Paternalism,respect and dialogue

Introduction

Although liberal thinkers have traditionally rejected state paternalism, recent studies on human irrationality in psychology and behavioural economics have opened up a space for a defence of paternalism that appears congruent with liberal commitments. According to this defence, since human beings are often prey to systematic cognitive biases and reasoning errors, the state may paternalistically interfere with citizens to help them make better choices, without violating their rights or respect for their preferences. This defence of paternalism is not only used to justify libertarian paternalism (Sunstein and Thaler, 2009) but is also sometimes used to justify coercive paternalism which is the main focus of this article (Conly 2012; De Marneffe 2005, 120; Van de Veer 1988, 438; Dworkin 1972; Rawls 1971, 249). ¹

Yet, some liberal critics of paternalism maintain that paternalistic policies are nonetheless insulting and disrespectful to citizens, who ought to be treated as moral equals capable of pursuing their own conceptions of the good (Quong 2010; Shiffrin 2000; Anderson 1999, 301–302, 330; Cornell 2015; Kleinig 1983, 38; Groll 2012). These critics argue that independently of cognitive irrationality or ignorance that may undermine citizens’ abilities to make the best choices for themselves, the disrespect that paternalism conveys always constitutes a strong presumption against paternalistic laws in a liberal state where all citizens ought to be treated as moral equals.

In this article, I argue that both the rationality-based defence of paternalism and the respect-based objection to paternalism are problematic for a similar reason: both views treat citizens disrespectfully because they conceptualize the citizens’ interests from the wrong standpoint. According to the rationality-based defence of paternalism, paternalistic interventions are compatible with respecting citizens, as long as they promote an idealized conception of rational choice that is determined from the third-person standpoint of experts ² and policymakers. But without presupposing a specific ranking of values, I argue that idealization alone cannot yield a determinate conception of an ideally rational choice. In fact, I argue that attempts to individualize an idealized conception of interest to fit an individual’s particular dispositions and preferences also fail. Therefore, I claim that coercing citizens on the basis of an idealized conception of rational choice is bound to be a disrespectful imposition of a particular ranking of values on citizens.

By contrast, the respect-based objection to paternalism assumes that simply accepting the citizens’ conceptions of their own interests, as defined from their own first-person standpoint, is the proper way to respect citizens as moral equals. In other words, this view assumes that a laissez-choisir (literally meaning, ‘let them choose’) policy on self-regarding matters is the proper way to respect citizens as moral equals who can effectively advance their own ends. ³ But for reasons I will explain, such a laissez-choisir policy actually amounts to disregarding citizens’ conceptions of their own interests, as opposed to respecting them as the presumptively rational beliefs of fully responsible agents.

In light of these criticisms, I seek to establish the following two claims. The first claim is that in order for citizens to be genuinely respected as moral equals, the correct conception of their interests must be determined from a second-person standpoint ⁴ (or, joint standpoint) that is forged through a process of mutual dialogue between the relevant experts, policymakers and the citizens themselves. In any proper dialogue or argumentation, all participants must take seriously everyone’s claims as presumptively rational claims and evaluate their soundness by exchanging reasons with each other, as opposed to simply accepting or rejecting them out of hand. This constitutive dynamic of argumentation compels every participant to seriously consider the validity of the reasons others adduce for their conception of interests, which I will argue, is precisely what is required to treat others as persons of equal moral standing.

The second claim, which builds on the first, is that properly respecting citizens as moral equals in the above sense opens the door to paternalism, rather than closing it off as most liberal thinkers assume. If treating citizens as moral equals requires understanding their conceptions of interests through a mutual dialogue with them about the correct interpretation of their interests, a possibility opens up for policymakers and experts to criticize and thereby initiate a revision in citizens’ conceptions of their own interests and their corresponding choices. This indicates that engaging others in a rational argumentation about their own interests is paternalistic, contrary to the traditional liberal view. Yet, unlike other kinds of paternalistic behaviour, I argue that this dialogical form of paternalistic intervention is a byproduct of properly respecting others as moral equals. In light of this argument, I conclude that if policymakers and experts are concerned about promoting citizens’ welfare, they should seek to arrive at a shared understanding of what citizens have most reasons to choose through rational discursive means, rather than unilaterally coercing them without regard to their own beliefs about what’s good for them.

This article is structured into five sections. First, I sketch out a general definition of coercive paternalism and explain the Equal Respect objection (i.e. what I referred to as the ‘respect-based objection’ above) against it, which says that paternalistic policies undermine respect for citizens’ status as moral equals who are capable of independently advancing their own interests. ⁵ In the second section, I show that the rationality-based defence of coercive paternalism cannot overcome the Equal Respect objection, because the choice between different conceptions of an ‘ideally rational choice’ is bound to be arbitrary. In fact, I argue that a suitably individualized conception of rational choice that is relativized to each individual’s character, disposition and stable preferences is also problematic. The third section is dedicated to showing that contrary to the assumptions of many liberal thinkers, laissez-choisir anti-paternalism is also incongruent with respecting citizens as moral equals because it involves disregarding, as opposed to respecting, citizens’ conceptions of their own interests as the beliefs of fully responsible agents.

In light of this double critique of coercive paternalism and laissez-choisir anti-paternalism, in the fourth section, I explain why genuinely respecting citizens as moral equals requires policymakers and experts to understand citizens’ conceptions of their own interests by engaging in a mutual dialogue with them about why their reasons are (or are not) sufficient grounds for their conception of interests. Then, I explain why the very act of engaging in this dialogical process of understanding citizens is a distinctive paternalistic intervention in its own right. On the basis of this argument, I conclude in the fifth section that paternalistic behaviour and respect for others are not always incompatible, contrary to what liberal thinkers generally assume.

The Equal Respect objection to paternalism

‘Coercive paternalism’, as I shall use the term throughout this article, refers to coercive interference with the liberty of individuals for their own good without their consent. While there are many different ways to treat others paternalistically without using coercion – say, by withholding certain information from others to protect them from harm (Buchanan 1978; Gert and Culver 1976) or nudging others into making healthier, more prudent choices (Sunstein and Thaler 2009) – this article will focus primarily on coercive paternalistic interference by the state.

By ‘coercive’ interference, I refer to any interference that involves restricting the agent’s option set without their consent, by removing or replacing one or more of the agent’s existing options. When the government bans hazardous waste disposal in public waters or slaps a prohibitive tax on the import of certain goods, it is a case of coercive interference. When such coercive interference is specifically directed at a person (P) to prevent harm to P themselves without their consent, it is an instance of paternalistic coercion towards P. Let me unpack this definition a bit further.

When P is coerced for reasons other than advancing P’s own welfare or interests, the coercion is not paternalistic. Although Seana Shiffrin claims that paternalistic action refers to any interference with matters that lie within one’s ‘legitimate domain of control’ which is typically larger than the pursuit of one’s own welfare, Shiffrin’s definition is too broad (Shiffrin 2000, 216–217). On Shiffrin’s definition, when a passerby interferes with a parent who is scolding their child too harshly, such interference is paternalistic towards the parent since the parent has legitimate control over their child’s upbringing. But notice, since the interference is aimed at protecting the child from harms caused by their parent, the interference is not really paternalistic towards the parent at all. Therefore, defining paternalistic interference as interference with any decision that is within another’s legitimate domain of control is too broad. In order for an action to count as ‘paternalistic’ towards P, that action must be directed at preventing harm to P themselves.

In addition, a paternalistic action must be undertaken in the absence of the recipient’s consent. When Ulysses’s crewmen tied Ulysses to the mast of the ship upon his request, the crewmen’s action was not paternalistic since they were merely carrying out the wishes of their leader. Whenever one voluntarily requests or consents to some interference prior to the interference, it is an instance of self-binding or precommitment, rather than paternalism.

By contrast, when someone removes or replaces one of your options for your own sake without your consent, their action is paternalistic even if you subsequently approve of what they did. If I throw away all the sweets in the pantry for the sake of your health before you know about it or have the chance to think about it, my action is paternalistic even if you later find out and agree that my action was ultimately for the best (Shiffrin 2000, 214). And of course, throwing out all the sweets in the pantry against your will, despite your protests and wails, is clearly paternalistic. These considerations suggest that ‘coercive paternalistic interference’ is best defined as a removal or replacement of another person’s options for their own good, in the absence of their prior or present consent. ⁶

A major objection against paternalism, thus defined, is that although paternalistic action is ostensibly intended to advance another person’s own good, it is actually motivated by one’s insulting judgement that since the other is too obtuse or weak-willed to do what’s best for themselves, they should be compelled by a third party to do what’s best for themselves. Indeed, this is why paternalism has been often criticized as being incompatible with the liberal commitment to ‘respecting others as equals’.

‘Respecting others as equals’ in this context refers to the Rawlsian or Kantian idea of treating others as persons who have the basic capacity to rationally set and pursue their own goals. According to Rawls, one of the defining features of moral personhood that entitles a person to equal justice is the capacity to rationally form and pursue their own goals and values (Rawls 2005, 19; Rawls 1999[1971], 442). On this view, treating someone paternalistically amounts to denying them the status of a moral equal, since such treatment implies that they lack the basic rational capacities characteristic of all moral persons. Drawing on this Rawlsian view, Jonathan Quong claims that treating others paternalistically is wrong because it denies them the status of moral equals.

Liberal political philosophy ought to begin with a conception of ourselves as free and equal. Following Rawls, we can characterize citizens as free and equal in virtue of their possession of two moral powers: a capacity for a sense of justice, and a capacity for a conception of the good… Paternalistic actions, however, appear inconsistent with this conception of our moral status. This is true because paternalism involves one person or group denying that another person or group has the necessary capacity, in a given context, to exercise the second of the two moral powers: the capacity to plan, revise, and rationally pursue their own conception of the good. To treat a person paternalistically is thus (at least, temporarily) to treat that person as if he or she lacks the second moral power. This means there is always a strong moral reason not to treat a sane adult paternalistically, and this is why paternalism is prima facie wrong (Quong 2010, 101).

Since Quong’s objection above hinges on how paternalistic treatment denies the subjects respect for their equal moral status, I will call this the Equal Respect objection to paternalism. The force of the Equal Respect objection becomes apparent when we consider what happens to the relational dynamic between two agents when one treats the other paternalistically.

Suppose that you decide to regularly attend an expensive yoga and wellness centre for the sake of your health. As your old friend, I believe that I know you very well and decide that you have been thoroughly misled by the fancy advertisements and media hype surrounding the new fashionable wellness trend. To save you from wasting your money and time, I go to the wellness centre and terminate your membership on your behalf without your consent. In this case, my behaviour is certainly paternalistic towards you, and you would understandably feel quite offended at my patronizing behaviour. Even if you end up agreeing that what I did was for the best, you would still have good reasons to feel insulted and disrespected. Not only am I implying that you are too obtuse or ignorant to recognize what’s best for yourself, I am also implying that I have superior abilities than you in judging what’s truly best for yourself. Therefore, my behaviour expresses contempt and lack of respect for your decision-making ability (regardless of whether you actually feel insulted). ⁷

Hence, treating a person paternalistically, by definition, appears incompatible with respecting them as a person of equal status. Moreover, the offence to the paternalized subject’s sense of dignity and equal status stands, regardless of the subject’s actual abilities or benefits of the intervention. Therefore, overcoming the Equal Respect objection seems to be a necessary condition for the justification of any paternalistic action.

The rationality-based defence of paternalism

According to some proponents of paternalism, however, the Equal Respect objection is misplaced because it is not disrespectful to override someone’s choice when that choice is caused by cognitive irrationality, error or ignorance. On this view, choices that are worthy of respect are expressions of the agent’s ‘genuine’ preferences, which are the preferences that the agent would have if they were perfectly rational and well-informed. ⁸ Since this view supports paternalistic interference only with choices that are in some sense irrational or mistaken, I will call this the ‘rationality-based defence of paternalism’.

According to the rationality-based defence of paternalism, it would be compatible with respect for citizens as moral equals to restrict the choice not to wear seatbelts, the choice to spend now rather than save for later or the choice to take addictive and dangerous drugs for immediate pleasure, since these choices are usually motivated by some glitch in the chooser’s rational system. ⁹ But for these glitches in their rational system, people would agree that banning these choices would be best for their own interests. Therefore, the argument goes, it is morally permissible to ban these choices for the sake of people’s own good. This version of the coercive paternalist view relies then, on something akin to a hypothetical consent-based justification of paternalistic interference (Conly 2012; De Marneffe 2005, 120; Van de Veer 1988, 438; Dworkin 1972; Rawls 1999[1971], 249). Sarah Conly describes this view most straightforwardly. ‘We should save people from doing things that are gravely bad for them when they do that only as a result of an error in thinking… Sometimes the only way to stop someone from making a terrible mistake is to intervene and prevent him from choosing freely’ (Conly 2012, 3–4). The core argument of coercive paternalism then, is that irrationality or mistaken judgement makes paternalistic interference permissible. ¹⁰

As it turns out, however, this rationality-based defence of coercive paternalism cannot work without relying on an arbitrary criterion of an ‘ideally rational choice’. Consider the case of professional boxing, which has for a long time attracted heated calls for a ban from various countries due to the alarming incidence of serious brain damage and death it produces, as well as its ruthless exploitation of young men from socioeconomically disadvantaged backgrounds (World Medical Association 2005; Simon 1991, 59–60; Wacquant 2001).

According to Loïc Wacquant’s ethnography of professional boxers in the South Side of Chicago, most ordinary boxers are keenly aware of their own exploited condition, as well as the current lack of proper safety regulations. Indeed, professional boxers are so intensely aware of their own exploited condition that they describe themselves as whores, slaves and stallions – all bodily commodities to be sold and bartered on the market (Wacquant 2001, 182). Yet, surprisingly enough, most boxers freely choose to remain in the boxing business without raising much complaint or clamouring for change. This is because, as Wacquant’s ethnography attests, professional boxers have their own distinct reasons which in their eyes outweigh their reasons to stop boxing. According to Wacquant’s ethnography, most boxers believe that continuing with boxing is their best choice because (1) exploitation is an inescapable fact of life; (2) that they, unlike workers in other professions, do not have to ingratiate themselves with the higher-ups to win the game and become a champion and (3) that among millions of boxers, he alone will be the exception to the general rule of extortion and achieve fame and success (Wacquant 2001, 186–189).

In response, physicians, ethicists and policymakers (i.e. the ‘experts’ for short) who support a ban might argue that, as a matter of fact, most ordinary professional boxers are not sufficiently aware of the high risks of serious brain damage, the egregious level of exploitation and injustice they are putting up with or the de facto lack of control and independence they enjoy in their profession. If the boxers were ideally rational agents with full awareness of the health risks and no susceptibility to irrational biases or ignorance, like the rational agents in Rawls’s Original Position, then the boxers would recognize that their reasons are not sufficient grounds for continuing with boxing despite its negative effects, or so they might argue. But unfortunately for the experts, this idealization (or hypothetical consent) strategy does not necessarily warrant their desired conclusion, that is, that there are no sufficient reasons for anyone to pursue professional boxing.

Suppose that Jim is an ordinary professional boxer in the real world and call his ideally rational counterpart, Jim". Unlike Jim, Jim" is not susceptible to cognitive errors or biases at all. Also, unlike Jim, Jim" does not know how much tolerance he has for physical pain, how much he enjoys risk-taking or the particular features of the sport of boxing, how poor he is or what alternative jobs he has access to. All that Jim" knows are general sociological, psychological and economic facts about the world, just like the rational agent in Rawls’s Original Position. In such a condition, would Jim" choose (1) a society where boxing is banned or (2) a society where boxing is legally permitted?

Given the high risks of permanent brain damage and death, low pay, exploitative working conditions and harsh training requirements, one might think that a perfectly rational Jim" would obviously choose a society where boxing is banned. In fact, Rawls explicitly states that a rational agent in the Original Position would choose to have laws that could protect himself (viz. himself as he would turn out in the real world) from the ‘weakness and infirmities of reason and will in society’ (Rawls 1999[1971]: 219). Thus, it might seem self-evident that Jim" would choose to have laws that could prevent himself from choosing boxing out of ignorance or misestimation of risk.

But notice that the opposite case could be made just as well. A perfectly rational Jim" might also think that since he does not know how much he would be risk-averse in the real world, he would be best off, all things considered, to have the extra career option of boxing. After all, he might just turn out to be an enthusiast for boxing, despite its attendant risks and low pay, just as some danger junkies enjoy extreme sports and wildlife adventures even if no one pays them. So a cool-headed, perfectly rational Jim" might choose to be in society (2), based on the reasoning that his desire for boxing might turn out to be greater than his desire to avoid risk in the real world. Moreover, Jim" might also think, ‘Even if I don’t turn out to be such an enthusiast for boxing, I don’t have anything to lose by choosing to be in society (2), because I can always choose not to box if I turn out to be risk-averse’. So, depending on which particular sense of rationality Jim" adopts behind the veil of ignorance, Jim" could choose to be in either society (1) or (2). ¹¹

Jim’s case demonstrates that idealizing about what a perfectly rational agent would choose is insufficient to determine what is truly in the boxers’ interests. Unless there is a master principle that could determine which particular idealization is right and how far idealizations would have to go, various different idealizations as well as different outcomes are equally valid. But in fact, even the search for such a master principle is hopeless because it would lead to an infinite regress of idealizations. The choice between different idealized accounts of Jim’s interests thus ultimately depends on the idealizer’s particular ranking of different values: in this case, [the value of protecting oneself against expected harms and risks] versus [the value of having more options]. Those who rank the former value higher than the latter would decide that Jim’s ideally rational choice is society (1), whereas those who have the opposite ranking of values would decide that Jim’s ideally rational choice is society (2).

This means that how one chooses to idealize, as well as the outcomes one derives from idealizing, is determined in the end by one’s specific ranking of values. But since presupposing any specific ranking of values without an independent justification is question-begging, it follows that the choice between (1) and (2) is arbitrary. Consequently, standard forms of coercive paternalism are bound to be impositions of alien values and ends on citizens, even when the experts conceptualize the citizens’ (real) interests in terms of the idealized preferences of a perfectly rational agent.

One might suggest that we can avoid this problem by relativizing the notion of idealized preferences to each individual’s stable preferences, their own unique disposition and tastes and their ‘normal’ capacities of deliberation (see Van de Veer (1988, 75) for the relevant notion of ‘hypothetical individualized consent’). But in order for this individualized version of idealization to work, we would need an independent criterion that could enable us to distinguish between factors that hinder each individual’s ‘normal’ deliberative capacities on the one hand, and factors that are simply part of their stable preferences and character on the other.

But identifying such an independent criterion seems to be an impossible task. In the second hypothetical choice situation of Jim" above, for instance, it is not clear whether Jim"’s high tolerance for risk should be attributed to his innate thrill-seeking tendencies or to his lack of appreciation of the risks of grievous injury or death in boxing. Put another way, there seems to be no plausible method with which we can verify that the individual’s so-called stable preferences and dispositions are not themselves the product of irrational factors that vitiate their ‘normal’ capacities of deliberation: lack of sufficient reflection, rationally indefensible (but stubborn) beliefs or social conditioning, to name a few. But if this is so, then it is nearly impossible to identify a consistent and independent criterion that could distinguish between one’s stable preferences and character on the one hand, and factors that vitiate one’s ‘normal’ deliberative capacities on the other.

Furthermore, it is not clear how we might distinguish between ‘normal’ and ‘ideal’ deliberative capacities, and what sorts of factors should be regarded as hindrances to the latter but not the former. It would seem that on a plausible view, the difference between ‘normal’ and ‘ideal’ deliberative capacities would be only a matter of degree, not a difference in kind. But if this were true, then hindrances to ideal deliberative capacities would also be hindrances to normal deliberative capacities, with the only difference between the two sorts of hindrances being a matter of degree. This would make it extremely difficult, if not impossible, to distinguish between ideally rational preferences and normally rational preferences, since there is no non-arbitrary cutoff point above which factors that threaten ideal deliberative capacities suddenly become severe enough to become threats to our normal deliberative capacities as well.

This seems sufficient to show that as long as experts define the citizens’ interests from the third-person standpoint of a detached observer who does not share the perspectives of the citizens themselves, they cannot avoid the substantial risk of imposing alien values on citizens and thereby disrespecting them. More generally, these arguments show that paternalistic coercion carried out in the name of promoting other people’s ‘rational interests’ is impermissible, because in most cases, it involves imposing an arbitrary ranking of values on others.

Does equal respect require a laissez-choisir anti-paternalist policy?

So far, I have argued that even when citizens are susceptible to irrationality (as most humans are), coercive paternalism led by experts is still objectionable, because it does not take seriously enough the conception of the good that citizens endorse from their own standpoint. Does this mean that we should have a laissez-choisir (i.e. ‘let them choose’) policy in the domain of individuals’ voluntary, self-regarding affairs? In other words, should each individual’s own first-person standpoint always be privileged over that of others when it comes to matters concerning themselves?

The most plausible version of the laissez-choisir view is the soft paternalist view, which is the main focus of this section. ¹² According to the soft paternalist view, interference with individuals’ self-regarding conduct is permissible only when the conduct is substantially non-voluntary or when interference is necessary to verify that the conduct is voluntary (see Feinberg 1986; Mill 2003[1859]; Van de Veer 1988). The reasoning behind this view is that because a non-voluntary choice does not reflect the agent’s own will, it is not truly the agent’s own choice and is therefore a target of permissible interference. But unlike coercive paternalism which focuses on respecting the counterfactual will of an ideally rational agent, soft paternalism focuses on respecting the actual will of the existing agent, since a ‘voluntary’ choice refers to just what the agent would actually choose in the absence of coercion, factual ignorance or gross cognitive incapacity. ¹³

This means that on the soft paternalist view, others may not interfere with one’s voluntary choice, no matter how dangerous or counterproductive that choice may be to oneself. In Feinberg’s words, the soft paternalist ‘must argue that paternalistic reasons [concerning the agent’s own good] never have any weight on the scales at all’ (Feinberg 1986, 26). The roots of this soft paternalist view can be found in the basic liberal principle, proposed by John Stuart Mill, that the individual has sovereign decision-making authority over themselves (Feinberg 1986, 68; Mill 2003 [1859]). And it is this basic liberal principle that is at the core of the laissez-choisir view.

Contemporary liberal anti-paternalists, including Quong and Shiffrin, endorse some version of this basic liberal principle on the grounds that such a principle is required to respect other people’s agency, autonomy or status as moral equals (Quong 2010, 80–81; Shiffrin 2000, 213). The crucial, but perhaps surprising point I will argue is that in fact, respecting others as equals is conceptually incompatible with this basic liberal principle characteristic of the laissez-choisir view. This is because blindly accepting other people’s claims or beliefs about their own interests, as the laissez-choisir view demands, is not what respect for others requires.

A plausible interpretation of Rawls’s idea that citizens are equal in virtue of their possession of the two moral powers is that citizens are equal in virtue of their capacity to think and decide responsibly about political matters of justice, as well as personal matters concerning their own conception of the good (Rawls 1985, 233–234). But this implies that treating someone as a moral equal requires taking the claims they make from their own standpoint seriously and considering them as presumptively sound claims that could be rationally explained in terms that make sense. The important point which has been overlooked in the literature is that, taking a person’s claims seriously in this sense requires trying to understand the reasons behind that person’s claims, rather than just accepting those claims offhand.

Accepting a person’s claims offhand, without even attempting to understand their reasons for those claims, is not to treat that person as a moral equal who can make legitimate reason-based claims about what is right/wrong or good/bad. Instead, just accepting a person’s claims casually amounts to ignoring the intended meaning of their claims. This is because simply accepting a person’s claims without even trying to make sense of them is tantamount to treating those claims as a joke or the utterances of a small child or a deranged person. When someone is joking around, we do not take their words seriously enough to enquire about their underlying reasons or argue over them, so we just tend to go along with whatever they say. Similarly, when a small child or a deranged person says something, we do not try to argue out with them as to why they said those things or whether what they said is true. Without even attempting to understand and assess the validity of their claims, we get right down to the business of managing them, curing them or training them; in Peter Strawson’s terms, we regard them from an objective stance (Strawson 2008 [1964], 9).

By contrast, when we are conversing with people whom we consider our intellectual peers about serious matters which presumptively have true or false answers, such as the correct explanation of a person’s behaviour, moral duties or what is truly in someone’s best interest, we cannot (and do not) just go along with whatever they say. This is because in such serious matters, the truth or falsity of the other’s beliefs has direct implications for the truth or falsity of our own beliefs (Scanlon 2000, 74; Habermas 1985, 119). ¹⁴ If you happen to be right that Jim’s choice mainly stems from his genuine passion for boxing, then it would be wrong of me to presume that Jim’s choice stems from his desire to escape poverty or delusions brought on by falsely glamorized portrayals of boxing in the media. Moreover, you would be correct to conclude that Jim has important reasons to continue professional boxing, and I would be wrong to conclude that Jim does not have sufficiently good reasons to continue boxing. ¹⁵ But both of our beliefs cannot be simultaneously correct: there must be some causal explanation that explains why we have reached different conclusions. Either you or I may have missed some relevant information or reasoned incorrectly.

This is why whenever we are conversing with those whom we consider our equals about serious intellectual matters, we operate on the presumption that we should all reach the same conclusion about the matter at hand. And in case we fail to reach an agreement, we tend to balk at the discrepancy and try to find out which one of us has gotten it wrong by reviewing the evidence again and arguing the matter out with each other (Pettit and Smith 1996, 430–431). Even if our arguments cannot decisively settle the question of who is right or wrong at the end of the day, we would be compelled to find out why our judgements differ insofar as we take the other seriously as our moral equal – as a person who is equally capable of recognizing what’s good or bad, right or wrong.

In fact, if you denied that the truth or falsity of other people’s judgement has direct implications for the truth value of your judgement and refused to care about your judgemental differences, you would not be taking others seriously as moral equals who are capable of making sound, reliable judgements about themselves and the world (Habermas 1985, 132–133). If you take others seriously as your equals, you would care enough about their claims to endeavour to understand why their judgements differ from your own, rather than ignoring those differences as unimportant. Therefore, respecting others as your equals necessitates a dialogue or argument with them whenever you disagree with them about a serious moral or intellectual matter.

Although respect requires this sort of dialogue with anyone we might get into a serious disagreement with, this point usually applies to those who stand in a specific relationship to us (e.g. colleagues, family and friends) since they are the ones we normally consider appropriate interlocutors for serious conversations about intellectual or moral matters. In fact, those who have prior duties to care about other people’s welfare, such as elected representatives, physicians or caretakers have downstream duties of respect to try to understand other people’s reasons for action through dialogue when they act in ways that are (seemingly) counterproductive or harmful to themselves. ¹⁶ But in any case, the bottom-line is that once you get involved in a serious conversation with others about the reasons behind their judgements, then respect requires you to try to understand why their judgements diverge from your’s through sustained discursive exchange.

But if it is true that respect entails making an effort to understand others’ reasons for their judgements (instead of simply accepting their judgements), the following question arises: what does understanding others’ reasons require, and why does it require something other than simply accepting their reasons? To grapple with this question, first note that a ‘reason’ for action, by definition, is a consideration that purports to provide a justification or rationale for doing something in a particular situation (Habermas 1985, 115–116; Scanlon 2000, 73–74). This means that the boxers’ judgement about their reasons for continuing with boxing is not merely an expression of their subjective preferences, but rather, a judgement about what is right or justified for them to do in their circumstances.

But matters about what one is justified in doing (or not doing) in a particular situation is something we presume has a right or wrong answer, unlike one’s preference for chocolate over vanilla ice-cream. The idea is not that there is a right answer that is empirically verifiable, but instead, that we make the presumption of a right answer that is reasonably acceptable to all whenever we argue over some serious normative question. ¹⁷ In fact, this presumption is what makes it possible for us to argue about normative questions in the first place, like whether it is justified to prohibit boxing or the recreational use of drugs to prevent harm to people. If we did not make such a presumption, there would be no point in arguing over such normative questions to begin with: we should just agree to disagree, like we do over ice-cream preferences. But we do in fact meaningfully argue over various normative questions every day.

To be sure, disagreements about serious normative matters (such the moral permissibility of prohibiting boxing or drugs) involve disagreements about individual preferences to some extent, but they cannot be collapsed into purely preference-based disagreements because normative questions deal with which preferences ought (or ought not) to be respected in a cooperative society for mutual benefit, given the balance of all the reasons at stake. And disagreement about what ought to be done essentially involves a conflict between different presumptions about what the right thing to do is. But there is no question about what ought to be done when it comes to purely preference-based disagreements: there is only a divergence between what you like and what I like. This shows that unlike purely preference-based disagreements, disagreements about serious normative matters are based on individuals’ conflicting presumptions about the ‘right’ thing to do.

Now the fact that the boxers’ reasons are judgements about what is ‘right’ or ‘justified’ for them to do implies that in order to understand the boxers as rational agents who are capable of making sound judgements about what is right or wrong, you must try to make sense of their reasons as considerations that, in some sense, make their choice to continue boxing a ‘good’ or ‘right’ choice in their circumstances. Otherwise, you can only dismiss their reasons as nonsense.

But in order to understand the boxers’ reasons properly in the above sense, you cannot blindly accept their reasons. Instead, you need to know (1) why the boxers take their reasons to be a good justification for their choice to continue boxing and (2) why they are right (or wrong) to take their reasons as a good justification for their choice. This implies that understanding others’ reasons necessarily involves judging their reasons as sound or unsound, as opposed to blindly accepting them. In Habermas’s words,

The interpreter would not have understood what a “reason” is, if he did not reconstruct it with its claim to provide grounds; that is, if he did not give it a rational interpretation in Max Weber’s sense. The description of reasons demands eo ipso an evaluation, even when the one providing the description feels that he is not at the moment in a position to judge their soundness. One can understand reasons only to the extent that one understands why they are or are not sound, or why in a given case a decision as to whether reasons are good or bad is not yet possible. An interpreter cannot, therefore, interpret expressions connected through criticizable validity claims with a potential of reasons (and thus represent knowledge) without taking a position on them. And he cannot take a position without applying his own standards of judgment, at any rate standards that he has made his own… But if he would not so much as enter upon a systematic assessment,…, he would not be able to treat reasons as that which they are intended to be. In this case, the interpreter would not be taking his subject seriously as a responsible agent (Habermas 1985, 116, 133). ¹⁸

As Habermas explains above, in order to understand other people’s claims about their own reasons qua the claims of equally rational agents, you need to at least try to make sense of why they think their reasons are ‘good’ reasons for their behaviour. If you blindly accepted their reasons without even trying to make sense of why their reasons are good (or bad) justifications for their (seemingly irrational) behaviour, you would be effectively ignoring the intended meaning of their reasons, and thereby disrespecting them.

In this sense, blindly accepting another agent’s claims is analogous to merely ‘going along’ with an incompetent agent’s utterances: both actions boil down to ignoring the agent’s reasons. Genuine respect for others requires taking others’ claims about their interests seriously and trying to understand the reasons why they believe their action is conducive to their interests when you think it is not.

The claim here is not that we have a moral duty to try to convince others that they must be wrong (and that we are right) whenever we disagree about what they should do. Rather, the claim is that if one operates on the presumption that others are moral equals who can make sound judgements about what is good or right for them to do, then one should (and would) expect others to recognize the same considerations that have convinced oneself that what they are about to do is wrong, bad or pointless. And if one starts out with this expectation of agreement among presumptively equal rational agents, then one would be naturally motivated to wonder and ask about the other’s reasons when, for instance, one sees the other about to do something that seems self-destructive or irrational. It is in this sense that the presumption of the other’s equal status entails engaging them in dialogue about their reasons when we see them about to do something that appears strange or harmful to themselves.

Now an anti-paternalist critic might object that it is misleading to suggest that ‘accepting’ a competent person’s claims on trust is analogous to merely ‘going along’ with what an incompetent person says. The critic might argue that respect for another agent silences or excludes all considerations in favour of interfering with their choice to X, provided the agent has voluntarily chosen to X in conditions free of coercion, incapacity or factual error (Feinberg 1986; Groll 2012; Enoch 2016, 46). On this view, respect requires deferring to the agent’s decision, no matter what your understanding of their reasons for that decision turns out to be. Based on such a view, the critic might argue that accepting the agent’s claims about her own interests is indeed a sign of respect. But there are several reasons why this objection is untenable.

First, this objection is based on a fundamental misconception of what respect for others as equals entails. When you observe someone doing something that seems self-destructive, pointless or outré, then your first natural reaction would be to try to understand why they are doing such a thing by asking them about their reasons for doing that thing. But notice that you tend to respond in this way only because you start off with the presumption that the other is just as competent and rational as yourself. If you presumed that the agent lacks competence or is unquestionably less competent than you are, you would not even bother to understand their reasons. Since respect for others qua equal rational agents is based on this presumption of equal competence, it follows that respect for others entails a disposition to try to understand why they are doing what they are doing, rather than simply deferring to them.

Second, it is implausible to assume that respect requires bracketing your beliefs and expectations about what is rational or sensible for others to do – even when it primarily concerns their self-regarding matters. If this assumption were true, then respect would make it morally impermissible, for instance, to judge that voluntary amputation of one’s healthy limbs or compulsive handwashing are irrational behaviours, possibly stemming from an underlying psychiatric condition. ¹⁹ In fact, this assumption would make it morally impermissible for even licensed physicians to render any psychiatric diagnosis based on unintelligible or erratic behaviour. But this is obviously implausible.

Third, this objection begs the question as to why the agent’s autonomy should trump all other values, including the agent’s welfare. Moreover, this objection is based on a view that is objectionably perfectionistic, since not all reasonable people share the view that autonomy should trump everything else. In fact, even Joel Feinberg who develops the most compelling defence of the right to autonomy concedes that individuals’ rights over themselves and their property may sometimes be justifiably infringed for the sake of other important considerations (see Shafer-Landau 2005, 188–189; Feinberg 1970, 75).

Finally, this objection entails some extreme consequences that even committed anti-paternalists would reject. For instance, on the proposed view, the mere fact that someone has chosen to end their life should bar anyone from interfering with their decision. This implies, for instance, that it is impermissible to interfere with the choice of a competent person who is severely depressed or anorexic (but otherwise healthy) to take their own lives. ²⁰ But most people, including committed anti-paternalists, would reject such a view as unnecessarily extreme and harsh. Hence, the proposed view is implausible. The claim I am making here is perfectly consistent with recognizing that individuals have a basic right to act against their own interests if they wish, since the claim is only that there is no sufficient basis for treating the right to autonomy as a moral trump card.

If the arguments so far are correct, it seems fair to conclude that the anti-paternalist policy of just accepting other people’s claims about their own interests regardless of their reasons for those claims is equivalent to a refusal to recognize others as our moral equals. Contrary to the traditional liberal views of Quong, Shiffrin, and Feinberg then (Quong 2010, 100–101; Shiffrin 2000; Feinberg 1986, 69), genuinely respecting citizens as moral equals does not require simply accepting the citizens’ first-personal account of their own interests. Quite the opposite: respecting others as our moral equals entails a disposition to actively engage them in dialogue about the reasons underlying their behaviour, with the aim of reaching a mutual understanding of their behaviour.

Towards dialogue: Reconciling respect and paternalism

If the arguments above are on the right track, then respecting others as our equals requires at least making an effort to understand their reason for action when their action seems irrational – as in the case of actions that are harmful to themselves. But this raises a question concerning what we need to do specifically in order to understand others’ reasons.

I have claimed above that understanding others’ reasons requires knowing why they regard their reasons as a sufficient justification for their action, and whether they are right or wrong to do so. But if understanding others in this manner simply requires asking for their reasons and knowing what caused them to do what they did, then it would seem that anti-paternalism might be compatible with respect for others, after all.

According to John Stuart Mill, for instance, it is permissible to intervene in another person’s self-harming choice, coercively if necessary, to ascertain that they are aware of the danger (Mill 2003 [1859], 165). This implies that Mill would certainly be in favour of intervening just to ask the agent why they are about to do something harmful to themselves. Feinberg and Quong, too, claim that intervening temporarily when the agent has no reasonable way of knowing the risks and dangers associated with their choice is not paternalistic and therefore permissible (Feinberg 1986, 159, 271; Quong 2010, 82). So if understanding others merely requires asking for and knowing what others take their reasons to be, then anti-paternalists could argue that their view allows for a genuine understanding of others’ reasons and is therefore compatible with respect for others.

But again, this thought is based on a misconception of what understanding other people’s reasons requires. Understanding other people’s reasons requires you to go far beyond simply asking for and knowing what their reasons are. Suppose that when you ask Jim the boxer why he chooses to persist with professional boxing despite its obvious disadvantages, Jim states that he chooses to stick with boxing because he would face the same kind of exploitation wherever he goes, and boxing offers the chance of succeeding with his own two fists unlike other professions. But merely knowing that these are Jim’s reasons is not sufficient for you to make sense of why Jim thinks those are ‘good’ reasons to choose boxing as an occupation. If anything, Jim’s reasons, at face value, appear to be clearly mistaken since other professions do not require putting up with a daily regime of gruelling physical training, permanent brain damage or the tyranny of boxing managers merely in exchange for a paltry wage and a very slim chance of success.

Instead, in order to understand why Jim, who is presumably just as sound-minded as anyone else, believes that his reasons are actually ‘good’ reasons for continuing with boxing, you need to know why, within the particular context that Jim is situated in, it makes sense for Jim to believe that his reasons are ‘good’ reasons. As Jürgen Habermas puts it,

In order to understand an utterance in the paradigm case of a speech act oriented to understanding, the interpreter has to be familiar with the conditions of its validity; he has to know under what conditions the validity claim linked with it is acceptable, that is, would have to be acknowledged by a hearer. But where could the interpreter obtain this knowledge if not from the context of the observed communication or from comparable contexts? (Habermas 1985, 115).

In order to understand Jim then, you must construct (within your own mind) the presumptive conditions of validity that operate within Jim’s context of action, so that relative to those conditions, Jim’s reasons appear to be sufficient grounds for his choice. Such presumptive conditions might consist of, for instance, a biased culture that glorifies violence; the collective myth that boxing, unlike other professions, uniquely fosters the virtues of self-reliance and independence; or a collective adaptation to exploitation and poor working conditions. In short, you might rationally explain that Jim came to accept his (seemingly implausible) reasons as ‘good’ reasons for continuing with boxing despite all of its disadvantages, because his reasons are valid relative to the biased culture, collective myth or adaptive preferences that constitute the shared background assumptions of Jim and his fellow community members. ²¹

Without making sense of Jim’s reasons by constructing such presumptive rational explanations so that his reasons appear sound and plausible to yourself, you cannot understand Jim’s reasons as they are intended – namely, as sufficient grounds for concluding that Jim (or anyone in relevantly similar circumstances) should continue boxing. In other words, without rationally interpreting Jim’s reasons, you can only take Jim’s reasons merely as his subjective opinion or belief about what is the case that does not have any real connection to what is actually the case.

But as mentioned above, this amounts to ignoring the intended meaning of Jim’s reasons, which is a claim about what is actually the case – and not merely about what appears to be the case in his own eyes. Therefore, if you are committed to respecting Jim as a moral equal whose judgements deserve to be taken seriously as they are intended, you must take Jim’s judgement about his reasons seriously as a claim about the sufficient grounds for his choice and not merely as a subjective expression of how things appear to himself (Habermas 1985, 132–133).

But notice that while your rational explanation might allow you to make sense of why Jim might have been led to think that his reasons are sufficient grounds for his choice, there is no conclusive evidence to conclude that your explanation and its associated judgement of Jim’s reasons are necessarily correct. After all, Jim might turn out to have important reasons for continuing to box that are completely unrelated to any of the factors mentioned above. Indeed, if you regard Jim as a moral equal who can competently interpret his own needs and wants and rationally weigh the pros and cons of his options, there is no basis for privileging your judgement over Jim’s own.

This means that insofar as you regard Jim as your moral equal, you cannot dogmatically privilege your judgement over Jim’s own or accept Jim’s judgement blindly, without first sorting out your judgemental differences through a rational procedure. But, as I have demonstrated above, such a rational procedure cannot be limited to idealized (i.e. ‘armchair’) reflection only, since the choice between different idealizations and their associated outcomes is bound to be arbitrary. It follows then that the only rational procedure that could, in principle, be used to sort out judgemental differences between yourself and other agents without arbitrarily privileging anyone’s particular view is a process of dialogue with the agents themselves. ²²

What I have argued so far shows that understanding other people’s reasons requires going far beyond simply asking for and knowing what their reasons are. Understanding others requires engaging in a sustained dialogue with them as to why they take their reasons to be sound, and whether they are actually right to believe that their reasons are sound. This makes it questionable whether a self-consistent anti-paternalist would be able to permit the higher level of intervention that is necessary to reach a genuine understanding of other people’s reasons.

In his famous discussion of the unsafe bridge example, for instance, Mill claims that you should let a competent person decide for themselves whether they have sufficiently compelling reasons to cross an unsafe bridge, if they have been warned that the bridge is unsafe.

Nevertheless, when there is not a certainty, but only a danger of mischief, no one but the person himself can judge of the sufficiency of the motive which may prompt him to incur the risk: in this case, therefore (unless he is a child, or delirious, or in some state of excitement or absorption incompatible with the full use of the reflecting faculty), he ought, I conceive, to be only warned of the danger (Mill 2003 [1859], 165).

On Mill’s view, although you may interfere with the agent to make sure that they have been informed about the dangers associated with their choice, you may not interfere with them further if it is certain that they have been sufficiently informed (see Arneson 1980, 484–485 for a similar interpretation of Mill). This is evident in Mill’s assertion above that since only the agent themselves can judge whether it is worth taking the risks, others may not interfere with them further if they have already been sufficiently warned.

But as I have explained above, understanding the agent does not stop at simply verifying that the agent is fully informed and aware of what they are doing. Understanding the agent requires you to get clear about why the agent is willing to cross the broken bridge even when they clearly know that the bridge is broken (and you have no doubts that they know this). Since Mill believes that individuals should have sovereign authority to decide the ‘sufficiency of the motive’ for incurring a given risk, it seems fair to conclude that Mill would (at the very least) consider it an unnecessary meddling in others’ business to enquire about their reasons for crossing the broken bridge.

Joel Feinberg’s view, which has its roots in Mill’s anti-paternalist view, is likewise inconsistent with the higher degree of intervention required to understand others. Feinberg argues that ‘an autonomous being has the right to make even unreasonable decisions determining his own lot in life, provided only that his decisions are genuinely voluntary (hence truly his own), and do not injure or limit the freedom of others’ (Feinberg 1986, 67). Indeed, for Feinberg, what counts as a valid reason for a particular action seems to be largely a matter of the individual agent’s subjective preferences (Feinberg 1986, 69, 106–112). This indicates that Feinberg too, would be opposed to any intervention beyond what’s required to inform the agent of the dangerous condition of the bridge – since for Feinberg, the agent’s reason for their decision is simply a matter of individual preference, which is no one else’s business.

Finally Quong, in a similar note, suggests that it is permissible to intervene only when there is no reasonable way for the agent to know about the dangers associated with their choice. If it is reasonable to expect that the agent should have known about the danger, then any further intervention would indicate a negative judgement of the agent’s capacity to appreciate the relevant facts and would therefore be objectionably paternalistic. On Quong’s view then, if you stop an agent attempting to cross an unsafe bridge even when there is a clearly visible sign saying ‘This bridge is broken’, your act would be paternalistic (Quong 2010, 82–83).

All of this suggests that anti-paternalists would generally hesitate to approve the higher level of intervention required to understand others. On the account I have given, understanding another agent requires you to try to get clear about the reasons why they are willing to cross the unsafe bridge even if you are certain that they know the bridge is unsafe. This is because the need for understanding arises precisely when you observe the agent trying to do something that appears irrational or self-harmful for reasons that are not yet clear to you. But the anti-paternalist views I have examined above require you to refrain from intervening once you have verified that the agent is aware of all the relevant facts, regardless of what the agent’s reasons might be. Therefore, it seems fair to conclude that liberal anti-paternalism is generally inconsistent with the sort of interventions required to understand others.

The surprising implication of this argument is that respecting others in the genuine sense of the word opens the door to paternalism, rather than closing it off as most liberal thinkers assume. If respecting others as equals requires trying to understand others’ reasons for action through dialogue, and this, in turn, requires going beyond the limits of what liberal anti-paternalists allow (as I have just argued), then it follows that respecting others opens the door to paternalism. In fact, contrary to the traditional liberal view that engaging other people’s rational faculties through dialogue is not paternalistic, engaging others in a dialogue with the aim of understanding their conception of interests turns out to be a distinctive form of paternalistic intervention in its own right.

According to the traditional liberal view, although engaging others in a dialogue or deliberation about their interests might trigger them to reflect upon and revise their conception of their own interests, that does not by itself show that deliberation invites paternalistic interference with others. Christian Rostbøll, for instance, argues that despite the fact that deliberation facilitates the rational transformation of participants’ preferences, deliberation still remains fundamentally opposed to paternalism, since contrary to deliberation, paternalism involves A forcefully imposing on B their judgement of what’s good for B against B’s wishes (Rostbøll 2005, 385–386). Unlike paternalistic intervention, Rostbøll argues that deliberation enables participants to retain their control as agents, since participants are allowed to reject or accept the claims of others on the basis of their own reasoning. Shiffrin and Quong both agree that engaging others in a rational argument about matters of their own business is not paternalistic, because it involves engaging with their rational agency instead of supplanting it (Shiffrin 2000, 213; Quong 2010, 81).

Although this argument may seem plausible at first glance, it drives a false wedge between engaging others in deliberation about their own affairs in the absence of their consent or inclination to do so and paternalistically interfering with others. Suppose Jim is getting ready to step into the ring for another boxing match, and I stop him and get him involved in a discussion about why he chooses to remain in the boxing business when he needs to risk his life every time and it barely pays him enough to make ends meet. My action seems paternalistic in this case, even if I am not forcefully imposing my judgement of what’s best for himself on Jim, and even if Jim retains the power to accept or reject my claims on the basis of his own reasoning.

For instance, Jim might not want to hear me recite all the reasons why continuing with boxing is a bad idea (although he knows what I am saying is all reasonable), or he might be unwilling to deliberate so carefully before doing what he wants to do. Indeed, even if I do not literally force Jim to talk to me (say, by blocking his path), my ‘offer’ to talk about Jim’s personal decision can still be paternalistic. ²³

To see why, note that in most ordinary situations, we casually ask others what they are doing when we spot them doing something that seems irrational or dangerous to themselves. And once people get involved in a conversation of this sort, they usually feel pressured to continue the conversation because they do not want to be misunderstood or give off the wrong impression that they are admitting that what they are doing is wrong or irrational. This is not because the one who initially started the conversation pressures or forces the other to continue the conversation: even if I do not pressure or force Jim to talk to me, Jim is likely to feel pressured to justify his choice to me, because otherwise, I might (wrongly) assume that he does not fully understand what he is doing or that he is not in his right mind. So even if Jim does not particularly welcome my advances, he might feel compelled to continue the conversation with me to prove that he knows what he is doing. The point is that the internal dynamic of rational justification has the effect of compelling participants to keep on engaging in the conversation once they get involved in it, whether they like it or not. In this sense, getting others involved in a discussion about their own business for their own sake is a type of paternalistic interference that does not involve (intentional) coercion.

In fact, even if we bracket this consideration, the very act of offering to talk to others about matters of their own business is paternalistic if the offer is a gratuitous offer intended to advance another person’s own good in the absence of their request (see the definition of paternalism defended in the first section, ‘The Equal Respect objection to Paternalism’). Consider a brother who asks his little sister whether things are going well with her boyfriend. Although his sister assures him that everything is fine, the brother is still doubtful and offers to talk with her about her boyfriend over coffee. The brother’s action seems paternalistic in this case, even if his sister ends up accepting the offer. ²⁴ Hence, even offers including offering others a chance to discuss matters of their own business is paternalistic, if it is aimed at improving their welfare or interests in the absence of their request. ²⁵

But while offering to talk with others for their own sake can be paternalistic, it is nevertheless compatible with respecting others as moral equals, since competent rational agents can also make mistakes from time to time. ²⁶ Just because I suspect that you might be mistaken on this particular occasion, I need not think you are generally incapable of deciding for yourself. This explains why, although the brother acts paternalistically in the example above, he might still have a lot of respect for his sister’s general capacity to deliberate and decide rationally. Indeed if what I have argued above is correct, the brother might actually be expressing respect for his sister by seriously making the effort to understand her reasons for staying in a seemingly toxic relationship, instead of presuming that she must be deluded or dismissing her reasons as just a matter of her idiosyncratic preference. This example shows that even though offering to talk with others about matters concerning their own welfare is paternalistic, it is not disrespectful as coercing others for their own good is.

Given all this, it seems fair to conclude that the dialogical intervention I suggest is paternalistic, even if the recipients are free to accept or reject the intervener’s suggestions and advice on the basis of their own reasoning. This suggests that it is misleading to assume that rational deliberation and paternalism are starkly opposed to each other. On the contrary, getting others involved in an unsolicited rational discussion about their own business for their own sake is a species of paternalism, even if it stems from respect for them as responsible agents who are appropriately responsive to reasons.

The arguments so far imply that although paternalistic coercion aimed at advancing others’ (so-called) ‘rational’ interests is unwarranted, paternalistic intervention through rational discursive means is warranted. In the context of state policy, this means that legislators, public health officials, economists and other experts who are concerned to promote citizens’ interests should try to arrive at a shared understanding of what citizens have most reasons to choose through a rational dialogue with them, instead of unilaterally coercing citizens for their own sake without regard to their actual beliefs about what’s best for themselves.

Conclusion

Current normative views on paternalism are divided by a false dichotomy between the idea that overriding citizens’ self-regarding choices is not disrespectful when their choices are mistaken relative to some idealized conception of rational choice, and the idea that the very act of suspecting that citizens might be mistaken about their (subjective) conception of their own interests is disrespectful. What I have tried to do in this article is to show that neither view gets it right on the question of what is required to genuinely respect citizens as equals.

Respecting citizens as equals requires taking citizens’ judgements about their own interests seriously as presumptively sound judgements that are grounded in valid reasons. In order to express respect for citizens as equals in this sense, however, I have argued that experts and policymakers must try to understand the reasons that underlie citizens’ judgements about their interests through a critical conversational exchange with citizens, instead of blindly accepting their reasons without knowing why they should be accepted (à la laissez-choisir anti-paternalism) or rejecting their reasons on the basis of some idealized conception of rational choice (à la rationality-based defence of paternalism). But engaging in such a critical process of understanding citizens, as I have illustrated above, functions as a distinctive form of paternalistic intervention in its own right.

So contrary to what many traditional liberals think, respecting others in the genuine sense of the word actually does seem to have a lot to do with paternalistic behaviour. This counterintuitive conclusion might be easier to swallow if we keep in mind that paternalistic behaviour is not always rooted in contempt for others but can sometimes be an expression of concern and respect for the fellow members of our moral community.