“Utilitarianism” can most generally be described as the doctrine that states that the rightness or wrongness of actions is determined by the goodness and badness of their consequences. This general definition can be made more precise in various ways, according to which we get various species of utilitarianism.
Act and Rule Utilitarianism
The first important division is between “act” utilitarianism and “rule” utilitarianism. If, in the above definition, we understand actions to mean “particular actions,” then we are dealing with the form of utilitarianism called act utilitarianism, according to which we assess the rightness or wrongness of each individual action directly by its consequences.
If, on the other hand, we understand actions in the above definition to mean “sorts of actions,” then we get some sort of rule utilitarianism. The rule utilitarian does not consider the consequences of each particular action but considers the consequences of adopting some general rule, such as “Keep promises.” He adopts the rule if the consequences of its general adoption are better than those of the adoption of some alternative rule.
Since, in this context, the word rule can be interpreted in two ways, to mean either “possible rule” or “rule actually operating in society,” there are actually two species of rule utilitarianism.
If we interpret rule simply as “possible rule,” we get an ethical doctrine strongly resembling that of Immanuel Kant. It is true that Kant is not normally regarded as a utilitarian, but nevertheless a utilitarian strain can be detected in his thought.
If we interpret his categorical imperative, “Act only on that maxim through which you can at the same time will that it should become a universal law,” as meaning “Act only on that maxim which you would like to see established as a universal law,” and if liking here is determined by the individual’s feelings as a benevolent man, then we get a version of utilitarianism which may usefully be called Kantianism. It is true that Kant would object to this appeal to feelings of benevolence and would wish to distinguish sharply between willing and “wanting or liking.
Nevertheless, it is far from clear how Kant’s distinction can be defended; and when he elucidates his general principle by means of examples, he does indeed tend to think in terms of the consequences that we should like to see brought about.
However, the word Kantianism is used here merely as a useful and perhaps not inappropriate label; whether Kant himself would approve of its present application is not an important issue in the present discussion.
If, in our definition of utilitarianism, we interpret the word rule as “actual rule,” or “rule conventionally operative in society,” we get a form of rule utilitarianism that has been propounded in recent times by Stephen Toulmin, who seems mainly concerned with the justification, and in some cases the reform, of rules of conduct that are actually operative in society.
When we think of the writers with whom the term utilitarianism is most naturally associated, namely, Jeremy Bentham, J. S. Mill, and Henry Sidgwick, we must think of utilitarianism primarily as act utilitarianism.
However, controversy has developed over whether Mill should not rather be interpreted as a rule utilitarian, and there has also been much discussion of the rival claims of act and rule utilitarianism to be viable ethical theories.
R.M.Hare, in his book Freedom and Reason (Oxford, 1963), has recently argued that there is no clear distinction between act and rule utilitarianism, since if a certain action is right, it must be the case that any action just like it in relevant respects will also be right. If these respects are then specified in detail, we get a rule of the form “Do actions of this sort.”
A defender of the distinction between act and rule utilitarianism could reply that since the situations in which actions occur are infinitely variable, and since no two actions have quite the same sorts of consequences, the act utilitarian may not be able to describe the “relevant respects” mentioned above in any less general form than “The action is of the sort that has the best consequences.”
But if this is so, Hare’s principle that if an action is right then any action which is like it in the relevant respects is also right does not yield a sufficiently particular form of rule to justify the assimilation of act and rule utilitarianism.
Egoistic and Universalistic Utilitarianism
Act utilitarianism, unlike rule utilitarianism, lends itself to being interpreted either in an egoistic or in a nonegoistic way. Are the good consequences that must be considered by an agent the consequences to the agent himself (his own happiness, for example), or are they the consequences to all humankind or even to all sentient beings? If we adopt the former alternative, we get egoistic utilitarianism; and if we adopt the latter alternative, we get universalistic utilitarianism.
Since what is best for me is unlikely to be what is best for everyone, it is clear that there is not only a theoretical but also a practical incompatibility between egoistic and universalistic utilitarianism. This was not always seen by the early utilitarians, who sometimes seem to have confused the two doctrines.
There is, in fact, even a pragmatic inconsistency in egoistic utilitarianism, since an egoist, on his own principles, would be unlikely to wish to be seen in his true colors, and so would have no motive for expressing his ethical doctrine. In this entry we shall be concerned with utilitarianism in the universalistic sense.
Hedonistic and Ideal Utilitarianism
Another distinction, which cuts across that between act and rule utilitarianism, is the distinction between hedonistic and ideal utilitarianism. Utilitarianism has been defined above as the view that the rightness or wrongness of an action depends on the total goodness or badness of its consequences.
A hedonistic utilitarian will hold that the goodness or badness of a consequence depends only on its pleasantness or unpleasantness. As Bentham put it, quantity of pleasure being equal, pushpin is as good as poetry.
An ideal utilitarian, such as G. E.Moore, will hold that the goodness or badness of a state of consciousness can depend on things other than its pleasantness. According to him, the goodness or badness of a state of consciousness can depend, for example, on various intellectual and aesthetic qualities.
In his calculations, the ideal utilitarian will be concerned not only with pleasantness and unpleasantness, but also with such things as knowledge and the contemplation of beautiful objects. He may even hold that some pleasant states of mind can be intrinsically bad, and some unpleasant ones intrinsically good.
|J. S. Mill|
J. S. Mill took up an intermediate position. He held that although pleasantness was a necessary condition for goodness, the intrinsic goodness of a state of mind could depend on things other than its pleasantness, or, as he put it, there are higher and lower pleasures.
It should be noted that we have assumed that the only things that can be intrinsically good or bad are states of consciousness. Other things can of course be extrinsically good or bad. For example, an earthquake is normally extrinsically bad, that is, it causes a state of affairs that is on the whole intrinsically bad.
Moreover, a utilitarian can hold that something that is intrinsically bad, such as the annoyance of remembering that we have forgotten to do something, is extrinsically good, for it is a means to a set of consequences that are on balance intrinsically good.
|G. E. Moore|
G. E. Moore held that states of affairs other than states of consciousness could be intrinsically good or bad. For an ideal utilitarian, this is a theoretically possible contention, but nevertheless, few ideal utilitarians would find the contention a plausible one, and we shall therefore ignore it in this article.
Normative and Descriptive Utilitarianism
Utilitarianism may be put forward either as a system of normative ethics, that is, as a proposal about how we ought to think about conduct, or it may be put forward as a system of descriptive ethics, that is, an analysis of how we do think about conduct.
The distinction between normative and descriptive utilitarianism has not always been observed. It is important to bear carefully in mind the distinction between normative and descriptive utilitarianism and to note that objections to descriptive utilitarianism do not necessarily constitute objections to normative utilitarianism.
Properly speaking, utilitarianism began with Jeremy Bentham (1748–1832), who was a universalistic hedonistic act utilitarian. He put forward his view essentially as normative ethics, but he was unclear about the distinction between normative and factual utterances and may justly be accused of committing what Moore later called the naturalistic fallacy—the fallacy of claiming to deduce ethical principles solely from matters of fact. (David Hume had in effect pointed out this fallacy before Bentham’s time.)
PRECURSORS OF UTILITARIANISM. Anticipations of Bentham are to be found in the history of ethics. In ancient times Aristippus of Cyrene and Epicurus propounded hedonistic theories. However, their doctrines approximate egoistic rather than universalistic utilitarianism, despite the fact that they were unclear about the difficulty of reconciling the two doctrines and hence tried to have it both ways.
The same might be said of Abraham Tucker and William Paley, the more immediate precursors of Bentham, who also injected certain theological conceptions into their systems. The tension between egoistic and universalistic hedonism can also be detected in the eighteenth-century French writer Claude-Adrien Helvétius, who appears to have influenced Bentham; also, the political philosopher William Godwin should be mentioned.
David Hume is often classified as a utilitarian, but he used utility not as a normative or even as a descriptive principle, but as an explanatory one: When asked why we approve of certain traits of character, he would point out that they are traits which either are useful or are immediately agreeable.
Both because he used the principle of utility in an explanatory way and because he was primarily concerned with the evaluation of traits of character (virtues and vices and the like) rather than with the question of what actions ought to be done, it is not advisable to regard Hume as a utilitarian.
J. S. MILL. As was mentioned above, there has been some controversy over whether J. S. Mill (1806–1873) ought to be regarded as an act utilitarian or as a rule utilitarian. Mill does not make his position on this issue very clear.
Probably he was not very well aware of the distinction, and in any case he would probably have thought it a fairly unimportant one, since he was mainly concerned with the opposition between utilitarianism in general and other systems of ethics that were quite nonutilitarian.
Although Bentham had on at least one occasion used the word utilitarian, it was Mill who introduced it into philosophy. He appropriated it, with some change of meaning, from a passage in the Scottish novelist John Galt’s Annals of the Parish (Edinburgh, 1821).
SIDGWICK. We can with some confidence classify Mill as a normative utilitarian rather than a descriptive one, but the first utilitarian philosopher who was very explicit on this issue was Henry Sidgwick (1838–1900).
Sidgwick understood that there is a distinction between normative and factual sentences, although, like G. E. Moore (1873–1958), he thought that ethical principles could be the objects of intellectual intuition. Sidgwick was a universalistic hedonistic utilitarian, but he was also strongly attracted by the claims of egoism.
He saw more clearly than earlier writers that there was a theoretical inconsistency in being both an egoistic and a universalistic utilitarian, and he considered the possibility that there might be theological sanctions that would reconcile the two views, if not in theory, then at least in practice.
LATER UTILITARIANS. Moore and Hastings Rashdall were ideal universalistic utilitarians, although Moore, with his principle of organic unities, and Rashdall, with his importation into the utilitarian calculations of the moral worth of the actions themselves, introduced considerations which, if taken seriously, would seem to vitiate the truly utilitarian character of their theories.
A subtle form of rule utilitarianism of the sort we have called Kantianism was propounded in 1936 by R. F. Harrod. Contemporary writers such as Stephen Toulmin, P.H. Nowell-Smith, John Rawls, K. E.M. Baier, and M. G. Singer have propounded views that either are or approximate rule utilitarianism. R. B. Brandt has been sympathetic to rule utilitarianism and has recently defended a rather subtle and complex version of it.
Analysis and Critique
UTILITARIANISM AS A DESCRIPTIVE ETHICS. It is fairly easy to show that both act utilitarianism and rule utilitarianism are inconsistent with usual ideas about ethics, or what can be called the common moral consciousness. For the principles of both systems will in some cases lead us to advocate courses of action that the plain man would regard as wrong.
Consider, for example, the case of a secret promise to a dying man. To ease his dying moments, I promise him that I will deliver a hoard of money, which he entrusts to me, to a rich and profligate relative of his.
|Analysis and Critique|
No one else knows either about the promise or the hoard. On utilitarian principles, it would appear that I should not carry out my promise. I can surely put the money to much better use by giving it, say, to a needy hospital. In this way I would do a lot of good and no harm. I do not disappoint the man to whom I made the promise, because he is dead.
Nor, by breaking the promise, do I do indirect harm by weakening men’s faith in the socially useful institution of promise making and promise keeping, for on this occasion no one knows about the promise. Normally, of course, an act utilitarian will keep a promise even when the direct results are not beneficial, because the indirect effects of sowing mistrust are so harmful.
This consideration clearly does not apply in the present instance. The plain man, however, would be quite sure that the promise to the dying man should be kept. In this instance, therefore, we have a clear case in which utilitarianism is inconsistent with the way in which, for the most part, people in fact think about morality.
The rule utilitarian, on the other hand, would probably agree with the plain man in the above case, because he would appeal to the utility of the rule of promise keeping in general, not to the utility of the particular act of promise keeping.
Nevertheless, cases can be brought up that will show the incompatibility of even rule utilitarianism with the common moral consciousness. For example, a riot involving hundreds of deaths may be averted only by punishing some innocent scapegoat and calling it punishment.
Given certain empirical assumptions, which may perhaps not in fact be true, but which in a certain sort of society might be true, it is hard to see how a rule utilitarian could object to such a practice of punishing the innocent in these circumstances, and yet most people would regard such a practice as unjust.
They would hold that a practice of sometimes punishing the innocent would be wrong, despite the fact that in certain circumstances its consequences would be good or that the consequences of any alternative practice would be bad.
In this instance, then, there is a conflict between even the rule utilitarian and the plain man. (This is not, of course, to say that in fact, in the world as it is, the rule utilitarian will be in favor of a practice of punishing the innocent, but it can be shown that in a certain sort of world he would have to be.)
ACT UTILITARIANISM AS A SYSTEM OF NORMATIVE ETHICS. Both act and rule utilitarianism fail, then, as systems of descriptive ethics. But act utilitarianism as a system of normative ethics would seem to have certain advantages over both rule utilitarianism and nonutilitarian, or deontological, systems of ethics (a deontological system of ethics is one that holds that an action can be right or wrong in itself, quite apart from consequences).
Moreover, the failure of act utilitarianism as a descriptive system is the source of its interest as a possible normative system: If it had been correct as a descriptive system, then the acceptance of it as a normative system would have left most men’s conduct unchanged.
No proof of utilitarianism. A system of normative ethics cannot be proved intellectually. Any such “proof” of utilitarianism as was attempted by Bentham or Mill can be shown to be fallacious. (Mill disclaimed the possibility of proof and spoke more vaguely of “considerations capable of determining the intellect,” but he presented an attempted proof nonetheless.)
Sidgwick and Moore were clearer on this point and saw that ethical principles cannot be deduced from anything else. They appealed instead to intellectual intuition, but recent developments in epistemology and other fields of philosophy have made the notion of intellectual intuition a disreputable one.
|C. L. Stevenson|
The tendency among some more recent writers, such as C. L. Stevenson, R. M. Hare, and P. H. Nowell-Smith, has been to regard assertions of ultimate ethical principles and valuations as expressions of feeling or attitude, or as akin to imperatives rather than to statements of fact. In this respect, they develop further the position held much earlier by Hume.
Now if we abandon a cognitivist theory about the nature of moral judgments, such as was held by Sidgwick or Moore, and adopt the view that ultimate ethical principles depend only on our attitudes, that is, on what we like or dislike, we must give up the attempt to prove any ethical system, including the act-utilitarian system.We may nevertheless recommend such a system. We may also try to show inconsistencies or emotionally unattractive features of various possible alternative systems.
Appeal to generalized benevolence. In putting forward act utilitarianism as a normative system, we express an attitude of generalized benevolence and appeal to a similar attitude in our audience. (The attitude of generalized benevolence is not the same as altruism.
Generalized benevolence is self-regarding, and other-regarding too—I count my happiness neither more nor less than yours.) Of course, we all have in addition other attitudes, self-love, and particular likes and dislikes.
As far as self-love is concerned, either this will be compatible with generalized benevolence or it will not. If the former, then self-love does not conflict with act utilitarianism, and if the latter, nevertheless self-love then will be largely canceled out, as among a number of people engaged in discussion.
Arguments against deontological systems. As to particular likes and dislikes, an important case concerns our liking for obeying the rules of some deontological ethics in which we have been raised. However, the following persuasive considerations can be brought up as arguments against the adherent of a deontological system of ethics.
It can be urged that although the dictates of a generalized benevolence might quite often coincide with those of an act-utilitarian ethics, there must be cases in which the two would conflict with one another.
Would the benevolent and sympathetic persons to whom we conceive ourselves to be appealing be happy about preferring abstract conformity with an ethical rule, such as “Keep promises,” to preventing avoidable misery of his fellow creatures?
It will be noticed that the above defense of utilitarianism against deontology is purely persuasive, an appeal to the heart and not to the intellect. It is based on the metaethical view that ultimate ethical principles are expressions of our attitudes and not the findings of some sort of intuition of ethical fact.
|W. D. Ross|
An intellectualist in metaethics, such as W. D. Ross, could well resist our appeal to feeling by saying that it is possible to see that his deontological principles are correct, and that whether we like them or not is beside the point.
Weakness of rule utilitarianism. In defending act utilitarianism, then, we appeal to feelings, namely, those of generalized benevolence. Since people possess other attitudes too, such as loyalty to a code of morals in which they have brought up, the possession of feelings of generalized benevolence is not a sufficient condition of agreement with the act utilitarian.
But it is a necessary condition. Now the rule utilitarian also appeals ultimately to feelings of generalized benevolence. Like the deontologist, however, he is open to the charge of preferring conformity with a rule to the prevention of unhappiness.
He is indeed more obviously open to such a charge, since he presumably advocates his rule utilitarian principle because he thinks that these rules conduce to human happiness. He is then inconsistent if he prescribes that we should obey a rule (even a generally beneficial rule) in those cases in which he knows that it will not be most beneficial to obey it.
It will not do to reply that in most cases it is most beneficial to obey the rule. It is still true that in some cases it is not most beneficial to obey the rule, and if we are solely concerned with beneficence, in these cases we ought not to obey the rule.
Nor is it relevant that it may be better that everybody should obey the rule than that nobody should. That the rule should always be obeyed and that it should never be obeyed are not the only two possibilities. There is the third possibility that sometimes it should be obeyed and sometimes it should not be obeyed.
Hedonistic act utilitarianism.We shall therefore neglect rule utilitarianism as a system of normative ethics, and consider only act utilitarianism, which will be conveniently put forward in a hedonistic form. The reader will easily be able to adapt most of what is said to cover the case of ideal utilitarianism.
Indeed, in many cases the differences between hedonistic and ideal utilitarianism are not usually of much practical importance, since the hedonist will usually agree that the states of mind the ideal utilitarian regards as intrinsically good, but which he does not, are nevertheless extrinsically good.
Bentham would say that Mill’s higher pleasures, if not intrinsically better than the lower ones, are usually more “fecund” of further pleasures. This is not to say, however, that there are no cases in which there would not be a significant difference between hedonistic and ideal utilitarianism.
The act-utilitarian principle can now be put in the following form: “The only reason for performing some action A, rather than various alternative actions, is that A results in more happiness (or more generally, in better consequences) for all humankind (or perhaps all sentient beings) than will any of these alternative actions.”
Since this principle expresses an attitude of generalized benevolence, we can expect to find a good deal of sympathy for it among the sort of people with whom it would be profitable to carry on a discussion about ethics.
It may therefore be possible to obtain wide assent to the principle, provided that we can develop its implications in a clear and consistent manner and that we can show that certain common objections to utilitarianism are not as valid as they are supposed to be.
We have already seen that certain objections, based on “the common moral consciousness,” fail because they are valid only against descriptive utilitarianism and not against normative utilitarianism.
Determining consequences. Utilitarianism would be an easier doctrine to state if we could assume that we could always tell with certainty what all the consequences of various possible actions would be, and if we could assume that very remote consequences need not be taken into account.
In applying the utilitarian principle, we would simply have to envisage two or more sets of consequences extending into the future, and ask ourselves, as sympathetic and benevolent men, which of these we would prefer.
There would be no need for any calculation or for any summation of pleasures.We would simply have to compare two or more possible total situations. Sometimes, indeed, the postulate that we need not consider very remote situations will not be necessary.
For example, if it be admitted that, on the whole, people are more happy than not, a man and woman who are left alive as sole representatives of the human race after some atomic holocaust could, as utilitarians, decide to have children in the hope that the world would once more be populated indefinitely far into the future.
This is because although the generations will extend indefinitely far into the future, there is reason to believe that each generation will be happy rather than unhappy, while if no children are had, there will be no succeeding generations at all, and so no possibility of happiness accruing in the future. In normal cases, however, we do need to assume that remote consequences can be left out of account.
Surely, however, this is a plausible assumption, for on the whole, the goodness and badness of very remote consequences are likely to cancel out. In any case, if this assumption cannot be made, also difficulties will arise for many deontological systems (for example, the system of W. D. Ross), which allow beneficence as one principle among others.
Unfortunately, however, we do not know with certainty what the various possible consequences of our actions will be. This uncertainty would not be so bad provided we could assign numerical probabilities to the various consequences.We could then still employ a method similar to that of envisaging total consequences. A very simplified example may make this clear.
Suppose that the only relevant consequences are, on the one hand, a 3/5 probability of Smith’s being in some state S, and on the other hand, if we do an alternative action, a 2/7 probability of Jones’s being in some state T. We simply envisage 21 people just like Smith in state S as against 10 people just like Jones in state T.
It should be evident how, in theory at least, this method could be extended to more complex cases. However, numerical probabilities can rarely be assigned to possible future events, and the utilitarian is reduced to an intuitive weighting of various consequences with their probabilities.
It is impossible to justify such intuitions rationally, and we have here a serious weakness in utilitarianism. It is true that this weakness also extends to prudential decisions, and most people think that they can make prudential decisions with some rationality.
But this is not of much help, since in propounding a normative system we are concerned with what we ought to think, not with how we do think. Utilitarianism is therefore badly in need of support from a theory whereby, at least roughly or in principle, numerical probabilities could be assigned to all types of events.
|rules of thumb|
THE PLACE OF RULES IN ACT UTILITARIANISM. Even the act utilitarian cannot always be weighing up consequences. He must often act habitually or in accordance with rough rules of thumb. However, this does not affect the value of the act-utilitarian principle, which is put forward as a criterion of rational choice.
When we act habitually we do not exercise a rational choice, and the utilitarian criterion is not operative. It is, of course, operative when we are deciding, on act-utilitarian principles, the habits or rules of thumb to which we should or should not school ourselves.
The act utilitarian knows that he would go mad if he deliberated on every trivial issue, and that if he did not go mad he would at least slow up his responses so much that he would miss many opportunities for probably doing good. He may also school himself to act habitually because he may think that if he deliberated in various concrete situations, his reasoning would be distorted by a selfish bias.