Chapter 7, part 1, 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.
- 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.
- indicates a link to material not in the original print version. [GPL].
Mill's rejection of paternalism forms part of his liberty principle, but it is such an important part that he explicitly spells out his anti-paternalism. immediately after stating that the prevention of harm to others is the only legitimate basis of interference with the freedom of individuals.
His own good, either physical or moral, is not a sufficient way-rant. He cannot rightfully be compelled to do or forbear because it will be better for him to do so, because it will make him happier, because, in the opinions of others, to do so would be wise, or even right. [p. 73]
But Mill's anti-paternalism does not extend to "children and persons under age" (On Liberty, p. 137). He also allows us forcibly to prevent a person from crossing an unsafe bridge (p. 152). But here he assumes that the person is unaware of the dangerous condition of the bridge, and that he does not wish to fall into the river. It is his ignorance which justifies a temporary restriction of his freedom. Interference is justified only if there is not time to warn the person about the danger. Mill generalizes by claiming that interference is unjustified unless the person is "a child, or delirious, or in some state of excitement or absorption incompatible with the full use of his reflecting faculty" (p. 152).
In trying to understand the nature and scope of Mill's anti-paternalism, it is useful to bear in mind two different aspects of a person's conduct: the decision-aspect and the consequence-aspect. The former refers to the different ways m which an agent's decision to act in a particular manner is vitiated or impaired, or his consent to certain acts is not "full and free". Thus a person may commit a dangerous act without realizing, or fully appreciating, its danger. This is the case with the person crossing the unsafe bridge. On the other hand, the consequence-aspect refers to the undesirable consequences of a person's act, as for example the fact that the act will harm him or produce other undesired effects. [109/110]
The cases which Mill excludes from the scope of his anti-paternalism are cases in which both the decision -- and the consequence-aspects are present. Thus when Mill refers to children, it is evident that he believes that they lack certain capacities that normal adults have. Children, especially very young children, have not developed sufficiently their emotional and intellectual capacities, and they also lack the degree of knowledge and experience needed for meaningful choices.
An analysis of Mill's arguments and examples will show that he is opposed to strong paternalism but that he favours a degree of weak paternalism [cf. Feinberg, Paternalism]. Strong paternalism maintains that we are justified in interfering to prevent a person from harming himself even when his decision is fully voluntary or totally unimpaired. Weak paternalism is the doctrine that we are justified in interfering to prevent a person from harming himself only when there is a defect in his decision to engage in the self-harming activity. Weak paternalism therefore justifies intervention when both the decision- and consequence aspects are present, whereas strong paternalism justifies intervention by reference to the consequence-aspect alone.
An agent's decision to perform a particular act may be affected by a variety of factors which reduces the significance to be attached to the decision. A brief discussion of some of these factors will help in the understanding of weak paternalism.
First, a person may lack some relevant knowledge. This can take several forms. An agent may do something without being aware of the harmful consequences of his acts, and there may be reason to believe that had he known these consequences, he would not have acted in the way he did. An obvious case of this is the taking of a medicine with harmful side-effects unknown to the patient. But a person may also be lacking in knowledge not with respect to the consequences of his act, but to the nature of the act. An example of this would be a patient whose modesty is outraged by a doctor who pretends that he is giving her some special medical treatment. Lack of knowledge is often, as in this case, the result of fraudulent or misleading representation by others.
Secondly, there is lack of control. Sometimes a person may be aware of the consequences of his act, but because of [110/111] temporary emotional unbalance, may be unable to appreciate the full significance of these consequences, and to exercise rational judgement with respect to them. Emotional unbalance may be caused by grief, distress, or severe strains. There is also another type of factor. Some drugs are severely addictive, and though the initial decision to take them is freely made, the subsequent "choices" to continue taking them are impaired by the agent's addiction.
Thirdly, there is undue influence. This can be exerted in different ways, in some of which there will be disagreement about whether or not undue influence was really present. Coercion is an example of undue influence, but there are also the pressures of economic inducements, and sometimes of customs and traditions. A person who would not otherwise perform an act may do so because of economic inducements. This is common enough, and does not normally raise questions about undue influence. There was, however, an interesting case in which an air-hostess, guilty of a breach of regulations, consented to be caned by the manager of her company as an alternative to dismissal or being grounded, which would have meant the loss of her "flying pay". The manager administered six cuts on her buttocks with a light cane. He was convicted of assault in spite of receiving her written consent (McCoy, cited by Hughes, pp. 683-84). The occasions on which economic inducements may be regarded as impairing a person's decision depend on a number of factors like the degree of reluctance of the consenting victim, the relative bargaining positions of the parties, the relationship between the parties, and the nature of the available alternatives. No sharp line can be drawn between the situations in which undue influence is present and the situations in which it is not. The pressures exerted by customs and traditions are also sometimes of such great strength as to be treated by some as a kind of "undue influence". Thus Glanville Williams observes that the consent to duelling was often given with great unwillingness and solely out of fear of being called a coward. When duelling was legally prohibited, unwilling duellists had an honourable excuse for not fighting, and were thereby freed from "the tyranny of custom" (p. 78).
Persons whose decisions are substantially impaired in one of the ways specified above are likely later on to regret [111/112] their decisions to engage in harmful conduct. Interference will, in retrospect, be accepted and appreciated. But the subsequent consent of the subject is not a necessary condition for the justification of weak paternalism. Suppose, for example, that the person crossing the unsafe bridge condemns our interference. This does not make the interference wrong. What it does is to make any further or subsequent interference wrong. But the original interference is still justified so long as it was not known at the time that the subject would not consent [for discussions of the connection between paternalism and the consent of the subject, see: Carter, Hodson]. Weak paternalism insists on always maintaining some contact with a person's own preferences and values. It promotes each person's own good as defined by that person himself. It is not a cloak for enforcing the values and preferences of the person interfering or society at large.
However, in some cases, because of the nature of the impairment, it is not possible to discover what a person's real preferences and values are. This is true of those special categories of persons children, the mentally subnormal, and the mentally ill who lack the capacities of normal adults. We may not know how particular persons would choose if they ceased to be children, or to be mentally ill, or if they were not mentally subnormal. There is here the danger that interference may be designed to shape and develop their values and preferences in line with those of the interfering party. The subject who is interfered with in this way will come to approve of the intervention; this problem is discussed by: Rawls, pp. 249-50; Murphy, pp. 482-83; Carter, pp. 136-39. In order to reduce the dangers of such manufactured consent, we need something like the Rawlsian device of choosing in a hypothetical state called the original position (Rawls, ch. 3). In Rawl's theory of justice, rational and mutually disinterested persons choose from behind "a veil of ignorance" the principles which are to govern the basic institutions of their society. The veil of ignorance deprives them of knowledge of their particular talents and abilities, their class interests or social positions, their conceptions of the good or their particular moral and religious views, their psychological make-up, and the stage of development of their society. It does not, however, rob them of information about general psychological laws, and general truths about the world, or the nature of human societies, or human nature. Persons in the original position will allow some [112/113] interference with their conduct if they turn out to be children, or adults who lack certain normal capacities.
In adopting the device of the original position for the limited purpose of helping to determine the scope of weak paternalistic intervention in the conduct of special categories of persons, we are not committed to Rawls's ambitious policy of deriving principles of justice from the original position. Rawls applies his theory to the problem of paternalism, see pp. 248-50; see also Richards, pp. 192-5, and Murphy.
There is a much greater likelihood of agreement about interference on weak paternalistic grounds than about the fundamental principles of justice. For example, there will be agreement about the need to interfere to prevent harm, and to protect and develop the capacities for choice between alternative patterns of conduct. But there will also be disagreement, and in particular about how children are to be treated as they grow older.
It is possible that the accumulation of evidence will show that children, except when they are very young, do not lack many of the relevant capacities of normal adults (on this point see Houlgate). If this is correct, then paternalistic intervention in children's conduct will in many cases be unjustified. But even so, there seems to be one important qualification. If we think that relatively young children are capable of making informed choices about sexual matters, we should then be prepared to tolerate sexual relations between children of about the same age. But it does not follow that we should also tolerate the same type of sexual relations between a child and an adult. If we left children completely free to enter into any relationship with anybody they wish, they are likely to be exploited by some adults with much greater experience.
This brief discussion of paternalism is sufficient to show that Mill is quite consistent in accepting weak paternalism while at the same time rejecting the right of society to impose its value on individuals. There are at least three differences between the two doctrines.
I. Weak paternalism involves the protection of individuals from harming themselves in situations where their decisions impaired. On the other hand, in order to enforce society's Zaitpvalues the law may interfere even when a person's decision to engage in an activity is fully voluntary and clearly Informed. Thus to prevent freely consenting adult homosexuals [113/114] from indulging in homosexual activities in private simply because these activities violate the shared morality of society, is to enforce this morality, but the prohibition cannot be justified on weak paternalistic grounds.
2. Weak paternalism is not concerned to punish moral wickedness. Indeed paternalistic intervention only takes place when the agent's decision is impaired in some way, from the moral point of view, this is generally regarded as reducing his moral culpability. A person who acts as a fully free moral agent in violating the accepted morality of society is outside the pale of paternalistic intervention, but, on the other hand, he is precisely the sort of person against whom society may wish to enforce its shared values.
3. The basis of paternalistic intervention is confined to an appeal to the interests of particular persons who are to be prevented from harming themselves. This is as true of strong Paternalism as it is of weak paternalism. But the enforcement of society's shared values often relies on appeals to all sorts of more general considerations, including, as we have noticed, the protection of the distinctive institutions of society. Paternalism focuses attention constantly on the interests of the person with whom we are interfering. The enforcement of society's shared values is only too likely to sacrifice these mterests to the often unreasoning and unreasonable anger and hatred of the community.
Carter, Rosemary. "Justifying Paternalism" Canadian Journal of Philosophy, 7 (1977).
Feinberg, Joel "Legal Paternalism" Canadian Journal of Philosophy, 1 (1971).
Hodson, John D. "The Principle of Paternalism" American Philosophical Quarterly, 14 (1977).
Houlgate, Laurence. "Children, Paternalism, and Rights of Liberty" Having Children, ed. Onora O'Neill / William Ruddick. Oxford, 1979.
Hughes, Graham. "Consent in sexual offences" Modern Law Review, 25 (1962).
Mill, John Stuart. "On Liberty". Utilitarianism, Liberty, Representative Government (Everyman edn).
Murphy, Jeffrie G. "Incompetence and Paternalism" Archives for Philosophy of Law and Social Philosophy, 60 (1974).
Rawls, John. A Theory of Justice. Oxford, 1972.
Richards, David A. J. A Theory of Reasons for Action. Oxford, 1971.
Williams, Glanville. "Consent and Public Policy" The Criminal Law Review (1962).
Last modified 22 April 2001