Chapter 7, 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.
- 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].
have suggested that selling oneself into slavery can be brought under Mill's principle which justifies the prevention of harm to others. There is another line of argument which stretches the category of conduct which harms others without their consent even further, and I wish now to explain and resist this argument.
D. H. Regan claims that some acts which are normally regarded as self-harming can be treated as acts which harm another person (Sect. iii). What we ordinarily regard as different time-segments of the same person can more appropriately be viewed as different persons. Thus the motor-cyclist at the time of his choosing to ride without a helmet, and the motor-cyclist who later meets with a serious accident, are two different persons. If there are two different persons, then the act of riding without a helmet is really an act which harms another person, the motor-cyclist's later self. So intervention in the act comes within the scope of Mill's liberty principle. This account draws on Derek Parfit's work on personal identity; cf. Parfit, "Identity," and Parfit, "Later selves;" see also Bernard William's comments, in Williams, "Persons." Parfit refers to the criterion of psychological connectedness for distinguishing between persons. This considers the degree of similarity in the psychological characteristic at different times. [119/120]
However, Regan seems to adopt a narrower criterion. He writes: "the motor-cyclist is a different person, in the relevant respect, if he is no longer the sort of person who would ignore his future wellbeing for the sake of small increments of present utility" (p. 205). So if the motor-cyclist is likely, after being banned in an accident, to regret the past act of riding without a helmet and to change his attitude towards risk, then there is a different person who has been harmed. But since a person's attitude towards risk is only one of his psychological characteristics, it is likely that a change here is not accompanied by a change in other psychological characteristics. This leads to a situation in which for some purposes the same person is present at two points in time, but for other purposes there are two different persons. Regan accepts this. He gives the example of someone who, at about the same time that he embezzles, also commits an unrelated act of aggravated assault. Suppose that after ten years he has grown more honest and would no longer embezzle. However he has not changed in his attitude towards physical aggression. Regan thinks that it is inappropriate to punish him for embezzlement, but proper to punish him for assault. This situation can be handled easily enough by the resources we ordinarily use of saying that one and the same person is to be punished for assault but not for embezzlement. But if we adopt Regan's criterion of when we have the same person, we face the problem of determining the relation between the assailant and the person who is not punished for embezzlement. They are not the same person, but there is no way of punishing the assailant which does not also adversely affect the other person. If the assailant is imprisoned, the other person will also find himself in gaol. If the assailant is fined, then the resources of the other person are also reduced. All their desires, interests, and values are the same. Even Siamese twins are not so closely bound together.
Similarly, if we apply Regan's criterion of personal identity, we would have to say that the motor-cyclist, who rides without a helmet, is not the only person affected by punishment. An ordinary person has many projects, not all of which may fit into a single, over-all plan. For example, he has made promises which he intends to keep, and these promises are [120/121] unrelated to his desire to ride without a helmet. By punishing him for riding without a helmet, the state may deprive him of the opportunity to keep some of his promises. Punishing the motor-cyclist will therefore also interfere with the freedom of an innocent person, the promise-maker. This seems unfair. So even if we accept Regan's view that the motor-cyclist's act can harm another person, the same reasoning also provides us with a basis for not punishing him, and it is not clear how we are to balance the case for punishing the guilty motor-cyclist against the claims of the innocent promise-maker.
Punishing the motor-cyclist is of course a preventive measure taken to reduce the harm caused if there is an accident. So the real offence of the motor-cyclist is that of engaging in a dangerous activity which is likely to cause harm to another person, namely, his later self. It is of course perfectly proper to prevent someone from performing a dangerous act even though not all such acts will actually cause harm to others. For example, not every single act of driving at high speed through a busy street will cause harm. It is enough that every such act is likely to cause harm. However, in the ordinary case of dangerous driving, although actual harm to others may not be inflicted, there are actual persons who are likely to he harmed, and who therefore need protection. But in the case of the motor-cyclist, if there is no accident, then it is not just the case that he does not harm another person. There will also be no other person to be harmed. The accident both gives birth to and harms the later self. But still, there are other analogies to draw on. It is certainly not the case that the only type of harm to others is harm to actual persons who already exist at the time of the relevant act. Environmental pollution may not harm the existing population, but it will harm future generations. No doubt in this case the people who will be banned will come into existence independently of the harmful act, but it is unclear whether this is a significant difference.
But there is one important difference. Death is not a relevant harm in the case of the motor-cyclist. For if the accident causes death, then it does not produce a later self who is harmed. Death, at least instant death which leaves no time for regret, ensures that there is no birth of the later self, [121/122] and so death makes the accident self-harming but not harmful to another person. This point is recognized by Terry S. Kogan who adopts Regan's position. Kogan argues that the account of later selves cannot be used to justify state intervention in suicide.
In order for Mill's justification of state intervention to apply intertemporally, it is necessary that one make a reasonable prediction that future selves will disidentify with the present self. But where it is clear that there will be no future selves, such prediction becomes meaningless. Since there will be no future intertemporal "others", suicide is a purely self-regarding action. [pp. 841-2]
But the same point applies to all risky actions which cause death, or severe brain damage in which the person, while still alive, is unconscious or is incapable of formulating new values and plans of life, and is therefore incapable of renouncing his old values and attitudes. From such a brain-damaged being no later self can emerge to be harmed. It is paradoxical that activities which are likely to cause these types of serious harm fall outside the scope of Regan's and Kogan's new defence of paternalism, whereas activities which are likely to cause lesser harm can be interfered with.
Kogan also claims that there is evidence to support the view that a person will reject his earlier goals as a result of the traumatic experience of a serious accident (p. 837). If this is true, then it presumably applies even when a helmet or seat belt was worn. The accident may, for example, be caused by driving at high speed, or by the failure to pay sufficient attention, and the person now regrets and rejects his previous attitudes. If it is the trauma of the serious accident which causes the change in him, sufficient for a later self to emerge, then this change can come about even when he had taken the precaution of wearing a helmet or seat belt. Now those who support the compulsory wearing of helmets and seat belts do not of course claim that the rate of accidents thereby reduced. What is rightly claimed is that when there are accidents, the wearing of helmets and seat belts reduce the extent of harm caused. No doubt death and severe brain damage will be among the serious harm that will be reduce. But we recall that on Regan's and Kogan's account, if there is death or severe brain damage, then there is no harm to[122/123] others as there are no later selves. This being the case, the probability that the wearing of seat belts and helmets reduces the incidence of these types of harm constitute a reason against requiring them to be worn. For when the harm caused is death or severe brain damage, the victim is the present self and not a later self. Accidents which result in these types of harm do not therefore cause harm to others, but are purely self-harming. If the wearing of seat belts and helmets reduces the number of deaths and cases of severe brain damage, then it increases the harm caused to others by reducing the number of self-harming actions. This is a surprising but inescapable implication of an appeal to the idea of harm to later selves.
Kogan, Terry S. "The Limits of State-Intervention: Personal Identity and Ultra-Risky Actions" The Yale Law Journal, 85 (1976).
Parfit, Derek. "Later selves and moral principles" Philosophy and Personal Relations. London, 1973.
_____ "Personal Identity" Personal Identity, ed. John Perry. Berkeley, Los Angeles and London, 1975.
Regan, "Justifications for Paternalism" The Limits of Law, ed. J. Ronald Pennock and John W. Chapman, Nomos 15. New York, 1974.
Williams, Bernard. "Persons, Character and Morality" The Identities of Person, ed. Amelie Oksenberg Rorty. Berkeley, Los Angeles and London, 1976.
Last modified 22 April 2001