I. INTRODUCTION
In her now famous attack on luck egalitarianism, Elizabeth Anderson takes luck egalitarians to task for abandoning negligent victims: by virtue of its commitment to choice-sensitivity, luck egalitarianism apparently is compatible with, or positively affirms, letting people who, as a consequence of their own choices, end up badly injured or in financial misery, fend for themselves.Footnote 1 One line of response to this problem endorsed by a number of luck egalitarians, although not developed in detail, is paternalistic, requiring people to take out insurance against the adverse consequences of their own bad choices.Footnote 2 According to Anderson, this is objectionable in that it involves restricting people's liberty in significant ways by appeal to the disrespectful rationale that they are incapable of running their own lives.Footnote 3
Against this background, Anderson proposes a non-paternalistic argument for providing individuals with certain non-renounceable goods, that is, goods that cannot irrevocably be given up or contracted away. In brief, we have an unconditional duty to respect the dignity or moral equality of others. This implies, inter alia, the duty not to enslave others and not to leave them dying by the roadside even if they have consented not to be offered assistance should an accident befall them. As it is a duty we owe to all our fellow citizens, we may rightfully be taxed for this good. An inalienable or non-renounceable right corresponds to the duty in question. However, the duty is not a demand to benefit individuals or to cater to their interests and to do so against their will (i.e. it is not grounded in paternalistic reasons). It is a requirement to treat people respectfully, that is, in a way that their inner worth makes appropriate, irrespective of its consequences for the person's well-being or interests.Footnote 4 The bits and pieces of this argument are scattered around Anderson's dense and long article. In comparison with other aspects of her argument, it has not received much attention.
In parallel effort with Anderson, and more elaborately, Paul Bou-Habib presents an alternative non-paternalistic justification for ensuing that certain goods are unconditionally available to individuals.Footnote 5 In a nutshell, he argues, like Anderson, that we have a duty that right-holders cannot renounce to come to the assistance of others; if this is so, people's risky uninsured conduct may become a liability to us in the sense that we will have to carry the costs of discharging our duty of assistance towards them. Accordingly, risk takers may rightfully be compelled to take out insurance to cover the costs of coming to their aid, should they be the victims of an accident. Thus, according to this argument, they are required to take up insurance, not for their own sake, but for our sake, or for the sake of those duty-bound to bail them out in case of an accident. Again, this would appear to show that the argument would pass unscathed by any potential charges of paternalism.
In spite of the ingenuity and initial plausibility of both proposals, I believe that they ultimately fail. A certain paternalist argument for compulsory insurance is much more plausible, and it is unclear that the non-paternalistic reasons suffice to justify compulsory insurance or taxation. Specifically, the non-renounceable right to aid (corresponding to the unconditional duty to aid people in need or jeopardy) that is an integral part of both arguments is much more plausibly accounted for by reference to a duty to promote the value of autonomy as an intrinsic well-being-related value; and it is unclear whether a duty that is not grounded in such values would suffice to justify the restrictions on liberty involved.
The structure of the article is as follows. First, I provide some background for the discussion by briefly addressing the concepts of paternalism and a non-renounceable right (section II). Second, I examine Bou-Habib's argument (section III). Third, I present my paternalistic counterview (section IV). Fourth, I demonstrate the applicability of the foregoing considerations to a rebuttal of Anderson's non-paternalistic argument (section V).
II. PATERNALISM AND NON-RENOUNCEABLE RIGHTS
By an agent B's paternalistic interventions in activities of an agent A, I will understand: (i) restrictions on A's liberty imposed by B; (ii) for A's own good, that is, to benefit agent A in certain respects, that is, to promote A's welfare, well-being, or interests; (iii) against A's will.Footnote 6 Challenges have been made to each of the stated conditions, but they are immaterial to the issues pursued below.Footnote 7 This is partly because the acts that will count as paternalistic on the more restrictive conditions I endorse would also count as paternalistic on the less restrictive conditions being offered as substitutes.
Following Peter de Marneffe, I add to this understanding of paternalism a further so-called justificatory condition.Footnote 8 It aims at capturing the gist of reconciliatory non-paternalistic justifications such as Anderson's and Bou-Habib's examined in this article, and does so, I believe, in an appropriate way. It says, roughly, that (iv) B's act cannot be justified satisfactorily on non-paternalistic grounds. Accordingly, if a plausible non-paternalistic justification of B's act may be offered, it is not a paternalistic act. This implies that even if a policy restricts A's liberty for his own good and against his will it does not count as a paternalistic act if a plausible non-paternalistic justification of it can be offered. One may, of course, in such cases, say that while the act in question is partly or entirely motivated by paternalistic concerns, there is a non-paternalistic justification for it.Footnote 9 However, on the proposed standard for what counts as a paternalistic act, which in addition to a motivational condition – (ii) – includes a justificatory condition – (iv) – the act in question is strictly speaking not a paternalistic act.
With the justificatory condition in place, the liberal strategy of reconciliation makes sense. This strategy aims to reconcile liberals' traditional resistance to paternalism with the fact that many apparently paternalistic regulations appear reasonable.Footnote 10 It does so by trying to present compelling non-paternalistic justifications for the regulations or acts in question. In the absence of the justificatory condition, the relevance of these reconciliatory justifications would be doubtful; all that liberals would seem impelled to do in order to show that a policy or act is not necessarily paternalistic would be to show that it might be non-paternalistically motivated, not that a plausible non-paternalistic justification for the act or policy is forthcoming.
The concept of a non-renounceable right to receive aid in an emergency is, as stated above, central to Bou-Habib's and Anderson's arguments and to the inquiry of this article.Footnote 11 A right being non-renounceable means that it cannot irrevocably be contracted away, not for a fixed period of time and, a fortiori, not permanently. Joel Feinberg provides the example of a spring rite in which players for the duration of the contest give up the protection of their rights to life for a fixed period of time (i.e. allow others to try to kill them).Footnote 12 Renouncing rights on a permanent basis he exemplifies by a person giving up for good her right to hold property, or contracting to become the permanent chattel-slave of another (thus renouncing permanently her right to liberty), or giving up her right to life by putting it in the hands of another (allowing him to decide whether or not she will live and for how long).Footnote 13 This is the kind of things people cannot do given that they hold rights that are non-renounceable.
A right being non-renounceable should be distinguished from it being non-waivable. Waiving one's right implies giving up one's right, releasing others from correlative duties, in a way that is revocable. That is, the right in question can at any time be reinstated and the duties reactivated. For example, I may ‘waive my right to exclusive enjoyment of my land by inviting a stranger’ (and hence release others, or certain others, from duties to stay off my land).Footnote 14 Similarly, I may waive my right against others that they refrain from exposing me to danger, take a nap in my bed, or touch me.Footnote 15 But I may at any time recall the waivers in question (telling the stranger to get off my land, not to expose me to danger, not to use my bed, and not to touch me). A right being non-waivable implies that a person holding it cannot give it up on a particular occasion.
I shall interpret Bou-Habib and Anderson to be saying that while a person may give up her right to be rescued on a particular occasion (releasing others from their duties to save me), retaining her right to recall her waiver at any time – that is, she may waive her right to aid – she is not permitted to relinquish that right irrevocably, for a fixed period of time (for example, as regards a certain activity and its potential consequences) or on a permanent basis – that is, she may not renounce her right to aid. This is crucial in that it means that there can never be any guarantee that others do not incur liabilities as a result of my risky activities (I may decide not to waive my right to aid should an accident occur, and I cannot irrevocably, for what concerns the risky activity and its consequences, give up my right to aid) – they consistently have a duty to be prepared to offer me assistance in an emergency should I choose not to waive my right to assistance on such an occasion.
III. BOU-HABIB'S ARGUMENT FOR COMPULSORY INSURANCE
As noted above, Bou-Habib's argument revolves around the idea that we have a moral duty to assist people in emergencies. According to this duty, we ought to come to the rescue of others in need, at least in so far as we can do so without incurring significant risks to ourselves. The intuitive appeal of this duty seems strong. It is unthinkable, even shocking to our moral consciousness, to leave people in need or jeopardy unassisted.
Given this duty, other people's risky activities may impose unwelcome costs on us. Consider as illustration John Stuart Mill's famous case of a man about to walk on a bridge that is likely to collapse and a bystander capable of interfering.Footnote 16One reason for a bystander to intervene pertains to his own good. If, what is likely to happen, the man falls in the cold water, it is the duty of the bystander to jump into the cold water to rescue him.Footnote 17
Faced with the prospect of others imposing costs on us in this way, we have a right to defend ourselves. That is, we may rightfully prevent others from foisting avoidable costs on us. One way of doing so would of course be by applying the severely liberty-curtailing measure of outright prohibition. Another, perhaps more compelling, policy would be to require risk takers to insure themselves. In this way, we let them pursue their activities without curtailment, but on condition that they take out insurance so that they can fully compensate us for the expenses we incur when discharging our duty of assistance or rescue that gets generated when their behaviour ends tragically, landing them in need or jeopardy the alleviation of which depends on our assistance. Restraining people's activities in this way, on the argument on offer, has nothing to do with catering to the well-being or interests of those who may impose risk on us; curtailing their liberty is something we do for our own sake, that is, to avoid having to bear the costs of discharging our duty of rescue or to avoid not being appropriately compensated for doing so. This is why the argument for compulsory insurance apparently is non-paternalistic.
Bou-Habib's argument can be recast in the following way:Footnote 18
The moral duty to aid in an emergency: If A is in an emergency that is the result of A's deliberate risky behavior, and B can aid A at a sufficiently low cost, then B has a moral duty to aid A.
The moral right to defend oneself from exploitation: If B has a duty to aid A in certain circumstances, then B may compel A to insure for the costs of providing aid to A in those circumstances.
Conclusion: B may compel A to insure for the costs of providing aid to A, when B can aid A at a sufficiently low cost, in an emergency that is the result of A's deliberately risky behaviour.
Bou-Habib's argument faces a crucial objection which threatens its status as a genuinely non-paternalistic argument for compulsory insurance. The duty of aid or assistance integral to it would seem to have to be one that the correlative right holder cannot irrevocably dispose of. That is, the right in question must be non-renounceable. If A could irrevocably give up his right to assistance as regards a certain risky activity and its potential consequences, he could guarantee B that he would incur no liabilities pertaining to that activity. Consider Bou-Habib's case of you coming across a stranger about to perform a near-suicidal gliding mission across the Grand Canyon.Footnote 19 In light of the liability you may incur from having to rescue him, you may want to prevent him from performing the glide or to make him take out insurance so that he can compensate you for your costs in discharging your duty of rescue. However, if the stranger could renounce his right to aid he could reassure you that no liabilities for you will arise from his risky endeavour irrespective of its outcome. Note that while Bou-Habib's argument for the reason just given appears to presume that the right to aid is non-renounceable, it need not presume non-waivabilty. The stranger may, standing on the ridge of the canyon, about to undertake the glide, waive his right to aid, and be permitted to do so, but this being revocable, and given that the stranger, should he incur an accident, would be likely to change his mind and thus reactivate your duty, you would still have reasons for restraining him or for compelling him to take out insurance.
The presumption that the right to aid is non-renounceable is what gets the avowedly non-paternalistic argument into trouble. It bars A from irrevocably giving up his right to aid – that is, restricts his liberty – and does so quite possibly against his will. For example, A may want to undertake a dangerous endeavour without any guarantee of assistance being available at his discretion should things go wrong; or A wants to be able to engage in risky activities without others thereby potentially incurring liabilities – costs on the basis of which they may justifiably restrict his liberty in various ways to perform the activities.Footnote 20 Moreover, the apparently most plausible, and perhaps only sufficient, reason for doing so is related to the interests of the right holder – it would not be in his interest irrevocably to give up his right to aid when he needs it. It is advisable at all times to retain the right to receive aid or to decline it, especially in light of the fact that one may see things differently should an accident occur. Accordingly, the presumption that the right to aid is non-renounceable seems paternalistic and hence to throw serious doubt on the non-paternalistic credentials of Bou-Habib's argument.
Bou-Habib's response to the paternalism objection emphasizes that disallowing A to renounce his right to aid need not involve catering to his good, interests or well-being. Bou-Habib appeals to ‘a duty to preserve autonomy as something of intrinsic value’.Footnote 21 In conceptualizing autonomy he appeals to Joseph Raz's notion of autonomy.Footnote 22 According to this, autonomy implies, roughly speaking, one being the author of one's own life – i.e. it is a life that is lived in accordance with the person's own considered convictions and values. Autonomy thus understood requires inter alia (a) appropriate mental abilities, (b) adequate options, and (c) the absence of coercion and manipulation. Bou-Habib holds that autonomy of this kind has intrinsic value in the sense that it is valuable independently of, critically, ‘its effects on the well-being of the persons whose autonomy it is’.Footnote 23 A may not renounce his moral right to receive aid ‘because his doing so would contravene his duty to preserve his own autonomy’,Footnote 24 the duty in question being a duty to preserve autonomy as something of non-instrumental value. Accordingly, B denying him the right to renounce his right to assistance need not be ‘paternalistically grounded in a concern for his well-being’.Footnote 25
Bou-Habib does not do much by way of spelling out the notion of the intrinsic value of autonomy. Given that his argument requires an account of autonomy that is independent of the well-being of the person whose autonomy it is, a Kantian account would seem to be appropriate. J. David Velleman, for example, argues that a person himself and his autonomy possess value independent of and prior to this person's interests or well-being.Footnote 26 To kill or destroy oneself or, perhaps, to run a severe risk of doing so for the sake of promoting one's interests or well-being (for example, in not experiencing pain) is, according to this view, to get things backwards morally speaking. That is, it would amount to treating what has value in itself (i.e. an end), namely the person, as a mere means to the promotion of one's interests or well-being.
When others, perhaps against the will of the person whose autonomy it is, cater to this value inherent in every person (but not for anyone), they are not acting in promotion of his good (against his will); they appropriately respect the value of autonomy possessed by the person in question.Footnote 27 In the case of B denying A the right to renounce his right to aid, Bou-Habib may point out that this is precisely what B does; that is, he is acting to preserve autonomy as a value unrelated to the well-being of A whose autonomy it is, or he is respecting this value in A. This is, most certainly, a non-paternalistic argument for a person having a non-renounceable right to receive aid, or at least an argument that eschews welfare-catering paternalism, which is arguably the standard version of paternalism.Footnote 28 Hence it may serve to rebut the paternalism objection against Bou-Habib's argument.
However, Bou-Habib's argument is open to a serious challenge. First, it is evasive of what appears to be the fundamental paternalist rationale for the policies in question.Footnote 29 Second, it is far from clear that the non-paternalistic rationale on offer is sufficient to justify the policies in question whereas the alternative paternalistic rationale clearly is (it could be the case that although the non-paternalistic rationale were not the most compelling or strongest justification, it would be sufficient to justify the policies in question which would be all that were required for the liberal reconciliatory strategy to succeed). Mounting these challenges is predicated on presenting a plausible alternative paternalistic rationale for non-renounceability and other apparently paternalistic policies. Accordingly, I begin the section below by presenting a view that is similar to Bou-Habib's in catering to autonomy as an intrinsic value but differs crucially in accounting for this value in a way that is closely related to individuals’ well-being – it takes, as Mill's has it, autonomy (or individuality) to constitute an important element of well-being.Footnote 30
IV. AUTONOMY, WELL-BEING AND PATERNALISM
The gist of my argument for the non-instrumental well-being-related-value of autonomy – a value which apparently paternalistic policies arguably cater to, and ought to cater to – is that autonomy has value independently of its tendency to promote pleasure or preference-satisfaction, and that the value in question is appropriately seen as a constituent of well-being or welfare (alongside other constituents, including pleasure or preference-satisfaction).Footnote 31
First, in many circumstances, autonomy is not a necessary or particularly efficient means of promoting happiness. Happiness is more likely to be promoted by a person unthinkingly following his natural inclinations and propensities or it may be manufactured by Brave New World architectonics. Second, the situation in which a person can choose between a wide range of adequate options (i.e. to act autonomously in this sense) is preferable to one in which he is faced with only one option even if this option is his most favoured one or the one that gives him most pleasure. He is exercising his autonomy in saying no to a number of inferior options, and this has value in itself (for reasons, as will be clear below, related to agency and self-esteem). Third, it might be that a committee of benign friends and relatives could make decisions on my behalf that would be superior, in terms of giving me what I want or making my life as pleasurable as possible, to the choices that I would make on my own. However, it would constitute a significant loss to me were this committee to run my life. It would in a sense not be my own life anymore, and its being so seems of significant and independent value.
So autonomy seems to be intrinsically valuable. It is so, arguably, partly by virtue of it being constitutive of self-esteem. When a person chooses among a range of options and is not subjected to the will of others (cf. the Razian notion of autonomy Bou-Habib appeals to above), he realizes to some degree (a higher one than the one realized by a non-autonomous person) the ideal of agency: he has an impact on the world or makes a difference to it. Furthermore, he imbues his own life with unity, arranging it in accordance with his own intentions, plans and projects (the degree of autonomy displayed varies among different persons). He also achieves some critical distance from his inclinations and propensities by deciding which one to follow and the appropriate degree. This is related to his ability to act on maxims or for reasons.Footnote 32 By virtue of this shaping of his own life and how it reflects certain traits and values of the person, the person may be seen to have a certain character. Now a person's character is the appropriate object of esteem or (more narrowly) ‘appraisal respect’, or, in the self-regarding versions of these attitudes, of self-esteem or appraisal self-respect. Self-esteem is an essential part of regarding our plans and projects as worth pursuing. In John Rawls's memorable phrase, without self-esteem or appraisal self-respectFootnote 33 ‘nothing may seem worth doing’.Footnote 34
In this way, autonomy, being constitutive of self-esteem, which forms an essential part of imbuing our life-plans with value, may be seen as a part of a good life or our well-being.
Turning now to the apparently paternalistic policies – including policies guided by our disinclination to allow people to renounce some of their important rights – at issue between non-paternalist liberal reconciliation theorists and paternalists, we shall see that they cater much more plausibly to individuals' autonomy of a kind that is related to their well-being (such as the one just outlined) than they do to individuals' autonomy of a Kantian non-well-being-related kind. In this way the Kantian view is evasive of the fundamental rationale of the policies in question (the first challenge to Bou-Habib's argument mentioned above). Following Gerald Dworkin, we may distinguish between the following three kinds of policies.Footnote 35 First, policies catering to people's safety, for example requiring motorcyclists to wear helmets (‘safety cases’); second, cases in which we bar people from renouncing important rights they have (‘slavery cases’); and, third, cases of collective actions such as putting fluoride in the community water supply (‘collective decisions’).
When we endorse such policies, I submit that our primary, even completely dominating, concern is to prevent various kinds of accidents and injuries, and to ensure that assistance is available should such occur, and to do so for the straightforward and compelling reason that people's lives can be affected in a very bad way, inasmuch as their interests and opportunities can suffer grave and irrevocable setbacks, by the occurrences in question or by the absence of appropriate sources of aid in case that they do occur. The particular importance of autonomy or of barring threats to autonomy and the related value of self-esteem, on this picture, is due to the profound negative impact on people's lives that deprivation with respect to their autonomy is likely to have.
I am appealing here to what seems to me a strong and basic intuition regarding our reasons for pursuing the policies in question. If we make motorcycle helmets compulsory, for example, our salient reason seems clearly to be the health and well-being of the biker. We care about how a crash without a helmet will affect his life, how it will make his life go worse in many ways. Of course, other concerns, such as the fact that we might incur some costs in relation to his crash or that his autonomy of a kind unrelated to his well-being might be badly affected, may certainly play a role, but such concerns seem to me to dwindle in comparison with the salient concerns which pertain to the well-being of the person in question.
In view of the salience of the individual well-being-related reason for restricting people's liberty in the policies in question, try undertaking the following exercise.Footnote 36 Assume that for some reason the Kantian non-well-being-related reason is mute, assume also that other non-paternalistic concerns are silent (for example, concern about the relatives of those potentially involved in accidents). It would seem to me that even under those conditions, the rationale for imposing the kind of restrictions at issue is strong, even unshaken, and that is arguably because the salient ground for these restrictions is paternalistic interests-related or well-being-related in the first place.
Conversely – and this is the second challenge to Bou-Habib's argument for the non-renounceability of the right to assistance mentioned above (regarding its sufficiency as a reason for restricting people's liberty) – undertake a similar exercise with respect to the proposed Kantian ground. That is, disregard paternalistic reasons for restrictions and disregard alternative non-paternalistic grounds. Would we in such circumstances still have reasons for restraining people in the ways suggested above? This is not, for the following reasons, clear to me.
Among the apparently paternalistic policies that the reconciliation theorists would like to defend non-paternalistically in the indicated way are some, we all agree, that involve significant restrictions on people's liberty. This includes collective schemes of insurance or ‘making mandatory the various universal social insurance programs characteristic of modern welfare states: social security, health and disability insurance, disaster relief, and so forth’;Footnote 37 and it includes, as noted above, disallowing people to renounce the right to assistance. Furthermore, these restrictions are ex hypothesi grounded on reasons unrelated to, and possibly against, considerations of the well-being or interests of the person whose liberty gets restricted. Instead, they cater to autonomy as an intrinsic value or to the person's rational-agency capacity. If this is true, it would seem to me that Bou-Habib's argument and similar arguments in the liberal reconciliation strategy ask people to incur unreasonably burdensome sacrifices to preserve something of impersonal value or of value unrelated to individual well-being. In my opinion, people cannot reasonably be asked to give up the requisite means for upholding ambitious collective schemes (which would plausibly involve contributing close to half of their income or even more) for reasons unrelated to individual well-being.
Consider the following example. I come into possession of a coin of profound intrinsic value, this value, say, being conditional on it preserving its coin-form. Let's say that I simply stumble on it on my way to work. Now, I happen to live in a society in which a regrettable trend has developed in which individuals and groups (private as well as public) seek to contrive coins of intrinsic value for the purpose of melting them, transforming them into fancy necklaces. In such circumstances, it seems to me, I would certainly be obliged to incur some costs and accept some difficulties in preserving the item with profound value which has happened to come into my possession and which I am presently the person most capable of guarding and preserving. For example, I may be obliged to hide it, possibly even from close relatives such as my partner, who is a bling-bling fanatic, and I may be obliged to polish it or in other ways act out of respect for it and the value it possesses.
But it seems to me that I am not morally obliged to withstand pressures from individuals and groups, having found out about the good I am hiding, of the kind that would allow me to keep the coin and thus preserve its value only if I accept deduction of close to half my income or more (i.e. it would not be unreasonably squeamish if I give in to proposals of the kind: ‘give me the coin, or I will appropriate half of your income’). It might be praiseworthy or at least heroic of me to do what it takes to protect the coin in such circumstances, but not something that is plausibly required of me.
So the burdens liberal reconciliation theorists ask citizens to incur in order to preserve something with purely intrinsic value, or because it possesses such value, seem to me to be excessive. Now, it is certainly true that we sometimes, justifiably it would seem, in some sense ask people to sacrifice liberties that are important to them for reasons unrelated to their well-being or interests. Acting rightly, one might say, sometimes involves acting in ways that are not maximally conducive to one's own well-being.Footnote 38 However, when we do so, this is arguably to an important extent justified by reference to the importance of protecting the rights of others to an equal share of liberty and the interests catered to by these rights. So the curtailment here of liberty for reasons unrelated to the well-being or interests of the persons whose liberty gets restricted is unlike restrictions aimed at protecting personhood or individual worth in that it is undertaken, most plausibly, not to preserve something of intrinsic value, but to prevent illegitimate curtailments of the liberty rights of others and the interests and well-being protected by such rights.
Less aggressively, but still causing problems for Bou-Habib's argument, one may say that it is at best somewhat inconclusive whether the Kantian reasons offered are sufficiently weighty to justify restricting people's liberties in non-trivial ways. Conversely, at least when focusing on cases that arguably involve a high risk of grave and irrevocable effects on one's autonomy and the related value of self-esteem and its role as facilitating rational life plans or appreciating their value, the well-being-related reason in question seems of sufficient weight to justify local curtailments of autonomy or liberty.
(In the considerations offered, I have chosen to set aside an objection to Bou-Habib's argument (and, for that matter, to Anderson's argument to follow) which many find telling. It says that denying the renounceability of the right to assistance by reference to a controversial conception of autonomy or personhood such as the Kantian is in itself a kind of problematic paternalism.Footnote 39 I tend to be less moved than others by this objection, perhaps because I share Bou-Habib's optimism with regard to accounting for autonomy in a way that is not disrespectful of competing, reasonable conceptions of the good, although my view on the appropriate content of such an account, especially regarding the question whether it is related to well-being or not, differs from his.)
I conclude that the Kantian-inspired reasons for restricting people's liberty as is done in the policies at issue are evasive and not necessarily sufficient. The welfare- or well-being-paternalist reasons are much more compelling. The upshot is that Bou-Habib's avowedly non-paternalistic argument for compulsory insurance appears to depend for its success on an assumption, the assumption of the non-renounceability of the right to receive aid, which can only be satisfactorily accounted for paternalistically. I will now turn to consider Anderson's non-paternalistic argument for providing individuals with certain goods on an unconditional basis. Like Bou-Habib's argument it is crucially dependent on the right to receive aid in an emergency being non-renounceable. I can deal with her argument more quickly in that its crucial defence is of a nature similar to that of Bou-Habib's. However, Anderson's argument is worthy of consideration in its own right in that it differs from Bou-Habib's in important respects and because, as noted in the introduction, it is a relatively neglected part of Anderson's broader argument.
V. ANDERSON AND THE NON-RENOUNCEABLE RIGHT TO AID
According to Anderson we have an unconditional duty not, for example, to enslave other people (even with their consent), and a duty to come to offer them assistance in the case that their deliberate risky behaviour results in an accident (even if they have permitted us not to do so):
[T]here are some things one may never do to other people, such as to enslave them, even if one has their permission or consent.Footnote 40
[T]he obligation to provide health care is unconditional and can't be rescinded, even with the permission of the person to whom the obligation is owed. We are not permitted to abandon people dying by the side of the road, just because they gave us permission to deny them emergency medical care.Footnote 41
We may all justifiably be taxed towards the purpose of discharging this duty: ‘Since this is an obligation we all owe to our fellow citizens, everyone shall be taxed for this good, which we shall provide to everyone.’Footnote 42 Corresponding to the duty or obligation in question, people have a set of inalienable or non-renounceable rights, for example not to be enslaved and to be offered assistance in an emergency. Like Bou-Habib, Anderson denies that non-renounceability rests on paternalistic reasons. That is, we are not disallowing people to renounce the right to be offered assistance out of concern for their well-being or interests, we are showing a certain appropriate respect for them or for the worth in them, irrespective of the effect our denials have on their well-being or interests:
Kant would put the point [about ‘the unconditional obligation of others to respect’ an individual's ‘dignity or moral equality’] as follows: every individual has a worth or dignity that is not conditional upon anyone's desires or preferences, not even the individual's own desires.Footnote 43
[Democratic equality is:] basing inalienable rights on what others are obligated to do rather than on the rights bearer's own subjective interest . . .Footnote 44
We recognize this [moral] worth in your inalienable right to our aid in an emergency.Footnote 45
Focusing on duties pertaining to emergencies, Anderson's argument can be recast in the following way:
The moral duty to aid in an emergency: If A is in an emergency that is the result of A's deliberate risky behaviour, and B can aid A at a sufficiently low cost, then B has an unconditional duty to aid A.Footnote 46
The non-interest-based justification of the duty to aid in an emergency: the duty to aid is grounded not in a concern for the well-being or interests of the rights holders (i.e. those who hold certain non-renounceable rights), but in an appropriate respect for the persons holding the rights, or for their worth.
Justified taxation: We may all justifiably be taxed towards the purpose of discharging this duty.
Conclusion: There is a non-paternalistic argument for universal taxation towards the purpose of discharging the moral duty to aid in an emergency.
There are certain notable differences between Anderson's and Bou-Habib's arguments. First, Anderson's argument is not, whereas Bou-Habib's is, a sort of ‘public costs’ or ‘public charge’ argument. That is, taxing people (or certain groups of people) or compelling them to insure is not justified as a way of making them set aside funds for covering expenses that their activities are likely to extend to others (by virtue of their having an unconditional duty of assistance). Instead, it is justified directly by appeal to a duty of assistance (and of refraining from certain ways of treating others) of universal applicability both in terms of its requirements and in terms of the individuals to whom this duty is owed.
Relatedly, Anderson's scheme is apparently less choice-sensitive than that of Bou-Habib. Bou-Habib would seem to justify compelling only certain groups to take out insurance, namely those who engage in risky behaviour, and as such are likely to become public burdens to the rest of us. Anderson's scheme would of course extract contributions from risk takers, which makes it more choice-sensitive than a scheme in which people may engage in cost-displacement.Footnote 47 Still, it is less sensitive, and less fair, at least if one is committed to a choice-sensitive account of justice, than a scheme in which only those who engage in risky behaviour are forced to take out insurance. Why, one might ask, should those who would never think of engaging in conduct that carries a high risk of them ending up in the cold water (cf. the bridge example mentioned above), in a motor-cycle accident, in a climbing accident, and so forth, contribute on an equal basis to a scheme geared to cater to such accidents? In contrast to risk takers, they do not engage in a lifestyle that carries costs that they rightfully ought to bear. At least it would seem that a consistent cost-sensitive approach would make required contributions sensitive to the likely costliness of one's behaviour to others. Those who take reasonable precautions not to end up in accidents and hence foist costs on others should, it seems, not make contributions in excess of what is necessary to finance the costs of assistance pertaining to accidents that happen in spite of people taking due precautions.
With these differences between the two arguments in mind, it is clear – given the notion of non-renounceable rights integral to Anderson's argument – that worries similar to those initially raised against Bou-Habib's argument regarding its status as non-paternalistic apply with equal force to Anderson's argument, and that her strategy of overcoming them is parallel to that of Bou-Habib. In a nutshell, her strategy is to say that the significant curtailments of liberty involved are justified not out of a concern for the interests or well-being of the person whose liberty gets restricted (i.e. not paternalistically); rather, the justification appeals to reasons unrelated to beneficence, namely, in Anderson's case, the importance of us not standing in a relation to others that is disrespectful or in infringement of strictures that apply to how we should treat others given their equal moral worth.
In light of my rejoinder to Bou-Habib's similar response above, it will come as no surprise that I find Anderson's strategy unsatisfactory or at least not clearly offering a justification for the restrictions of liberty involved. I do so because it involves imposing significant burdens on individuals for reasons unrelated to their well-being or interests, and for reasons unrelated to the well-being or interests of others. Whilst I do not, and need not, deny that the kind of considerations appealed to by Anderson justify some restrictions on liberty, it is at best indeterminate whether they justify the significant restrictions at issue. Furthermore, reasons related to people's well-being or interests appear to me to offer much more compelling grounds for restricting people's liberty in the proposed way. We want to bar people from making major mistakes such as renouncing rights that are necessary for the continuation of their autonomy; and we want to do so because we are concerned about the adverse impact of such mistakes on their lives – that they are highly likely to make their lives go worse.
Part of the strength of Anderson's criticism of luck egalitarianism, with which this article began, is due, I submit, to the way in which it highlights a concern with imprudent persons which we ought to have, and many of us indeed have, but which luck egalitarians apparently do not have. It is not, or not primarily, that luck egalitarians fail to treat such people respectfully in a way that is unrelated to their well-being. And if what I have said here is correct, luck egalitarians were on the right track in responding to these concerns: paternalistic reasons are indeed the salient and plausible ground for restricting people's liberty in certain ways with a view to preserving their autonomy, and being thus focused they may be seen as considerations that can plausibly temper while not jettison the choice-sensitivity of luck-egalitarian theories of justice.Footnote 48
