases in which punishment is justified[13], where any of these unjust actions are consistent with the principle of utility.[14]  The act utilitarian cannot answer that conflicts between utility and justice are only apparent or that they are separate issues.  Kasachkoff argues that these responses to the charge are insufficient since 1) not all just acts produce happiness, and not all acts that produce happiness are just, and 2) even though it is true that unjust punishment is allowed under act utilitarian principles, the latitude allowed for unjust punishment is unclear.  Thus, act utilitarianism lacks a notion of the moral significance of punishment according to desert, and not according to consequences.[15]

            The rule utilitarian fares no better when the issue of punishment is referred to that of justice.  We see the problems clearly in the example of the sheriff who must make a decision whether to turn over an innocent man to an angry mob who believes he is guilty, or protect him from them and suffer the effects of a resulting riot.  For the act utilitarian, we must admit that the sheriff should turn the man over, because the benefit of avoiding a riot and injury to the innocent far outweigh the problem of turning an innocent man over to a mob.  So much for act utilitarianism and the problem of justified punishment.

            Kasachkoff[16] claims that no form of rule utilitarianism will be successful in balancing the principle of utility and the requirements of justice, regardless of the rule utilitarian’s (possible) confident reply that the problem can be solved.  The rule utilitarian may attempt to solve the problem by referring to general rule utilitarianism and ideal rule utilitarianism.

            First, 'general'[17] rule utilitarianism will require that the sheriff determine what would happen if all sheriffs were to turn over the innocent to a mob.  He will conclude that the harm that would result from people believing that by force they can have their way will far outweigh the harm that would come from a riot. Second, the ideal rule utilitarian will claim that there is some ideal rule that determines the rightness of all actions, thus the solution to the sheriff's problem is clear.  Third, the ideal utilitarian could refer to intrinsic value in the determination of the right course of action, and, supposedly, this would solve the problem.

            Kasachkoff doubts that any version of utilitarianism can solve the problem.  First, the 'general' rule utilitarian makes a mistake in determining the outcome of the generalization principle involved in making a decision concerning what would happen if all sheriffs gave up their prisoners.  The mistake lies in the rule utilitarian making the specific case into a more general case than is warranted by the situation.  If the generalization were made relevant to the specific example, one would find that it would be right to turn the prisoner over, because the conclusion will be reached, on relevant grounds, from a consideration of all sheriffs in the exact same circumstance, not simply the general action of all sheriffs turning over their prisoners.  The rule utilitarian, in this case, ends up in a situation no better than that of the act utilitarian.

            Second, the ideal rule utilitarian really has no solution to the problem at all.  An ideal rule applies to all cases at all times, and to appeal to the principle for any particular case is to remove oneself from the sphere in which ideal rule utilitarianism has any application.  In other words, we cannot appeal to the ideal rule utilitarian for the solution to any particular case since this type of rule utilitarianism is not utilitarianism.  It fails as a strictly utilitarian principle.  It relies on the notion of ideal rules and their utility, not the application of the principle of utility to the rightness or wrongness of actions.

            Third, and finally, the ideal utilitarian fails because he cannot make a choice between intrinsic goods and evils produced by two alternative actions. If the sheriff is an ideal utilitarian and considers only intrinsic goods and evils, he will find that he is at a stalemate in the determination of right action.  If he chooses to turn the man over to the mob, an innocent man will be punished (an intrinsic evil); and if he does not, many other innocent people will die (an intrinsic evil).  If he does not turn the man over to the mob, an innocent man will not be punished (an intrinsic good); and if he does turn him over to the mob, many other innocent people will not be harmed (an intrinsic good).  The problem here, in short, is that ideal utilitarianism does not offer us a method of making a choice between these intrinsic goods and evils, and leaves us in the same position again (that of not being able to make a choice).

            William Frankena gives a much more general account of the problems with act and rule utilitarianism.[18]  Act utilitarianism can justify seemingly immoral acts and rule utilitarianism cannot distinguish between justice and utility.

            We can imagine, first, a case in which a particular action, A, produces 100 units of good or pleasure.[19]  We can also imagine another action, B, which produces the same amount of good or pleasure, but involves telling a lie.  A consistent act utilitarian cannot make a choice between the two, because the notion of a rule not to lie is independent of, and can be overridden by, the maximization of utility.  Thus, a seemingly immoral action (under ordinary notions of morality) can be made to seem justifiable (or in this case, it can be ignored since both actions have the same net amount of utility).

            A second case is one in which the act utilitarian must choose between two actions, again A and B, but in A, 100 units of good are produced while no lie is involved. In act B, 105 units of good are produced, but a lie is involved.  The consistent act utilitarian must now choose B (which involves telling a lie), which most clearly does involve the claim that a seemingly immoral action is consistent with the demands of utility.

            Similar examples and criticisms can be lodged against rule utilitarianism.  The most important example that Frankena gives is that of two rules, Rule 1 and Rule 2, both of which cannot be made part of our ethical system.  Rule 1 distributes 100 units of good to 5 people (for example)[20] who do not merit it, but the application of Rule 2 distributes 100 units of good to 100 people.  Since the rule utilitarian may be acting under the assumption that he must produce the greatest good for the greatest number, he will choose Rule 2.  Rule 2 is preferable to him because it is distributing the good more widely.  According to Frankena, this makes the rule utilitarian base his ethical theory not on the principle of utility, but possibly on a principle of equality.  But does this not turn the rule utilitarian into an act utilitarian, since he must now decide between the two rules, not on the basis of these specific rules, but on the basis of the rule that we must maximize utility in general?  He must, in short, choose between two rules, and the choice itself amounts to act utilitarianism.[21]

            It seems, at least on the basis of the preceding claims, that utilitarianism in either of its forms cannot meet the challenges demanded by the requirements of a rational or consistent ethical theory.  It would fail in consistency if it continues either to resolve itself (as rule utilitarianism) back into act utilitarianism, and it would fail if it did not consistently use the principle of utility as a principle of maximization regarding particular actions.  Further, it would not stand the test of rationality if it required adherence to rules that are not logical rules, and it would not stand the test of rationality if it amounted to a theory that ignored the requirements of justice when justice is recognized as one of the basic elements of (social) morality.



    [1]J.L. Mackie, Ethics: Inventing Right and Wrong, (New York: Penguin Books, 1977), p. 127.  This notion will be explored more fully in the next section.

    [2]This concept will be explained more fully in the next section.

    [3]Mackie, p. 127.

    [4]J.J.C. Smart, "Extreme and Restricted Utilitarianism" in Utilitarianism with Critical Essays, ed. Samuel Gorovitz, (Indianapolis: The Bobbs-Merrill Company, 1971), pp. 195-203.

    [5]Smart, p. 195.  I use 'good' and 'happiness' interchangeably.

    [6]Smart, p. 199.

    [7]M.A. Kaplan, "Some Problems of the Extreme Utilitarian Position", in Utilitarianism with Critical Essays, ed. Samuel Gorovitz, (Indianapolis: The Bobbs-Merrill Co., 1971), pp. 273-278.

    [8]Kaplan, p. 275.

    [9]Kaplan, p. 274 n 3.  See also R.B. Brandt's claim that moral rules are not institutional rules in "Some Merits of One Form of Rule Utilitarianism" in Introduction to Ethical Theory, ed. Kenneth F. Rogerson, (Fort Worth: Holt, Rinehart, and Winston), p. 88.  I am not sure that there is a direct connection between the concept of instrumental relationships and institutional rules, or of diffuse relationships and moral rules, as Brandt defines moral rules, but there does seem to be some relationship.  It might be said that the family is like a larger society containing moral rules, but they do not, as rules, imply specific duties on the part of the members (in the same way that Brandt claims that moral rules do not amount to rules of an institution).  Further, instrumental rules are rules for the production of a specific result, and appear only within institutions, in which case instrumental rules are subordinate to moral rules.

            Not to belabor the point, but this sounds rather similar, and quite closely related, to Searle's distinction between brute facts and institutional facts.  See, for example, his article, "How To Derive 'Ought' from 'Is'", reprinted in Phillipa Foot, Theories of Ethics.

    [10]Kaplan, p. 277.

    [11]Smart, p. 202.

    [12]Tziporah Kasachkoff, "Utilitarianism and Justice", in Kenneth F. Rogerson, Introduction to Ethical Theory, (Fort Worth: Holt, Rinehart, and Winston, 1991), p. 92.

    [13]That is, act utilitarianism appears to be capable of the justification of deception regarding whether a punishment determined to be appropriate has been carried out.  Let us suppose a case in which an innocent person has been sentenced to some specific punishment (death) in order to provide for the greatest happiness.  The claim that the utilitarian can be deceptive regarding punishment would be illustrated by a case in which an innocent person is condemned, the public is told that he has received his punishment (death), and the greatest good or happiness has been produced even though the person is set free, and is given a new name and identity by the government so that the deception will never be discovered.  If this is the case, then Kasachkoff is right when she claims that the act utilitarian lacks two essential principles of justice:  "(1) recognition of guilt as necessary for the justification of punishment, and (2) the requirement that the punishment fit the crime, i.e. be proportional to the gravity of the offense" (p. 93).  But I think that Kasachkoff misses something here, and that is that she does not make more reference to an example like that above, and does not explain clearly that there is something wrong, utilitarianism aside, with convicting an innocent person and then deceiving the public.  Not only must the punishment fit the crime, but the charge that a government or social system may find deception to be of the greatest utility is in itself seemingly immoral regardless of the utility in it.

    [14]Kasachkoff, pp. 92-93.

    [15]Kasachkoff, p. 94.

    [16]Kasachkoff, pp. 95-101.

    [17]Kasachkoff does not give this form of rule utilitarianism any other name than "rule utilitarianism," so I am identifying it as "general rule utilitarianism" for the sake of clarity.

    [18]William Frankena, Ethics, (Englewood Cliffs, NJ: Prentice Hall, 1963), pp. 32-33.

    [19]Frankena notes, and I must emphasize, that there is no way in which to assign a numerical value to the production of happiness or good.  These examples simply serve to clarify the problems inherent in the utilitarian's attempt to justify an action and his attempt to account for a principle of justice or of just distribution.

    [20]Frankena does not give numbers of people who will be affected by the application of the rule.  I have added these numbers to make my explanation simpler.

    [21]One point that Frankena does not discuss, but that I find relevant to the case, is the notion of the interpretation of the rule "to produce the greatest happiness (or good) for the greatest number."  Could it not be the case that the rule utilitarian could choose R! as the optimific rule because it distributes the greatest happiness to the greatest number in this case?  That is, it could be said that giving 5 people 20 units of good each is preferable to giving 1 unit of good each to 100 people since the 5 people receiving 20 'units' would be able to make better use of it, or be more happy, than 100 people would be collectively?  The case would be similar to determining whether it would be better to give 5 starving people 20 dollars each, or if giving 100 starving people 1 dollar each would be better, considering that the 100 people will continue to starve in any case since $1 cannot possibly feed them sufficiently, while giving $20 each to five people will make a difference in their conditions.