An important component of the idea of freedom is that a part of every person’s life should be protected, the “private sphere”, in which they are sovereign and no one else has the right to decide over them. One of the most famous expressions of this idea is John Stuart Mill’s dictum that “[t]he only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others. His own good, either physical or moral, is not a sufficient warrant”.¹ This has been called Mill’s “antipaternalistic principle”.²

Some of the harms that paternalists wish to prevent have a complex causal background. Often, the self-harming actions of the affected person combine with actions of others to produce the harmful outcome. Hence, the drug addict cannot use drugs unless someone provides him with the drugs; the “voluntary prostitute” cannot sell her services unless someone chooses to demean her by buying them; the smoker can smoke only if companies continue to sell products that kill half of their customers, etc. It may of course be claimed that the addict and the prostitute only harm themselves, but the same cannot be said about the drug dealer or the sex buyer.

The case of smoking is particularly instructive. To defend their activities, tobacco companies have sponsored antipaternalist campaigns focusing on the right of smokers to smoke.³ The (implicit) inference seems to be that if the smokers have a moral right to harm themselves, then the tobacco company has a moral right to provide them with the means to do so. In this way, antipaternalism extends to acts that harm others. I will use the term extended antipaternalism to denote the use of antipaternalist arguments for accepting actions and activities that harm or contribute to harming (consenting) others. It should be distinguished from antipaternalism proper—that refers to non-interference with self-harming actions and activities.

It does not take much moral reflection to realise that the inference from antipaternalism to extended antipaternalism does not follow automatically. We all have the right to kill ourselves with cyanide, but few would wish to allow cyanide companies to sell convenient doses of the substance to private people. As this example illustrates, a person’s right to harm him or herself does not necessarily imply the right for others to facilitate or contribute to their self-harming activity.

In spite of the problematic nature of the inference from antipaternalism to extended antipaternalism, the predominant trend in debates on public policy is to take its validity for granted. This inference has also been surprisingly little discussed by philosophers writing on paternalism. Feinberg is one of the few exceptions. He distinguishes between harms that people inflict on themselves and harms that others inflict on them with their consent. Feinberg does not make much of the distinction, but he points out that “[a]pplied to voluntary slavery, the principle of non-exploitation might say that it isn’t aimed at preventing one man from being slave so much as preventing the other from being a slave-owner”.⁴

The purpose of the present contribution is to investigate under what conditions it is reasonable to take the step from antipaternalism proper to extended antipaternalism. This will be done primarily by comparing cases in which extended antipaternalism is generally accepted with cases in which it is not, in search for underlying, consistently applied moral principles.

TOOLS FOR THE ANALYSIS

For the purposes of the present paper I will assume that the standard antipaternalist argument against intervening with self-harming action is valid in a wide variety of cases. (For overviews of arguments for and against paternalism, see references 5 and 6.) The focus will be on these cases—that is, on situations in which people have a moral right to perform some action or activity that harms themselves. The issue under investigation is whether or not such a right carries with it moral permissions or perhaps even obligations for others to contribute to the harm.

Generally speaking, a right is an involved complex of legal or moral relations. A legal right typically includes “a host of legal liberties, powers, and duties of various officials playing diverse rôles in the legal system that creates, protects, and enforces the legal right”.⁷ In strict terms, what I would call my “right” to get back the car that I lent to a friend is not a single, simple legal relation. Instead, it is a complex or cluster of legal relations of various types, including my friend’s obligation to return my car and various claims that I have on legal authorities to act on my behalf in the case of his non-compliance. In general, such a complex has one component which is the mainspring of the whole complex (in this case, the friend’s obligation to return the vehicle). The other components would have no motive in the absence of this one component, which will be called the central legal relation of the complex. The other legal relations in the complex are subsidiary.^8,9 Moral rights are analogous to legal rights in these respects. The same distinctions can be drawn between the complex of moral relations comprising a moral right, and the central and subsidiary moral relations that are included in the complex.

In discussions of paternalism we are concerned with a person’s right to act in ways through which he or she may harm themselves. The central moral relation of such a right consists of the person’s permission to harm him or herself (in certain ways). The subsidiary moral relations that should interest us are those that refer to actions by others that either prevent or facilitate the self-harming actions. Hence, we should look at permissions and obligations not to prevent the self-harming action and also at permissions and obligations to facilitate it. Obligations and permissions of non-prevention will be discussed in the next section, followed by obligations to facilitate and permissions to facilitate.

OBLIGATIONS AND PERMISSIONS NOT TO PREVENT

Obligations of non-prevention are common subsidiary components of rights, and they are also essential for the value of rights to the rights’ holders. My right to read any book I want would not be much worth if the local magistrate had the right to prevent me from reading whatever book they felt was bad for me. More generally speaking, if a person has a right to act in a certain way, then it is reasonable to assume that there are subsidiary obligations for others not to prevent them from doing so. This holds not only for rights to promote one’s own interest, but also for rights to act against them. It is doubtful whether a person can be said to have a right to burn down their own house if the chief officer of the local fire brigade is authorised to stop them from doing so.

Obligations of non-prevention do not cover competitive situations in which one person’s exercise of a right may prevent another’s exercise of that same right. Each of us has a right to sit on any bench in Hyde Park at noon tomorrow, but in the event that two people have decided to sit on the same spot, the person who comes second has no right to have the first removed. Similarly, in liberal societies, anyone has the right to become a successful businessperson, but this does not include protection from being competed out of business by others who exercise that same right. Since competition in self-harm is much less common than competition in self-furtherance, this limitation to obligations of non-prevention does not seem to have much relevance in the cases that concern us here.

Hence, rights to perform self-harming actions are typically accompanied by subsidiary obligations for others not to prevent these actions. However, an obligation not to prevent an action does not imply permission to facilitate or contribute to it. Therefore, obligations of non-prevention do not support extended antipaternalism and for the same reason, nor do permissions of non-prevention.

OBLIGATIONS TO FACILITATE

Before dealing with obligations to facilitate self-harm, it is instructive to consider the more common case of obligations to facilitate self-advancing activities. Some but not all rights to perform actions that promote one’s own interests are supported by subsidiary obligations for others to facilitate such actions. I have a right to climb Mount Everest, but no-one seems to have an obligation to help me achieve this (except as the result of an agreement giving rise to contractual obligations). I also have a right to walk unassaulted in the main street of my hometown. The local police is under a (moral and legal) obligation to make this possible.

Studies of other such examples will reveal a fairly consistent pattern: rights to perform an action are not in general accompanied by obligations for others to facilitate that action or to make it possible. Subsidiary obligations of this type are only attached to those rights that are considered to be particularly important for a person or for society as a whole. Rights to harm oneself have not in general been placed in that category. I am allowed to smash my furniture to smithereens, endanger my life by free climbing, donate all my money to the mafia, or behave in numerous other ways that are clearly against my own interests. Only in exceptional cases are there others who have an obligation to facilitate such actions.

The most important of these exceptions applies to disabled people. In a compassionate society, a severely handicapped person who cannot eat or move has a recognised right to be fed and to be transported to places they want to go to. More generally speaking, there are widely recognised obligations to compensate people for their physical disability. It is, however, controversial how far such obligations go. Most of us would agree that they do not include expensive habits, such as participation in a luxury cruise, unless the person can pay for it with their own means. Nor do they include illegal acts or, arguably, acts that are immoral albeit not prohibited by law.

It is also a matter of controversy to what extent such obligations cover acts that are generally recognised to be against the disabled person’s own interest. In such cases the caregiver is caught between two conflicting obligations: to further the disabled person’s interests and to compensate for their physical inabilities. There are situations in which almost everyone would agree that the latter obligation overrides the former, and also situations in which the reverse is the case. A disabled person should not be denied the opportunity of seeing a TV programme that is known to make them very sad. On the other hand, anyone has the right to hit him or herself on their head with a hammer. A handicapped person has the same right, but may be physically unable to exercise it. Few of us would claim that there is someone who has an obligation to help such a person by hitting him or her on the head with a hammer. Unfortunately, many of the cases encountered in practice are more difficult than these to adjudicate, such as when a disabled person wishes to take a higher dose of a prescription drug than what the doctor recommended, when they wish to drink excessively, or when they have made a well considered decision to commit suicide.

Hence, obligations to perform actions that facilitate self-harm can arise as a result of our general obligation to make handicapped people capable of autonomous action. However, the obligation at work here is not an obligation to facilitate self-harm but an obligation to make the person capable of autonomous action. It is important to distinguish between outcomes that one has an obligation to bring about and outcomes that are unavoidable but undesirable effects of actions one has an obligation to perform.¹⁰ It should also be observed that the obligation to compensate disabled people for their disability is a prima facie obligation. The fact that a particular execution of this obligation harms the handicapped person never strengthens the obligation but in some cases it may contribute to overriding it.

Of course, obligations other than those that compensate for disabilities can have other-harming side effects. Suppose that I have borrowed my new neighbour’s chainsaw and promised to give it back on Saturday morning. On Friday evening a reliable friend tells me that this neighbour has the habit of using the chainsaw in a way that is very dangerous to himself. My act of handing it back will therefore facilitate his performance of potentially self-harming actions. I nevertheless have an obligation (at the very least a prima facie obligation) to hand back the chainsaw. However, this does not mean that I have an obligation to facilitate self-harm. Again, a distinction must be made between outcomes that one has an obligation to bring about and outcomes that are side effects of carrying out an obligation.

PERMISSIONS TO FACILITATE

The meagre result from the previous section should not come as a surprise. For the antipaternalist argument, obligations to facilitate self-harm are not necessary; permissions are sufficient. The major line of inference from antipaternalist to extended antipaternalist standpoints should, expectedly, take the form of inferring a subsidiary permission to facilitate or contribute to someone else’s self-harm from that person’s right to harm him or herself.

Obviously, what can be inferred in this way will not be categorical permissions but rather permissions triggered by the right holder’s consent. (In my terminology these are grantable permissions.¹¹) My neighbour has a right to have the windows of their car smashed with a sledgehammer. It does not follow that I also have a permission to demolish the windows. It only follows that I am permitted to do so if they let me do it. Consent is a necessary requirement for the step from antipaternalism proper to extended antipaternalism. The notion of consent referred to here is not consent as a state of mind but consent as a performative action that can trigger permissions.^12,13

It is easy to find cases in which consent makes harming others allowable, but it is equally easy to find cases in which it does not. Neither law nor ordinary moral intuitions would admit consent as an excuse for mayhem or homicide, for taking someone as a slave, for bigamy, or for carrying out dangerous medical experiments with no foreseeable clinical value. On the other hand, vasectomy, and removal of a kidney for donation are accepted if consent has been obtained, and the same applies to minor acts of violence or nuisance such as pinching someone’s cheek. The general pattern seems to be that consent is taken to be an acceptable excuse for harming others only when this is done for some socially accepted purpose or when the harm is not considered to be serious (cf reference 14).

There is a striking contrast between the notion of consent or voluntariness used in biomedicine (including the pharmaceutical industry) and that used in most other social contexts. Before a subject is enrolled in a medical experiment, the research leader has to make sure that their consent is well informed—that is, based on both oral and written communication that the subject has received and understood. Furthermore, consent must be free in the sense of not resulting from payments or relations of dependence. Finally, the risks to which the subject is exposed have to be minimised, and they must be proportionate to the positive effects expected from the experiment.¹⁵ There is almost no likeness between these requirements and the notions of consent or voluntariness that are referred to, for instance, in discussions of the sale of addictive products.

A hypothetical example can clarify how the impact of consent depends on social conventions. Suppose a major soft drink company comes up with a new product that customers will become addicted to. The new soft drink has no serious immediate health effects, but in the long run it will give rise to cancer and cardiovascular diseases, thereby ultimately shortening the lives of about half of the consumers who become addicted to it. Few would claim that such a product should be allowed, yet its properties are analogous to those of cigarettes. The difference is of course that cigarettes are socially accepted and that it is considered politically impossible to prohibit them. As this example shows, consent or voluntariness does not per se provide us with sufficient moral underpinnings for extended antipaternalism, since the impact of consent depends on social conventions bound by traditions.

Another factor in addition to consent needs to be taken into account, namely the contribution that harmed people themselves make to the harms that befall them. Most of the other-harming actions defended by proponents of extended antipaternalism are only contributory. Without a self-harming action by the affected person her or himself, the harmful effect will not materialise. As an example of this, the manufacture and distribution of tobacco products is harmful only if people actually buy and use these products. Therefore, the smoker who is harmed has some responsibility for the harmful outcome. Tobacco advocates tend to take it (implicitly) for granted that the self-harmer’s responsibility for the outcome pre-empts any responsibility on the part of others who contributed to it. This distribution of responsibilities seems to be a result of social conventions rather than of consistently applied moral principles. Other cases with a similar structure are treated differently. In particular, the manufacturers and distributors of heroin are commonly held responsible for the effects that their products have on the health and wellbeing of those who choose to buy and use them. (Cigarettes are legal and heroin is illegal but that does not settle the moral issue.)

Boxing provides another example of how extended antipaternalism depends on convention-bound ascriptions of responsibility (cf reference 16). Discussions on professional boxing are typically couched in terms that would have been adequate if the two combatants were standing together in the ring, each punching their own face rather than that of their opponent. A more realistic analysis will reveal that there are at least three types of action conferring responsibility that are involved in boxing. Firstly, each boxer performs the self-harming action of letting the other hit them. Secondly, each boxer performs the (reciprocal) other-harming action of hitting their opponent. Thirdly, managers and promoters perform the (one-sided) other-harming action of inducing the boxers to hit each other. In the parallel examples of duels, street fights, and unauthorised prizefights, the corresponding other-harming actions are considered to confer responsibility for resulting bodily injuries (in spite of the concomitant and causally necessary self-harming action). Again, the difference is one of social convention, not of consistently applied moral principles.

Our ascriptions of moral responsibility are not based on causality alone but depend largely on social conventions and traditions. Proponents of extended antipaternalism tend to assign guilt and responsibility to the affected people rather than to others who contribute to their predicament. This practice of blaming the victim may have the effect of averting attention from potential targets of social reform. “Placing the focus in the individual’s role in causation may decrease the efforts spent in dealing with the social issues involved.”¹⁷ (Cf references 18 and 19.)

CONCLUDING REMARKS

Once the distinction has been made between antipaternalism proper and extended antipaternalism, it becomes clear that no general inference from the former to the latter is possible. We may well approve of Mill’s dictum that the only reason to interfere with someone’s behaviour is to protect others from harm, without accepting the marketing of tobacco or heroin, boxing promotion, driving with unbelted passengers, buying sex, or other actions that contribute to harming (consenting) others.

It emerges from the above analysis that our conventional moral attitudes to actions that harm the health of consenting others cannot be reconstructed in terms of consistently applied moral principles. Unless moral incoherence is accepted, this is a reason to intensify the ethical discussion on responsibilities in public health, with the aim of moving in the direction of a reflective equilibrium.