A strange form of entertainment appeared among French youth at the beginning of the 1990s. Local discotheques started to organize ‘dwarf tossing’ events, in which a dwarf, wearing suitable protective gear, would allow himself to be thrown short distances onto an air bed by the clients. French authorities soon prohibited these events, on the basis that the practice violated the principle of human dignity. In a series of appeals, the case was brought before the French Conseil d’Etat and the UN Human Rights Committee.  The applicant, Mr. Wackenheim, one of dwarf employees in question, claimed that banning him from his work had an adverse effect on his life and represented an affront to his dignity, since human dignity consists in having a job. Although his petition was rejected by both forums, Mr. Wackenheim’s reasoning raises interesting issues both from an ethical and legal point of view. In this paper, we shall investigate the central question of the case, namely: What are the acceptable limits of state intervention to protect competent individuals from their own self-harming conduct? Put it more directly, what are the rules and limits of paternalistic interventions?
The origins of the expression
The word ‘paternalism’ has strong negative connotations attached to it. Paternalism is something we often accuse people (or systems) of. It is, however, not an inherently immoral concept. Rather, the word has acquired its bad connotations throughout the course of history. Paternalism is a concept forged during the historical process of the Enlightenment and it is closely connected to the birth of modern liberalism.
Paternalism is commonly traced back to the medieval concept of patriarchalism, which used the paternal analogy to justify the privileges of the feudal ruling elite over the common people. The main ‘ideologist’ of this approach, Sir Robert Filmer, argued that the Sovereign has power over his people just as father has power over his children. In his book Patriarcha, Filmer made great efforts to trace back the lineage of kings to Adam and the Biblical patriarchs in order to prove that rulers are in a certain way parents of their people (since they are direct descendants of Adam, the father of all people), and therefore have natural paternal rights over them. 
With the coming of the Enlightenment, divine rights theories were gradually replaced by contractarian models of sovereignty. Locke provided an extensive critique of Filmer’s Patriarcha in his First Treatise of Government, rejecting the analogy between parental power and political sovereignty. It might be that the term paternalism was originally coined to combat the use of parental analogy by the former ruling elite, but in the process of conceptualization, paternalism acquired the same negative connotations.  Paternalism was not welcomed in a time that prized individuality and autonomy over other values. Paternalism implies unjustified authority and suggests that the state stands to its citizens as a father stands to his children, treating competent adults as if they were children. If adults are treated as children, it is fearful that they will never emerge from their ‘self-imposed immaturity’ and easily fall prey to the oppression of tyrannical political regimes.
Kant echoes this concern in the following way: ‘If a government were founded on the principle of benevolence toward the people as a father’s toward his children – in other words if it were a paternalistic government (imperium paternale) […] – such a government would be the worst conceivable despotism.’ Immanuel Kant, On the Old Saw: That May Be Right in Theory, but It Wont’t Work in Practice, (1974) 58-59.”
Although the present attitude towards paternalism is determined by these modernistic presuppositions, paternalistic behavior has not always been associated with pejorative connotations. Aristotle, for example, suggested that the polis does not only have the right to act paternalistically towards its citizens, but often has a duty to do so. ‘Any polis which is truly so called, and is not merely one in name, must devote itself to the end of encouraging goodness. […] What constitutes a polis is an association of households and clans in a good life, for the sake of attaining a perfect and self-sufficing existence’.
The concept of paternalism
There is a huge variety of definitions readily available for paternalism in the philosophical literature.  Paternalism, in its crudest form, can be defined as coercive intervention to the behavior of a person in order to prevent an individual from causing harm to his or her self.  This definition has two basic components, namely (1) coercive intervention and (2) the aim of preventing self-harm. We need to take into account the following objections at this point.
Ad (1): Paternalistic conduct is not always coercive and does not always interfere with the explicit will of the subject. One can plausibly argue that the concept of paternalism does not presuppose coercion or direct interference with the subject’s liberty of action. 
Consider, for example, the cases of benevolent deception, misinformation or withholding information. A doctor who lies to a mother on her deathbed about her son (telling her that he is doing well, although he has just lost his life in a car accident) definitely acts paternalistically. Paternalism does not even need to involve an attempt to control the behavior of the person. Giving blood transfusion to the unconscious victim of a traffic accident, while knowing of his religious objections to transfusion, is paternalistic, yet it does not coerce or seek to change the behavior of the subject. See Gert – Culver (1976) 46; Archard (1990) 36.”
Thus, it seems more promising to conceive paternalism as being connected to the violation of the protected person’s autonomy, where autonomy encompasses both internal (i.e. mental capacity, rationality, skills for preference formation, etc.) and external factors (i.e. adequate set of options).
Ad (2): The aims of paternalism are not only prevention from self-harm but also the promotion of the welfare and interests of its subjects. Compulsory primary education or compulsory social security membership does not directly prevent self-harm but aims to create long-term benefits for citizens. Consequently, we can distinguish between benefit-promoting and harm-preventing forms of paternalism.
Overall, there seem to be two elements that form the core concept of paternalism: (1) the interference with the subject’s autonomy and (2) the benevolent aim of preventing self-harm or promoting benefit.
The element of interference with autonomy
There are many examples of paternalism that are non-coercive but still interfere with personal autonomy. Incentives expand rather than limit the number of available options, thus they cannot be seen as violating liberty in the traditional sense. However, the father who offers his teenage son monetary rewards for earning good grades in school treats his son paternalistically;  the violation of autonomy being manifested in the father’s ‘tampering’ with the child’s autonomous preference-formation. Other ‘nudges’ operate by modifying choice- architecture and exploiting people’s inertia. They aim to structure others’ options in a way that the default option is the best option and less beneficial options can be selected only by opting-out of the best one. Changing the default rule from non-enrollment to automatic enrollment for an otherwise non-compulsory training course will increase the participation of employees, since most of them will probably not opt-out after being enrolled. Putting the apples and oranges before the snacks at a school cafeteria would probably increase the consumption of fruits and decrease the consumption of snacks.
Autonomy-based definitions of paternalism seem sufficiently broad. Arguably, certain forms of paternalism still fall outside the ambit of such definitions. Consider the following ‘hard case’ from Heta Häyry.
‘Dr. Smiley is a dentist and in the waiting room of her office there are magazines for the patients. Now, if Dr. Smiley decides to add to the reading selection some pamphlets propagating the importance of dental hygiene, her behavior will hardly constitute violations of the patient’s autonomy or other transgressions of major moral rules. Nevertheless […] it might well go against the common usage of language to say that her behavior cannot be prescribed as paternalistic unless there are violations of some sort present.’ Häyry (1991) 61.”
It is a matter of discussion whether such a practice is paternalistic or not. Mere provision of information, like warnings on cigarette boxes or television commercials drawing attention to the harmful consequences of smoking, does not substantially affect the subject’s autonomy.Rational persuasion is also nonvolatile to freedom, if both negative and positive arguments are impartially communicated. 
The element of benevolence
The other, generally underemphasized, component of the definition of paternalism is the element of ‘benevolence’. It requires that the paternalistic intervention (1) be aimed at (2) preventing self-harm or (3) promoting the benefit of the subject. The first constituent (‘be aimed at’) stresses the importance of the motives and intentions of the paternalistic actor. Although subject to disagreement, it seems plausible to accept that the paternalistic agent does not necessarily have to be successful in achieving good for the subject. 
Consider, for example, the case of an overprotective mother whose over-caring distorts her child’s psychological development. Her behavior is paternalistic even though not actually beneficent. Conversely, an actually beneficent intervention that lacks benevolent motives does not qualify as paternalism. One may protect others from self-harm for non-paternalistic reasons. For example, a master may prevent a slave from killing or disfiguring herself in order to use her as a prostitute. Kultgen (1995) 62.”
The second and third constituents of the element of benevolence are (2) the prevention of self-harm and (3) the promotion of benefit. Conceptually, this distinction does not seem to be overly important. Harm-preventing and benefit-promoting paternalism can be thought of as the two sides of the same coin. Protection from harm is benefit promotion in the long run and unrealized benefits can be conceived as a form of harm.
Self-harm is not the only reason that can justify intervention with individual liberty. Generally, four main liberty-limiting principles are distinguished: the harm principle, the offense principle, paternalism and moralism.  Since these principles often overlap, conceptual clarity requires that we make a clear distinction between them.
Reference to harm caused to others is a common argument to support intervention with individual liberty. In the pivotal essay of John Stuart Mill, On Liberty, ‘harm to others’ appears as the only possible reason for state coercion.
The most striking feature of Mill’s principle is its exclusive character. It sets a blanket prohibition to both paternalism and moralism. Hart aptly remarks that ‘Mill carried his protests against paternalism to lengths that may now appear to us fantastic’. 
Legal moralism is the possibility of states to interfere with a person’s liberty on the sole ground that the conduct is inherently immoral. Since self-harming conduct is often considered as immoral (at least in the Judeo-Christian tradition), moralistic and paternalistic reasoning tend to go hand in hand. To distinguish them, it is possible to say that when a law restricts the individual’s choices for his moral good, it represents ‘moral paternalism’,  which is a form of benefit-promoting paternalism. When a law restricts the individual’s choices in order to enforce social morality, we can speak of ‘legal moralism’.
The scope of the harm principle depends on how we understand the concept of harm. Harm must be distinguished from actions that cause feelings we don’t like but do not necessarily harm us. One such category of actions is the category of offense.  Offense is the production of a non-painful, unpleasant or uncomfortable ‘universally disliked mental state’ in a person (e.g. unpleasant or loud sounds, shocking or disturbing public behavior, etc). The offense principle is generally used as an auxiliary principle to the harm principle: it is not only the prevention of physical harm but also offensive behavior that can justify legal coercion. The offense principle can also overlap with other principles, becoming a tool of legal moralism or paternalism. Prohibiting alcohol consumption in public places is paternalistic but the original aim of such regulations is to safeguard public order. ‘Immoral’ behavior can also be banned under the offense principle; sexual intercourse in private is not the state’s business but it is prohibited if performed publicly due to its offensive character.
Two practices are frequently confused with paternalism in colloquial speech. The first kind of practice has the benevolent character of paternalism but it is not connected to self-harm. When the state creates special legal regimes to prevent child abuse or to protect vulnerable people from exploitation, it acts in such ‘quasi-paternalistic’ way. The so-called ‘good’ and ‘bad Samaritan laws’ also belong here. Bad Samaritan laws require third parties to provide help to victims of accidents, natural catastrophes or crimes.  Good Samaritan laws provide protection against liability for persons who cause harm in the course of providing aid. The second category often confused with paternalism contains practices that are not even remotely benevolent. Take the example of the management of a factory, which, for greater efficiency, treats employees as if they were school children (keeping a constant close watch on them, requiring them to ask permission to go to the toilet, etc.). Such rules are not paternalistic but they are sometimes confused because their authority reminds us of parental authority.
Different subcategories of paternalism
Paternalism has different types: we can distinguish between soft-hard, direct-indirect, mixed-unmixed, coercive and non-coercive, harm-preventing and benefit-promoting forms of paternalism. The soft – hard distinction is based on the voluntariness of the subject. Since it plays an important role in the justification of paternalism, I am going to discuss this category separately in the next chapter.
In the case of direct paternalism, the class of persons whose autonomy is restricted is identical with the class of persons whose benefit is intended to be promoted by such restrictions. Laws prohibiting suicide or requiring passengers in automobiles to wear seat-belts are typical examples. Indirect paternalism involves the restriction of other persons besides those who are benefited. Consider, for example, consumer protection laws or licensing requirements for doctors or lawyers.
While the direct-indirect category focuses on the subject’s autonomy, the mixed-unmixed distinction emphasizes the motives of paternalistic intervention. In practice, paternalistic regulations are not exclusively motivated by protection from self-harm. Other considerations, such as the protection of others, morals, public order, etc. are also taken into account. There are very few unmixed cases of paternalism. Even in a seemingly pure case (e.g. prescribing motorcyclists to wear crash helmets), one can refer to the indirect harm caused to others by the additional social security expenses.
Finally, distinction must be made between coercive and non-coercive forms of paternalism. The original formulation of the harm principle extended only to coercive interferences with liberty. However, there are many examples of non-coercive paternalism; the recent trend of libertarian paternalism operates almost exclusively with such methods.
There are two main approaches towards the justification of paternalism, corresponding to two main trends of normative ethics: autonomy-based (also known as ‘deontological’) and consequentialist justifications.
As we have seen, autonomy has a central role when it comes to the definition of paternalism. It is also important for justifications based in deontological ethics which commands an absolute respect for the autonomy of the individual:
For if the essence of men is that they are autonomous beings – authors of values, of ends in themselves, the ultimate authority of which consists precisely in the fact that they are willed freely – then nothing is worse than to treat them as if they were not-autonomous, but natural objects […] 
Paternalism, especially hard paternalism, seems to deny individual autonomy. In order to explore what kind of paternalistic interferences violate autonomy, I distinguish between two forms of autonomy-based justifications: the voluntariness and the consent-based approaches.
From the viewpoint of justification, the most significant distinction is between soft (weak) and hard (strong) paternalism. This distinction is based on the voluntariness of the subject. Hard paternalism advocates coercion to protect competent adults against their voluntary self-harming choices. Soft paternalism allows protection from self-regarding harmful conduct, if ‘the conduct is substantially nonvoluntary, or when temporary intervention is necessary to establish if it is voluntary or not’. Soft paternalism seems morally acceptable because it respects autonomy and does not impose external values on the individual.
Therefore, the central concern here is how to decide if a self-harming act is substantially voluntary or not. Determining voluntariness is a complex issue that requires the consideration of numerous factors. For example, there is direct harm and there are cases when only a risk of harm is present. A person who knowingly swallows a lethal dose of arsenic or chops his left hand off with an axe directly causes harm to himself.  On the other hand, to smoke cigarettes or to drive at an excessive speed is not directly harmful. It increases the risk of harm. We also have to distinguish between reasonable and unreasonable risks. Five considerations help us to determine the reasonability of risk: (1) the probability of self-harm, (2) the probability of the desired goal, (3) the magnitude of the harm, (4) the magnitude of the desired goal (‘the value or importance of achieving the goal’) and (5) the existence or absence of an alternative, less risky means to the desired goal. Although these elements might prove helpful in the analysis of ‘reasonable’ risk-taking, reasonableness remains essentially subjective: we can attribute different preferences to different desires and it is, at least, unsure if it is possible to measure preferences by a common understanding of reasonableness. Thus, introducing a distinct measure of voluntariness for self-harming actions seems to be not at all unproblematic.
Consent-based justifications require the subject’s consent to the paternalistic intervention. The idea behind this approach is that the person authorizes paternalism by waiving his or her right to autonomy through an act of valid consent.  We can distinguish between five types of consents: contemporaneous, prior, subsequent, anticipated and hypothetical consent.
In the case of contemporaneous consent, the subject consents to the agent’s paternalistic restriction at the time of restriction.  Autonomy is ‘retained’ in the voluntariness of the consent which establishes a connection between consent-based theories and the hard-soft justificatory scheme. Contemporaneous consent can be given explicitly (by words or by conduct) or tacitly: the form of consent is irrelevant for its justificatory power.
Prior consent to paternalism is exemplified in cases of weakness of will, referred to as akrasia in ancient Greek. This type of consent is thought to justify paternalistic intervention because it represents the subject’s settled preferences in contrast to his short-term, episodic ones. According to the classical example, Odysseus gave his prior consent to paternalistic coercion when commanding his crew to tie him to the mast of his ship and not let him go under the spell of the Sirens. The same rationale makes a drug addict agree to a medical treatment that involves the limitation of his or her liberty or an elderly person to sign a Lasting Power of Attorney that authorizes paternalistic support.
The issue of subsequent consent typically comes up in the case of children whose liberty is restricted by parents, with an appeal to the children’s best interests. Such paternalistic interventions are often said to be justified by the subsequent consent of children once they have grown up. However, justification by subsequent consent seems a ‘hazardous business’. How can we know that the consent will be actually given? What if the paternalistic intervention turns out to be useless or harmful? Or what if the child dies before he or she could have acknowledged the things parents had done for him? There is also the threat of manipulation: it is possible to structure the child’s preferences during the upbringing so that he will acknowledge paternalistic interventions that were not truly beneficial to him.
The anticipated consent argument comes up in cases when the subject did not have the opportunity to consent to the paternalistic intervention but consent can be inferred from the circumstances. It is similar to the subsequent or the hypothetical rational consent argument in that the element of consent is missing at the time of intervention. Stopping someone to cross an unsafe bridge or giving medical treatment to an unconscious victim of an accident are possible examples here.
According to the ‘hypothetical consent scheme’, a paternalistic act is permissible if a fully rational individual would consent to having the given restriction imposed upon him. People often act irrationally because they either attach seemingly ‘incorrect’ weights to their values or simply neglect to act in accordance with their real (i.e. higher-order) preferences.
To demonstrate the first case, Gerald Dworkin argues that if people who don’t fasten their seatbelts or wear crash-helmets on motorbikes were to be involved in an accident, they would admit in retrospect that the inconvenience of using the seatbelt or wearing the helmet wasn’t as bad as they originally thought [cf. Feinberg (1986c) 135]. He claims that this is a clear instance of attaching ‘incorrect’ weights to one’s preferences. For the second case, Dworkin mentions an instance of akrasia: the smoker, who is aware of the harmful consequences of smoking but is unable to quit due to addiction. ”
Dworkin argues that paternalism is permissible in the second case but he is hesitating in the first one. He implies that there is a distinct standard of rationality when assessing the correct weight of different values. This standard can be established by equating the goals of the subject with the life-plan of an averagely rational person. But is it possible to measure incommensurable value choices by introducing an objective standard of rationality?
Consequentialist justifications of paternalism
Consequentialist justifications focus on the outcome of paternalistic intervention. Simply put, paternalism is morally justifiable if it leads to ‘good’ or ‘desirable’ consequences. The question is what we consider as ‘desirable’ consequence. Classical utilitarianism, for example, has a hedonistic character: pleasure is the ultimate good that measures the morality of all actions. Ever since Bentham, there have been many different approaches as to how the concept of good shall be established. At the two extremes, consequentialist theories can either adopt a desire or an ideal theory of the good. Ideal theories claim that some states of affairs are objective components of the person’s good independently of what the person desires (cf. perfectionism). Desire theories focus on the satisfaction of individual desires to the maximum extent possible (cf. welfarism). From the perspective of paternalism, much depends on which theory we subscribe to. If promoting one’s well-being consists exclusively in satisfying one’s actual preferences, for instance, then paternalism can hardly be justified under any circumstances. In practice, desire and ideal theories are usually mixed: the good to be promoted by paternalistic intervention has subjective and objective components. The question is how to create an adequate theory of the good that respects individual autonomy but stays away from ethical relativism at the same time. 
So what kind of observations can we make about the relation of paternalism and consequentialism in general? Firstly, one can plausibly argue that consequentialism does not command principled anti-paternalism. A long-time presupposition of neoclassical economics was that the market always produces the best (i.e. most efficient) result and that economic efficiency and paternalistic interventions are incompatible with each other. This presupposition has increasingly been questioned by recent economic theory. Mill’s attempt to justify the harm principle (and its blanket prohibition on paternalism) on utilitarian grounds is also considered to be unsuccessful nowadays.  It seems obvious that paternalism can increase utility sometimes, both in the short and the long run. Secondly, it shall be emphasized that consequentialist theories are not necessarily hostile to autonomy. Autonomy can be a component of well-being; it can even hold a central place amongst the goods constituting welfare. However, trade-offs between freedom and other values might still be possible. In fact, a common criticism against consequentialism is that it might justify ‘too much’ coercion if autonomy-based side constraints are disregarded. As Gert and Culver warns us, ‘in discussing the justification of paternalism, it is very easy to fall into the error of supposing that all that we need do is compare the evils prevented with the evils caused and always decided in favour of the lesser evil.’ 
The argument from freedom maximization also rests on utilitarian grounds. In short, this argument claims that paternalism can be justified if it increases freedom in the future, even if it restricts actual freedom in the present. Freedom maximization seems to attribute an objective value to autonomy, thus comes very close to (liberal) perfectionist value theories.
Paternalism in Practice – Selected Issues and Cases
In constitutional adjudication, paternalism generally appears as a limitation of the right to privacy. Personal autonomy is primarily protected by the right to private life which is intended to ensure, with the words of the European Court of Human Rights, ‘the development, without outside interference, of the personality of each individual in his relations with other human beings’.  Since not necessarily enumerated as a fundamental right in all jurisdictons, it is often established by means of constitutional interpretation. In a wider sense, one can contend that any fundamental right serves the interest of preserving autonomy either by providing freedom from state intervention (i.e. civil and political rights) or freedom to pursue an authentic life (i.e. economic, social and cultural rights). In this context, paternalistic limitations are usually justified by reference to the protection of rights of others or the protection of public interest. If the harm is entirely self-regarding, certain non-derogable rights (i.e. human dignity and the right to life) will also trump the right to privacy and authorize paternalistic interventions.
Dangerous substances and smoking regulations
Until the early part of the twentieth century, contagious diseases were a common cause of death.  Today, in developed countries, death from contagious diseases is relatively rare and more people die from illnesses such as cancer or heart disease, often associated with unhealthy lifestyle (e.g. smoking, excessive drinking, unhealthy diet, etc.). In the case of contagious diseases, public health measures can be justified by appeal to the harm principle. The legal restriction of smoking, alcohol consumption or drug use – while being evident examples of paternalism – exhibits a mixed rationale for limitation. The sources related to these issues are extensive; this report shall only focus on the rules of smoking regulations.
The ban on smoking in public places is usually justified by reference to harm caused to others.  Today, it is well known that environmental tobacco smoke causes physical harm to non-smokers ‘varying from minor irritations to demonstrable respiratory, cardiovascular, and carcinogenic effects’.  Besides its negative effects on health, restrictions often refer to the indirect social harm (e.g. increased medical costs) and to the nuisance caused to others by smoking.  The harm and offense principles, taken together, seem to provide sufficient basis for limiting smoking in public – a statement that is being increasingly confirmed by recent policy developments throughout Europe. 
A widespread smoking ban was introduced in England in 2006, making it illegal to smoke in all enclosed public places from 1 July 2007.  Certain exemptions were nevertheless granted, including the possibility to smoke in psychiatric institutions until 1 July 2008. A recent case before the High Court involved a claim by mental health patients that the ban on smoking in the psychiatric institution where they were detained violated their right to privacy as enshrined in Article 8 of the ECHR. But does the right to privacy provide protection from non-smoking policies? In other words, is there a ‘right to smoke’?
The Counsel of the petitioners argued that ‘respect for private life involves permitting someone to do what they want to do, however foolish others might consider the activity’. Lord Justice Pill, however, refused to attribute the same scope to the right to privacy as to the harm principle:
We do not accept that the respect required by Article 8 is coextensive with the right of absolute independence contemplated by John Stuart Mill. The law may place restrictions on a person’s freedom of action without necessarily interfering with the right to respect required by Article 8. The expression ‘personal autonomy’, used by the ECtHR in Pretty, undoubtedly resonates with Mill’s philosophy but […] the protection ‘falls some way short of protecting everything a person might want to do even in that private space’.
In other words, the Court was convinced that to prevent someone from smoking does not affect that individual’s ‘physical or moral integrity’ and it rejected the idea that the right to privacy could involve an ‘absolute right to smoke’ wherever one is living. The Court of Appeal, upholding the decision of the High Court, argued that both the nature of the activity and the nature of the place where the activity is exercised must be taken into account when determining the scope of privacy. Although not denying that the mental health institution can be considered as ‘home’ to the patients, the Court established that the ‘inner circle’ of personal autonomy ‘whilst not destroyed, is significantly penetrated by reason of the very fact that a person is confined within a secure hospital’.  Arguably, ‘[t]here is no basis for distinguishing the loss of freedom to choose what one eats or drinks in such an institution and the ban on smoking’.  While not implying that the final judgment is incorrect, connecting the notion of autonomy to the nature of the place where it is being exercised – especially in cases when people have no influence over where they want to reside – carries the possibility of extensive paternalistic treatment.
The flipside of the ‘right to smoke’ question, namely the positive obligation of states to protect non-smokers (i.e. the right to be protected from smoking) was considered by the European Commission of Human Rights in 1998. The Commission concluded that even though Member States have a positive obligation to protect the life and privacy of individuals from third-party interventions, the ‘choice of the means’ is a matter that falls within each State’s margin of appreciation.  German anti-smoking regulations were found to be sufficient in the present case and the Commission ruled that ‘the absence of a general prohibition on advertising of tobacco products and on smoking does not amount to a failure on the part of the German State to ensure the applicant’s rights under Articles 2 [right to life] and 8 [right to private and family life] of the Convention.’
Prohibiting smoking in cases where no other regarding harm is present is outright hard paternalism.  Thus, legislators often resort to non-coercive means to discourage people from smoking, like raising the prices of cigarettes or imposing restrictions on trade and advertising. Since these ‘nudges’ do not straightforwardly interfere with personal autonomy, the area of legal scrutiny shifts from the privacy rights of the protected person to the economic freedoms of the paternalistic actor. There are three legislative trends aimed at reducing tobacco consumption in the European Union: the regulation of tobacco advertising, the approximation of excise duties  and the labeling regulations raising awareness to the risks of smoking on the packaging of tobacco products.
The problem of paternalism appears in quintessential form in the case of voluntary active euthanasia.  The European Court of Human Rights had to discuss the legality of this issue in the Pretty v United Kingdom case. 
The core of Mrs. Pretty’s complaint involves Article 3 of the Convention. She submitted that the suffering she faced due to her illness qualified as ‘degrading treatment’ and she was forced to subsist in a condition that is incompatible with human dignity. Two non-qualifiable rights clash here: the right to life and the right to human dignity. The Court gave preference to the right to life because it observes that the protection of human dignity does not demand positive action from the State in this case; Member States have an obligation to refrain from inhuman or degrading treatment but should not actively assist in suicides.
In relation to Article 8, the Court remarks that the concept of private life is a broad term not susceptible to exhaustive definition. It covers, for example, the physical and psychological integrity of a person, aspects of social identity, gender identification, sexual orientation and sexual life.  Article 8 also protects a right to personal development, and the right to establish and develop relationships with other human beings and the outside world. The Court states that ‘the notion of personal autonomy is an important principle underlying the interpretation of the Convention’s guarantees’ and it can be understood as ‘the ability to conduct life in a manner of one’s own choosing’ Accordingly, the judges acknowledge that the UK limitation on assisted suicides constitutes an interference with Mrs. Pretty’s privacy rights, but it is justified as being ‘necessary in a democratic society’ for the protection of the rights of others [being an assisted suicide]. Eventually, the Strasbourg Court found the UK’s policy on assisted suicides compatible with the European Convention and rejected all claims of Mrs. Pretty.
Other Cases of Physical Protection
Paternalism that aims to protect bodily integrity can take various forms. Consider, for example, regulations requiring the use of safety-belts or crash helmets; prescribing brightly colored jackets for hunters or safety equipment for construction workers; establishing restrictive rules for dangerous sports (e.g. boxing, motor racing), or simply prohibiting dangerous activities (e.g. base-jumping).  The traditional objection against these restrictions makes an appeal to the autonomy of the individual, although most mentioned cases go undisputed since they have only instrumental functions and do not interfere with the basic life plans of individuals. However, even such regulations can exhibit conflicts of values. Consider the issue of compulsory vaccination of children.  The petitioners in this case refused to vaccinate their children because it allegedly offended their religious convictions to ‘keep the body healthy by injecting antigenic materials into the veins’ and argued that ‘the State acts as an unwanted guardian and makes decisions about the citizens […] for its own objectives without providing an opportunity for deliberation’. 
 John Kleinig argues that moral condemnation is not inherent in the word ‘paternalism’, unlike, for example, in the word ‘murder’. John Kultgen replaces paternalism with the word ‘parentalism’ in his book. It is tempting to follow this option, not only because of the expression’s gender neutral character, but also because it is independent from the negative connotations paternalism carries. John Kleinig, Paternalism (Manchester: Manchester University Press 1983) 4; John Kultgen, Autonomy and Intervention – Parentalism in the Caring Life (New York: Oxford University Press, 1995).
 Consider the following quote from Filmer: ‘It may seem absurd to maintain, that Kings now are the Fathers of their People, since Experience shews the contrary. It is true, all Kings be not the Natural Parents of their Subjects, yet they all either are, or are to be reputed the next Heirs to those first Progenitors, who were at first the Natural Parents of the whole People.’ Robert Filmer, Patriarcha and Other Writings (Cambridge: Cambridge University Press, 1991) 10 (Chapter 1, Section 8).
 Locke accepts paternalism as legitimate in parent-child relations. Thus, the negative connotations of paternalism probably stem from its negative perception in public relations (i.e. patriarchalism). Consider the following excerpt on paternal power: ‘Children, I confess, are not born in this full state of equality, though they are born to it. Their parents have a sort of rule and jurisdiction over them, when they come into the world, and for some time after; but it is but a temporary one. The bonds of this subjection are like the swaddling clothes they art wrapt up in, and supported by, in the weakness of their infancy: age and reason as they grow up, loosen them, till at length they drop quite off, and leave a man at his own free disposal.’ John Locke, Second Treatise of Government, section 55.
 The following is a non-exhaustive list of possible definitions. Gerald Dworkin, Paternalism, in Gerald Dworkin (ed.),Mill’s On Liberty, (1997) 62; Joel Feinberg, Legal Paternalism, in Joel Feinberg, Rights, Justice and the Bounds of Liberty, (1980) 110; Jeffrie G. Murphy, Incompetence and Paternalism, 60 Archiv für Rechts- und Sozialphilosophie 465 (1974) 465; Bernard Gert, Charles M. Culver, Paternalistic Behavior, 6 Philosophy and Public Affairs 45 (1976) 49; John Hospers, Libertarianism and Legal Paternalism, 4 Journal of Libertarian Studies 255 (1980) 255; David Archard, Paternalism Defined, 50 Analysis 36 (1990) 36; Ernesto Garzón Valdés, On Justifying Legal Paternalism, 3 Ratio Juris 173 (1990) 173; Simon Clarke, A Definition of Paternalism, 5 Critical Review of International Social and Political Philosophy 81 (2002); Kalle Grill, The Normative Core of Paternalism, 13 Res Publica 441 (2007) 442; Danny Scoccia, In Defense of Hard Paternalism, 27 Law and Philosophy 351 (2008) 353.
 Compare with the similar definitions of Gerald Dworkin and Jeffrie Murphy. Paternalism can be understood as ‘the interference with a person’s liberty of action justified by reasons referring exclusively to the welfare, good, happiness, needs, interests or values of the person being coerced’. ‘Paternalism is the coercing of people primarily for what is believed to be their own good.’ Dworkin (1983) 20; Murphy (1974) 465.
 Scoccia (2008) 353. As Kultgen explains, ‘rewards’ can violate personal autonomy because they alter the conditions of choice. A highly paternalistic society based on positive reinforcements disregards autonomy, just as a coercive dictatorship does. Kultgen (1995) 70.
 Cass Sunstein, Richard Thaler, Libertarian Paternalism is not an Oxymoron, 70 University of Chicago Law Review 1159 (2003) 1166. For more examples of libertarian paternalism, see Cass Sunstein, Richard Thaler,Nudge: Improving Decisions About Health, Wealth and Happiness (London: Penguin, 2009). For a collection of ‘nudges’, see http://nudges.org/ (05.05.2011)
 It is possible to argue that a ‘seed’ of interference with autonomy is still present in these cases. Consider the parental analogy in the case of information giving. Due to the hierarchical relation between parents and children, even factual information will be received with more attention by children. Similarly, the advice of the surgeon general on the harmful consequences on smoking is supposed to have more deterrent effect than a simple warning without medical authority. Kultgen (1995) 70.
 ‘The definition of paternalism does not require that the paternalistic agent be successful in achieving good for the subject.’ Kleinig (1983) 76. ‘It is not essential that it achieve that end.’ Archard (1990) 37. See also Pope (2004) 697. It is important to keep in mind that intended or unintended consequences can still play an important role in the justification of paternalism.
 ‘Do you not know that your bodies are temples of the Holy Spirit, who is in you, whom you have received from God? You are not your own; you were bought at a price. Therefore honor God with your bodies.’ 1 Corinthians 6 19-20.
 The other category is the category of ‘hurts’. Hurts be defined as non-harmful physical and mental pains (e.g. twinges, aches, stitches, ‘heartache’, grief, despair, etc.) or physical discomforts (nausea, dizziness, tension, etc.). Feinberg (1986a) 47.
 Feinberg calls this category ‘presumptively nonblamable paternalism’. It consists of defending relatively helpless or vulnerable people from external dangers, including harm from other people when the protected parties have not voluntarily consented to the risk. Feinberg (1986c) 5.
 For the distinction between good and bad Samaritan laws and for an in-depth analysis of the moral justifiability of such regulations, see H. M. Malm, Liberalism, Bad Samaritan Law, and Legal Paternalism, 106 Ethics 4 (1995) and H. M. Malm, Bad Samaritan Laws: Harm, Help, or Hype? 19 Law and Philosophy 707 (2000).
 Feinberg calls this category ‘blamably nonbenevolent paternalism’. It means treating adults as if they were children, by forcing them to act or forbear in certain ways for the good of other parties, whatever their own wishes in the matter. Feinberg (1986c) 5.
 Gerald Dworkin uses the words pure and impure for direct and indirect paternalism. It can be easily mistaken with Feinberg’s category of mixed and unmixed paternalism which is referred to as pure and impure paternalism in John Kleinig’s terminology. Gerald Dworkin (1983) 22; Feinberg (1986c) 9; Kleinig (1983) 12.
 Gerald Dworkin argues that indirect paternalism requires stronger justification, because it involves third-parties ‘who are losing a portion of their liberty and they do not even have the solace of having it done in their own interest’. Others claim that indirect paternalism is not genuinely paternalistic because such limitations can be subsumed under the harm-to-others principle: by preventing cigarette manufacturers to sell cigarettes, we prevent them from causing illness to others in the same way we prevent factories from releasing pollutants into the air. Dworkin (1983) 22; Michael Bayles, Criminal Paternalism, in J. Roland Pennock, John W. Chapman (eds.), Nomos XV: The Limits of Law, 174-188.
 Joel Feinberg, Legal Paternalism, in Joel Feinberg, Rights, Justice and the Bounds of Liberty, (1980) 118. Some argue that soft paternalism is not even paternalism because the harm is not self-inflicted but comes from ‘external’ factors to the subject’s will (incapacity, lack of information, etc.). As Beauchamp notes ‘weak paternalism is not paternalism in any interesting sense since it is not a liberty-limiting principle independent of the harm-to-others principle’. Tom L. Beauchamp, Paternalism and Bio-Behavioral Control, 60 The Monist 62 (1976) 67. See also Feinberg (1986c) 13; Kleinig (1983) 8.
 A distinction must be made between consent to harm and consent to paternalistic intervention. Consenting to harm transforms the harm-to-others to a harm-to-self issue. It is possible to argue that external harm becomes voluntarily inflicted self-harm through consent and limitations introduced in such situations qualify as unjustified (hard) paternalism. The central question is where the limits of the volenti non fit iniuria principle are located. For an overview of the issue of consent, see Tom O’Shea (lead author), ‘Consent in History, Theory and Practice,’ Essex Autonomy Project Green Paper Report (University of Essex: Essex Autonomy Project, 2011) [available:
]. See also Vera Bergelson, ‘The Right to Be Hurt: Testing the Boundaries of Consent,’ 75 The George Washington Law Review 165 (2007) and Feinberg (1986c) 172-269.
 Kultgen and VanDeVeer calls this type of consent ‘current consent’. Kultgen (1995) 115; VanDeVeer (1986) 49. The lack of interference with autonomy leads some commentators to conclude that such cases are not cases of paternalism at all. Pope (2005) 685; Kultgen (1995) 116.
 Gerald Dworkin (1983) 28ff. For the discussion of Dworkin’s theory, see Feinberg (1986c) 135ff; VanDeVeer (1986) 70ff; Kultgen (1995) 122; Pope (2005) 698ff. Rawls also applies the hypothetical consent scheme to the question of paternalism: ‘Others are authorized and sometimes required to act on our behalf and to do what we would do for ourselves if we were rational, this authorization coming into effect only when we cannot look after our own good. Paternalistic decisions are to be guided by the individual’s own settled preferences and interests insofar as they are not irrational, or failing a knowledge of these, by the theory of primary goods.’ John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1999) p. 219 quoted by Pope (2005) 700. Dworkin abandons the hypothetical consent scheme in his later book and tries to justify paternalistic interventions with nonpaternalistic reasons (e.g. additional social security costs). Dworkin (1988) 125.
 Eyal Zamir, The Efficiency of Paternalism, 84 Virginia Law Review 229 (1998) 240. One way to reconcile paternalism with actual preferences theory rests on the idea of second-order preferences. People have preferences with regard to their own preferences. Second-order preferences may be paternalistically promoted to overcome first order preferences.
 Dan Brock, for example, places a rationality constraint on desire theories to prevent ethical relativism. Rationality is supposed to steer his theory in the direction of ideal theories but Brock admits that ‘different conceptions, and so constraints, of rationality are possible’. Brock (1983) 251.
 Douglas Husak aptly formulated: ‘It also is fair to say that most critics of paternalism have abandoned the attempt to formulate utilitarian objections to paternalism. It seems beyond dispute that many paternalistic interferences promote the good or welfare of the agent who is coerced, and do so without introducing disadvantages that outweigh the benefits. John Stuart Mill’s On Liberty is thought to contain several utilitarian arguments showing that a great number of paternalistic interferences are unjustified. Yet the prospects of formulating a general utilitarian case against paternalism appear so remote that many philosophers who combine sympathy for Mill with an anti-paternalistic bias are prepared to read much of On Liberty as a curious departure from utilitarianism.’ Husak (1980) 27.
 The US Supreme Court derived the right to privacy via a series of rulings, including issues of contraception [Griswold v Connecticut 381 US 479 (1965)], interracial marriage [Loving v Virginia 388 US 1 (1967)] and abortion [Roe v Wade 410 US 113 (1973)]. The German Constitutional Court derived the ‘general personality right’ from the interpretation of human dignity.
 For the moral dilemmas surrounding the issue of smoking regulations, see Robert E. Goodin, ‘Permissible Paternalism: Saving Smokers from Themselves,’ in Hugh LaFollette (ed.), Ethics in Practice: an Anthology, (2002) 308; Thaddeus Mason Pope, ‘Balancing Public Health against Individual Liberty: The Ethics of Smoking Regulations,’ 61 University of Pittsburgh Law Review 419 (2000).
 Although smoking imposes serious costs on society (in the form of health care costs and lost future earnings), its net economic impact seems to be positive due to increased tax income. Ibid. 444.
 To date, 12 EU countries have comprehensive smoke-free laws (i.e. complete bans in all public places including bars and restaurants) in force: Ireland (2004); Italy (2005); Sweden (2005); United Kingdom (2007); Finland (2007); Estonia (2007); Lithuania (2007); The Netherlands (2008); France (2008); Latvia (2010); Spain (2011); Greece (2010, problems with implementation). See www.smokefreepartnership.eu (04.05.2011)
 The prohibition came into force earlier in Wales (2 April 2007). Smoking was banned separately in Northern Ireland and Scotland by the Smoking (Northern Ireland) Order 2006 and by the Smoking, Health and Social Care (Scotland) Act 2005. They took effect 30 April 2007 and 26 March 2006, respectively.
 R (G) v Nottinghamshire Healthcare NHS Trust  EWHC 1096
 Ibid. para. 50. The Counsel explicitly cited John Stuart Mill’s harm principle: ‘The only part of the conduct of anyone, for which he is amenable to society, is that which concerns others. In the part that merely concerns himself, his independence is, of right, absolute. Over himself, over his own body and mind, the individual is sovereign’.
 R (G) v Nottinghamshire Healthcare NHS Trust  EWHC 1096, para. 100.
 For the positive obligation of Member States to protect from third-party interventions, see the seminal cases of McCann and Others v UK  21 ECHR 97, López Ostra v Spain  ECHR 46 and Guerra and Others v Italy  ECHR 7.
 The Television Without Frontiers Directive (Directive 89/552/EEC amended by Directive 2007/65/EC) banned all tobacco advertising and sponsorship on television in the European Community. This ban was extended by the Tobacco Advertising Directive (2003/33/EC) to cover other forms of media such as the internet, print media, radio and sports event. However, the directive does not prohibit indirect advertising on merchandising products, billboards or in local cultural and sporting events. An attempt (Directive 98/43/EC) to prohibit these was struck down by the European Court of Justice in its judgment Germany v. Parliament and Council, C-376/98  ECR I-8419. The ban of such advertisements was considered to be unnecessary for the free movement of goods and free provision of services within the EC since these products and services do not have cross-border effects.
 Directive 92/79/EEC on the approximation of taxes on cigarettes; Directive 92/80/EEC on the approximation of taxes on manufactured tobacco other than cigarettes; Directive 95/59/EC on taxes other than turnover taxes which affect the consumption of manufactured tobacco. The latest amendment on excise duties was adopted by Directive 2010/12/EU which foresees a gradual increase in the EU minimum taxation levels by 2014.
 The provisions of the first and second Labelling Directives (Directives 89/622/EEC and 92/41/EEC) were deemed to be insufficient due to the warnings’ small size and lack of visibility. The currently valid European regulations are based on the Directive on Packaging and Labelling of Tobacco Products (2001/37/EC). See, for example: www.who.int/tobacco/training/success_stories/en/european_community_labelling.pdf (10.05.2011)
 Conceptual confusion often surrounds the meaning of the term ‘euthanasia’. Distinction can be made between voluntary, nonvoluntary and involuntary forms of euthanasia. Involuntary euthanasia is imposed on the subject against his or her will or without his or her consent. It qualifies as murder and it is rarely discussed, let alone defended by anyone. In the case of nonvoluntary euthanasia, the patient is not mentally competent to make an informed decision (due to, for example, being in a state of coma). Passive euthanasia entails the withholding of treatment, while active euthanasia entails the use of lethal substances or forces to kill. See, for example Feinberg (1986c) 345. See also Peter Singer, Practical Ethics (Cambridge: Cambridge University Press, 1979) 128.
 Pretty v United Kingdom , 29 April 2002, application no. 2346/02. More exactly, the case qualifies as assisted suicide since Mrs. Pretty intended to commit suicide with the assistance of her husband. Voluntary active euthanasia involves an external person’s direct contribution to the patient’s death, while in cases of physician assisted suicide (PAS), the patient is the ‘final link’ in the causal chain and the doctor merely assists in the process of suicide (e.g. by providing the necessary drugs). From the viewpoint of criminal law, assistance in suicide is separately criminalized compared to active euthanasia, which dogmatically qualifies as homicide.
 Article 3 can involve positive obligations in other situations. In D v. the United Kingdom, a non-British citizen, suffering from AIDS, was threatened with removal from the UK to his home country (Saint Kitts) where no effective medical treatment was available for his illness. The European Court ruled that the removal of the person would amount to degrading treatment since