Chapter 2, part 4, of the author's Mill on Liberty, which Clarendon Press published in 1980. It has been included in the Victorian Web with the kind permission of the author and of the Clarendon Press, which retains copyright.
- Numbers in brackets indicate page breaks in the print edition and thus allow users of VW to cite or locate the original page numbers.
- Where possible, bibliographical information appears in the form of in-text citations, which refer to the bibliography at the end of each document.
- Superscript numbers link only to documents containing substantialbibliographical information; the numbers do not form a complete sequence.
- Non-bibliographic notes appears as text links.
- This web version is a project supported by the University Scholars Programme of the National University of Singapore. Scanning, basic HTML conversion, and proofreading were carried out by Gerhard Rolletschek, a Postgraduate Visiting Scholar from the University of Munich, working under the direction of George P. Landow, who added links to materials in VW.
n a recent paper entitled "John Stuart Mill and the Limits of State Action" Richard Wollheim gives a subtle and comprehensive account of Mill's notion of self-regarding actions13. According to Wollheim, self-regarding actions are "those actions which affect either the agent alone or other people solely insofar as they believe such actions to be right or wrong" (p. 9). Although, as we shall see at the end of this section, Wollheim omits some acts that Mill would wish to include in the self-regarding category, his account of what Mill means by such acts is substantially correct. However, he fails in his ingenious attempt to reconcile his interpretation of self-regarding actions with utilitarianism. He does not succeed in showing that Millian liberalism, as embodied in the claim that there are never good reasons for prohibiting self-regarding actions, is compatible with utilitarianism.
Wollheim points out that at first sight there may appear to be a conflict between his account of self-regarding conduct and the utilitarian doctrine. This is because, on Wollheim's interpretation, Mill is committed to disregarding as irrelevant the pain or distress of outraged moral and religious sensibilities. But a utilitarian must surely take into account any pain produced by an act. Rawls has given a vivid illustration. Suppose that the majority in a society has such an intense abhorrence for certain religious or sexual practices that even the very thought that these practices are going on in private and out of their view is enough to arouse anger and hatred.
Seeking the greatest satisfaction of desire may, then, justify harsh repressive measures against actions that cause no social injury. To defend individual liberty in this case the utilitarian has to show that given the circumstances the real balance of advantages in the long run still lies on the side of freedom; and this argument may or may not be successful. [Rawls, p. 450]
However, Wollheim uses two arguments to show that the utilitarian is ultimately justified in ignoring the pain caused solely by the belief that an action is wrong. If these two arguments succeed, then the utilitarian does not have to go in for the kind of balancing of pleasures and pains that Rawls suggests.
I shall begin with the second argument. Wollheim maintains [19/20] that Mill has two criteria for distinguishing a preference from a moral belief. First, a preference is based on feelings and emotions. It is not supported by reasons, and according to Wollheim, Mill was only prepared to accept reasons which appealed to the consequences of the act. The second criterion is that a preference is personal in the sense that it is not about how others ought to behave, but only about the individual's own conduct.
Wollheim points out that many, though not all, of the beliefs in question are mere preferences according to the first criterion. The beliefs we are considering are those which condemn actions as wrong even though these actions cause no harm or pain to others independently of their being thought wrong. Since the actions cause no independent harm, no reason of the relevant utilitarian kind can be given to show that they are wrong. Hence the belief that they are wrong must be a preference by the first criterion. We cannot, for example, say that a self-regarding action is wrong because it pains others, for the pain does not exist independently of the belief that the conduct is wrong, and hence cannot be cited in support of it. However, Wollheim shows that there are some beliefs of the type in question which are not preferences. In these cases the actions condemned by the beliefs also cause no independent harm, but it is genuinely believed that they cause such harm, and this sincere, though mistaken, belief is the reason for the condemnation. Wollheim gives the example of someone who believes that smoking marijuana causes impotence, and this belief is his reason for supporting the legal penalties against marijuana smoking. His belief, then, is not a preference according to the first criterion, because the belief is supported by a reason which appeals to the consequences of the act.
The next step in Wollheim's argument is to show that for Mill the two criteria for preferences are linked in that if a belief is a preference by the first criterion, then it will also satisfy the second criterion. He argues that in Utilitarianism Mill holds the view that a judgement can only be considered a moral judgement if it is based on an appeal to consequences. If, in justifying his act, an agent is not prepared to give a reason which invokes the consequences of the act, then his [20/21] judgement on the act is not a moral judgement. But the agent who cannot back his judgement with relevant reasons has merely stated his preference, according to the first criterion of a preference. But, at the same time, the judgement is also not a moral belief about what all should or ought to do in similar situations. So what is the nature of the judgement? Wollheim. concludes that Mill would seem to have only one answer: it is a judgement of personal taste or inclination about what the agent himself should do, and hence it satisfies the second criterion of a preference. If this linkage between the two criteria for preferences is granted, then the rest of Wollheim's argument can proceed.
Suppose X is a self-regarding action which "causes" pain to another person simply because he believes X to be wrong. Since this belief is not supported by a utilitarian reason, it is a preference according to the first criterion. But, because of the linkage between the two criteria for preferences, the belief will also satisfy the second criterion for a preference. From this, it follows that the belief cannot be about the conduct of others. For according to the second criterion for a preference, a preference only refers to what the person who has the preference would like to be or to do. So X cannot really violate his preference since X is the conduct of another agent. The pain that the person experiences when he sees, or comtemplates, someone else's performance of X cannot therefore be attributed to X. He is mistaken in thinking that X is the cause of his pain. Since X causes no other pain, there is no utilitarian reason for interfering with it.
The person who thinks that X is the cause of his pain in fact mistakes his preference for a moral belief. For example, he thinks that he holds the moral belief that homosexuality is wrong, and that is why the homosexual acts of others distress him. In fact his belief, being a preference, relates only to his own conduct. Perhaps he believes that a homosexual life is not the life he would like to lead. Wollheim suggests that the association in his mind between his pain and the homosexual acts arises very likely from his own desires and fears about homosexuality. Perhaps he himself both desires to indulge in homosexual acts, and is afraid to do so.
However, Wollheim's attempt to link the two criteria for [21/22] preferences fails. The linkage exists only when an agent is making a judgement about his own conduct. But of course we make judgements not only when we are agents, but also when we are spectators. In the case of a judgement about other people's conduct, from the fact that it is not a moral judgement but a preference according to the first criterion, it cannot be inferred that it is also a preference according to the second criterion. For my judgement may be personal in one of Wollheim's senses, in the sense that it stems from my own feelings and emotions, and yet not be personal in the required sense that it relates only to my conduct. My feelings and emotions may be related to the conduct of others. So the link between the two criteria for preferences is present only in the case of a person's judgements on his own conduct. But in the context of self-regarding conduct, it is our judgements on other people's self-regarding conduct, and not on our own, that raise a problem for the utilitarian.
Certainly Wollheim is right in drawing attention to the importance for Mill of the distinction between a preference and a moral belief. But there is no evidence that Mill accepted the second criterion for a preference. In a long footnote to his edition of his father's Analysis of the Phenomena of the Human Mind, Mill states the distinction between a moral belief and a preference:
This strong association of the idea of punishment and the desire for its infliction, with the idea of the act which has hurt us, is not in itself a moral sentiment, but it appears to me to be the element which is present when we have the feelings of obligation and of injury, and which mainly distinguishes them from simple distaste and dislike for anything in the conduct of another that is disagreeable to us, that distinguishes, for instance, our feeling towards the person who steals our goods, from our feeling towards him who offends our senses by smoking tobacco." (My emphasis.) [Mill's Ethical Writings]
Here is an explicit statement that a preference, or a "simple distaste or dislike", can be about "the conduct of another". such as his smoking tobacco. The same view is also expressed in Mill's essay on Bentham, where he speaks of liking or disliking other people's conduct which "neither does good nor harm", as opposed to having the moral sentiments of [22/23] approval or disapproval [Mill's Essays, p. 284]. Again there is not the slightest hint that preferences are to be restricted in scope to the conduct of the person who has the preference.
With these two passages from Mill's other works in mind, we can return to the essay On Liberty and see whether there is anything there which suggests a different account of the nature of preferences. But, curiously, the only passage cited by Wollheim which bears on the scope of a. preference goes against his account. For in it, as Wollheim himself acknowledges, Mill refers to the existence of preferences about what others should do: specifically, Mill mentions a religious bigot's distaste for the religious beliefs and practices of others [Wollheim, p. 22]." Wollheim treats this remark as an aberration from Mill's main characterization of preferences as limited in scope to the conduct of the person having the preference. But the belief that preferences can be about the conduct of others is, as we have seen, very much in line with Mill's comments elsewhere, and far from being an isolated remark, it seems to be part of his considered view of the nature of preferences.
But if preferences can refer to the conduct of others, then Wollheim's argument collapses. For the admission that a belief is a preference does not imply that another person's conduct will not violate it. So a self-regarding act can violate other people's preferences, and can therefore cause them pain. Without further argument a utilitarian would not be justified in placing it completely outside the scope of social intervention.
I now turn to Wollheim's first argument. This can be divided into two parts. The first part of the argument is that if the only effect an action has on others is through their belief that the action is wrong, then the belief in question must be false. This is because the utilitarian calculation of the pleasures and pains caused by an action must be made as if in a world prior to the adoption of moral attitudes. Otherwise one would have to include the pleasures and pains caused by the action through various moral beliefs. And if these pleasures and pains are included, then the rightness or wrongness of an action will, to some extent, be determined [23/24] by whether people feel or believe ihe action to be right or wrong. If, for example, people feel strongly enough that a particular action is wrong, and they are severely distressed by the mere contemplation of the action, then these pains may be bad enough to make the action wrong. So if one counts the pleasures and pains caused by various moral beliefs, then it is possible for an action to be right or wrong simply because a sufficient number of people feel strongly enough that the action is right or wrong. But this is absurd. If one now disregards the pain which arises simply from the belief that an action is wrong, then a self-regarding action has no painful consequences, and hence cannot be wrong on utilitarian grounds. The belief that it is wrong must therefore be a false belief.
Having established that the beliefs in question are false, Wollheim enters the second part of his argument by asking: why should the utilitarian Mill ignore the pain caused by the false beliefs about an action? After all, these pains are no less real than the pains arising from true beliefs. Wollheim says that Mill was perhaps inclined simply to disregard the pain caused through false beliefs in the same way that he looked down on the comforts of unreason and error. This inclination is supported by Mill's view that as intellectual inquiry progresses, false moral beliefs will be eliminated.
However, these considerations only show Mill to be a very imperfect utilitarian, and this goes against Wollheim's general thesis. Perhaps Mill is right that false moral beliefs will vanish with the progress of intellectual inquiry, but a good utilitarian will still take into account the pain brought about by false beliefs so long as the pain exists. Of course if he thinks that certain false beliefs will soon disappear he will not give much significance to the distress to which they give rise. But here what discounts the distress is not the falsity of the beliefs, but the short-lived or temporary nature of the distress. And if Mill was disposed bluntly to discount both the pleasures and pains arising from false beliefs as such, then he would have abandoned the utilitarian axiom that all pleasures as such are good, and all pains as such are evil.
However, even though in the second part of his argument Wollheim. does not succeed in explaining why a utilitarian [24/25] would ignore the pain caused through false beliefs, there was no need for him to have embarked on this task. The first part of his argument, if sound, already gives him all that is necessary. For he has already explained why the pain caused by the belief that the act is wrong should be ignored: it is because otherwise an act can be made wrong simply by a sufficient number of people feeling strongly enough that it is wrong. But if any pain so caused may be disregarded, then of course it follows that any pain caused by a false belief about the wrongness of the act may be ignored. No additional argument is needed. However, what is crucial to Wollheim's argument here is not the fact that the belief is false, but rather the fact that the belief is one concerning the morality of the act. As Wollheim maintains, the utilitarian calculation must be made prior to the adoption of moral attitudes. The distress in question is to be ignored not because it stems from a false belief, but rather because it stems from a belief about the morality of the act. In a different case, a person may hold the false belief that someone is going to kill him, and as a result live in constant fear and trembling. The distress caused by this kind of false belief must be taken into account by the utilitarian, and nothing that Wollheim has said shows otherwise. But this is not the kind of distress caused by self-regarding conduct. A self-regarding action causes no relevant pain because the only pain it causes arises from the belief that it is wrong, and this pain, we have seen, is to be ignored. So it does seem that, in most cases, there would be no utilitarian reason for interfering with such actions.
However, I add the qualification 'in most cases' because some self-regarding actions may cause harm to the agent himself, either immediately or in the future. And given that the harm is bad enough, there will be a good utilitarian reason for paternalistic intervention to prevent the agent from harming himself. So further argument is needed to show why the utilitarian would reject paternalism. But this qualification aside, the first part of Wollheim's argument, if sound, would appear to show that there is no utilitarian basis for interfering with self-regarding conduct. But his argument is unsound.
Wollheim maintains that the utilitarian must disregard the [25/26] pleasures and pains which arise simply from the adoption of moral attitudes towards an action, for otherwise the morality of the action will, in certain circumstances, be crucially determined by the mere strength of people's feelings about its rightness or wrongness. But this argument, as elaborated by Wollheim, suffers a little from an ambiguity in the scope of the pleasures and pains that are to be ruled out by the utilitarian. In one place Wollheim refers to "pleasurable and unpleasurable moral responses", and here what are pleasurable and unpleasurable respectively are the favourable and unfavorable moral responses themselves. But Wollheim also refers to a causal connection between the pleasures and pains on the one hand, and the associated moral beliefs on the other hand. And in general he refers to the effects on others of selfregarding actions as effects which proceed only "via certain beliefs they hold" (p. 8), or as effects which are brought about "solely insofar as they believe such actions to be right or wrong" (p. 9). But the extent of the effects on others that self-regarding actions can have will depend on whether or not we are to include among such effects all those caused by the beliefs these people hold about the morality of the actions. For example, it may be part of being a sincere and committed Sabbatarian that one is distressed, to some extent at least, by the licensing of Sunday entertainments. But different Sabbatarians will be distressed to different degreees. Suppose that some are so distressed that they are physically ill. Such illness is causally linked to their beliefs about the wrongness of Sunday entertainments. Is the physical illness therefore to be considered a part of the effects of Sunday entertainments? Or do we, as Wollheim seems to suggest on one occasion (pp. 12-13), confine the relevant effects to the mental anguish that the contemplation of the act causes in those who think it wrong?
However, there are several arguments that the utilitarian may use to try to show that, in spite of the harm in question, self-regarding conduct should never be interfered with. The first argument is to deny that the harm -- the serious illness of the morally sensitive person -- is to be attributed to the self-regarding conduct of others. In other words, it is not the violations of his moral beliefs which give rise to the harm. [26/27] The cause of the harm is the person's own peculiar personality. A normal person, who holds the same moral views, will not be as seriously affected as this sensitive person is by the mere awareness that others are committing immoral acts. The difference must therefore be attributed, not to these acts, but to the person's unusual, and indeed, extreme sensitivity. It is this extreme sensitivity which is the real cause of his illness [Feinberg, pp. 104-5].
But the argument, so far, does not provide the utilitarian with a reason for never prohibiting self-regarding acts. For granted that such acts do not cause the illness, it remains true that the acts are causally relevant in the production of the illness. What the utilitarian needs to show is that such a prohibition is never the most economical or effective way of preventing the illness, or that the cost in utilitarian terms of such prohibition will always outweigh the evil of the illness. These he may or may not be able to show. For example, it may be relatively easy to shield a sensitive person from the knowledge that sexual immoralities are committed every night in his society, or to steel him to face such knowledge with calmness. But if it proves to be impossible to do either, the utilitaxian could still plausibly argue that it is better for the sensitive person to be seriously ill rather than for the state to prohibit all the self-regarding actions which will offend his moral sensibilities. The cost in terms of human misery, and limited police resources, of such prohibitions will considerably outweigh the evil of one person's serious illness. However, all that the argument attempts to show is that the harm is not sufficient to justify the suppression of self-regarding actions. But the harm must always be taken into account in any utilitarian calculation of whether certain self-regarding actions should be permitted.
But I have still not come to grips with Wollheim's argument that "the Utilitarian calculation must be made as if in a world prior to the adoption of moral attitudes." (p. 12.) His argument here fails to make a couple of distinctions. First, it is necessary to distinguish between those pleasures and pains which arise from the adoption of utilitarian moral attitudes, and those pleasures and pains which arise from the adoption of non-utilitarian moral attitudes. Prior to the adoption of any moral attitude, a self-regarding action [27/28] causes no pain to others, and once again, leaving aside the effect on the agent himself, there is no utilitarian basis for believing it to be wrong. Hence if a utilitarian, who is aware of this fact, still regards the action as wrong, and it thereby distressed by it, then there is something irrational about him. But from the non-utilitarian point of view, the action may properly be regarded as wrong even though it has no undesirable consequences independently of its being so regarded. For the ground of the moral judgement that the act is wrong may have nothing to do with the consequences of the act. This point is of some significance because most of those who are likely to be distressed as a result of their beliefs about the wrongness of some self-regarding actions will be non-utilitarians. Now, the distress of non-utilitarians can be treated as something that occurs prior to the utilitarian moral judgement bn a self-regarding action. This being the case, there is no reason why the utilitarian should ignore the distress. The distress occurs prior to the adoption of utilitarian moral attitudes, and therefore constitutes part of the data which the utilitarian has to take into account before he decides on the morality of the action. So even a self-regarding action can be wrong if it causes enough distress to non-utilitarians.
But now, suppose I am wrong about this, and the distress of the non-utilitarian may be ignored by the utilitarian. It follows, then, that no self-regarding act is morally wrong from the utilitarian point of view. But it does not follow that, on utilitarian grounds, it would always be wrong legally to prohibit such acts, or to punish persons for engaging in them. This is because punishing an act X is itself an act that is distinct from X, and may have consequences different from the consequences of X. This distinction is of course analogous to that between the morality of an act and the morality of praising or condemning a person for performing that act, a distinction that utilitarians themselves have emphasized [see J. J. C. Smart, pp. 197-8]. Now armed with this distinction, we can accept Wollheim's claim that "the Utilitarian calculation must be made as if in a world prior to the adoption of moral attitudes", and yet reject his unqualified view that the utilitarian may ignore the distress which arises by way of the belief that a self-regarding act is wrong. For now there are different levels [28/29] from which the distress may be viewed. With respect to the morality of X, we are to ignore the distress which proceeds from the belief that X is wrong. But if the question concerns the morality of punishing or prohibiting X, then the distress produced by the belief that X is wrong is relevant to the utilitarian calculation. For though the distress does not exist prior to the adoption of moral attitudes about X, it exists prior to the adoption of moral attitudes about the punishment or prohibition of X.
I conclude therefore that Wollheim does not succeed in reconciling Mill's defence of self-regarding conduct with utilitarianism. Moreover, any attempt at reconciliation will have to face an additional difficulty because Wollheim's account of self-regarding conduct is defective in at least one respect. He says that if there are actions that produce "immediate revulsion or disgust", then these actions are not selfregarding. This is because, on his account, self-regarding conduct affects others only in so far as they believe the conduct to be wrong. But the immediate revulsion or disgust is independent of any belief about the wrongness of the acts in question. Wollheim is right that the utilitarian would have to include such immediate revulsion and disgust in his calculation. But this places the utilitarian even further away from Mill. For it seems quite obvious that Mill would not regard the revulsion or disgust as a relevant reason for interfering with self-regarding conduct. In the essay he attacks those who tried to regulate the conduct of others simply on the basis of their feelings, their mere likes and dislikes, and he argues against the imposition on others not only of one's opinions or beliefs, but also of one's inclinations. Wollheim's exclusion of conduct which arouses immediate revulsion or disgust has the following odd consequence. A person who is immediately disgusted with another's conduct, and who can give no reason at all for his disgust, can none the less rightly claim that his disgust is to be taken into account by the utilitarian. But, on the other hand, Wollheim would allow the utilitarian to ignore the same person's disgust if it is based on a deeply held, and even clearly articulated, non-utilitarian moral belief about the wrongness of the act.
Feinberg, Joel. " 'Harmless Immoralities' and Offensive Nuisances" Issues in Law and Morality, ed. Norman S. Care / Thomas K. Trelogan. Cleveland and London, 1973.
Rawls, John. A Theory of Justice. Oxford, 1972.
Mill's Ethical Writings, ed. J. B. Schneewind. New York and London, 1965.
Mill's Essays on Literature and Society. ed. J. B. Schneewind. New York and London, 1965.
Smart, J. J. C. "Extreme and Restricted Utilitarianism" Mill: Utilitarianism, with Critical Essays, ed. Samuel Gorovitz. Indianapolis, 1971.
Wollheim, Richard. "John Stuart Mill and the Limits of State Action" Social Research, 40 (1973).
_____ "John Stuart Mill and Isaiah Berlin: The Ends of Life and the Preliminaries of Morality" The Idea of Freedom: Essays in Honour of Isaiah Berlin, ed. Alan Ryan. Oxford, 1979.
Last modified 18 April 2001