1 Introduction

With the core distinction between choice (responsibility) and circumstances (luck) in hand, luck egalitarianism claims that a distribution is just only if everyone’s share results from their own choices. This theory was seen as one of the most powerful egalitarian theories until Elizabeth Anderson posed objections to it. Among her several challenges to luck egalitarianism, the harshness objection is perhaps the best-known, most representative, and most powerful. This is because the harshness objection appeals to our egalitarian view about the appropriate treatment of people, which appears to conflict with the seemingly excessive emphasis on personal responsibility in luck egalitarianism.

Anderson (1999: 295–302) puts forth the harshness objection by raising eight cases. The most frequently referred to and significant case against luck egalitarianism concerns the treatment of the imprudent, that is, the case of a reckless driver:

An uninsured driver negligently makes an illegal turn that causes an accident with another car. Witnesses call the police, reporting who is at fault; the police transmit this information to emergency medical technicians. When they arrive at the scene and find that the driver at fault is uninsured, they leave him to die by the side of road (Anderson, 1999: 295).

As an alternative egalitarian theory, Anderson (1999: 315) presents a theory of relational (democratic) egalitarianism, according to which the social conditions of people’s lives should be such that they can hold an equal status to all other citizens. Since, in her argument, relational egalitarianism requires society to secure the basic standing of individuals—irrespective of their own personal responsibility for having become better or worse off—it is not vulnerable to the harshness objection, and thus it is more readily applicable to society’s problems than luck egalitarianism.

The harshness objection has been intensively discussed in egalitarian arguments (e.g., Voigt 2007; Segall, 2010: Ch. 4; Knight 2015; Lippert-Rasmussen, 2016: 189–91). Very recently, Andreas Albertsen and Lasse Nielsen provided a meticulous analysis of the harshness objection with the aim of explicating its biting edge (Albertsen & Nielsen, 2020). They argue that there are at least four possible ways of taking harshness as an objection to luck egalitarianism and subsequently show that only the inconsistency objection—the objection that luck egalitarianism is inconsistent with the ideal of basic moral equality—has real force.

More specifically, their argument unfolds the harshness objection in such a way as to divide it into the following objections: (1) the counterintuitiveness objection: luck egalitarianism involves a conflict with our intuition about the specific treatment of the imprudent— e.g., reckless drivers; (2) the badness objection: luck egalitarianism allows the level of advantage held by the imprudent to decline irredeemably; (3) the disproportionality objection: luck egalitarianism disregards a disproportional association between imprudent acts and negative consequences; and (4) the inconsistency objection: luck egalitarianism cannot meet the egalitarian ideal that people ought to be treated as moral equals. Albertsen and Nielsen maintain that the first three objections can be avoided even in the existing lines of luck egalitarian argument, such as revised luck egalitarianism (e.g., all-luck egalitarianism) and, more promisingly, pluralist luck egalitarianism. However, since, as Anderson (1999: 295) notes, all egalitarians would accept the ideal of basic moral equality, the inconsistency objection remains the most important for luck egalitarians to overcome.

The argument of Albertsen and Nielsen makes an enormous contribution to egalitarian debates for the following reason: as seen above, the harshness objection has been seen as a critical test for the relevance of egalitarian theories. Nevertheless, the harshness objection has been treated as a mere intuitive challenge against luck egalitarianism, even though it contains within it several distinct claims against luck egalitarianism, and not all these constituent disputes are based on intuition. Understanding the harshness objection in this way undermines the utility of the harshness objection in egalitarian debates, because appealing to particular intuitions plays a limited role in constructing general theories of egalitarian justice. Hence, the attempt to unravel which version of the harshness objection is truly decisive is of enormous importance in developing theories of egalitarian justice.

In this study, however, I demonstrate that the inconsistency objection is not as strong a challenge to luck egalitarianism as Albertsen and Nielsen believe. Before presenting my evidence, I will mention three preliminary notes. First, to see how strongly the inconsistency objection holds against luck egalitarianism, I construe the proposed objection in a comparative manner. As I see it, Albertsen and Nielsen treat the incompatibility between egalitarian theories of justice and the ideal of basic moral equality in a non-comparative manner: whether the incompatibility obtains is determined in an all-or-nothing manner, simply in terms of whether each egalitarian theory fails to treat people as equals. This is not a reasonable way of pressing the objection at stake, because an abstract ideal (including the ideal of basic moral equality) should allow for different ways of meeting the ideal in egalitarian conceptions of justice. We thus require a comparative perspective on how the ideal of basic moral equality has been satisfied in egalitarian conceptions of justice. This leads to the view that the proposed objection concerns the degree to which the targeted egalitarian theories mismatch the ideal. As shown below, by focusing on the degree rather than the binary fact of mismatch, we can see that the most decisive version of the harshness objection does not hold against luck egalitarianism in the proposed manner.

Second, given the importance of the comparative evaluation of the degree to which the relevant egalitarian theories meet the ideal of basic moral equality, I will use the term “incoherence objection” in place of “inconsistency objection” throughout this analysis. This makes it possible to see the degree to which each egalitarian theory—more specifically, luck egalitarianism and Anderson’s relational egalitarianism—fulfils the ideal of basic moral equality. This is because, unlike inconsistency, incoherence can be quantified in degrees.

Third, it is important to note that the aim of Albertsen & Nielsen (2020: 441–42) is to disentangle the different formats of the harshness objection to show that, if the harshness objection has real bite, it must be the incoherence objection—which is often overlooked as a format of the harshness objection—that is responsible for this bite. In this paper, I intend to challenge the antecedent of their conditional claim. In doing so, I do not presume that Albertsen and Nielsen put forward the inconsistency objection as a decisive objection to luck egalitarianism simpliciter, or as their claim against luck egalitarianism.

The argument proceeds as follows: First, I re-examine the point of the badness objection by a precise exegesis of Kasper Lippert-Rasmussen’s counterargument against Anderson’s relational egalitarianism, on which Albertsen & Nielsen (2020: 437–38) rely in order to make the incoherence objection distinct. This urges us to see that only a specific—deontological—version of luck egalitarianism is relevant to the incoherence objection, when compared with Anderson’s relational egalitarianism. Next, I show that the luck egalitarian conception of equal treatment does not endorse harsh policies, such as excessive responsibility-sensitive healthcare, that would be disrespectful to the imprudent. Finally, I demonstrate that deontological luck egalitarianism is not troubled by the case that involves a lack of respect for the prudent, which redirects the incoherence objection originally posed by Anderson against luck egalitarianism so that it becomes an argument against Anderson’s relational egalitarianism. I thereby claim that the harshness objection is not a truly decisive objection against the luck egalitarian project.

2 Badness Objection and Incoherence Objection

To grasp the thrust of the incoherence objection, we should first review how Albertsen and Nielsen divide the harshness objection into the different versions. We first notice that Albertsen and Nielsen, in the process of demonstrating the importance of the incoherence objection, distinguish between the objection based on harshness as a counterintuitive implication (i.e., the counterintuitiveness objection) and the objection based on harshness as a result of bad consequences such as harm or suffering (i.e., the badness objection). The formation of this distinction boils down to whether the badness of consequences can be evaluated independently of intuitive judgments about the particular cases (e.g., the case of how to treat reckless drivers in egalitarian theories). In egalitarian debates, it is often claimed that consequences are bad if they place individuals in dire straits. In this case, for example, the moral salience of meeting basic needs may account for the badness in question. A pluralist defense of luck egalitarianism makes use of this evaluation of the badness in that other moral values or principles, such as the humanitarian principle, embrace our obligation to save those who are very badly-off (Barry, 2008; Segall, 2010; Tan, 2012).

The case of Bullet Bob may be regarded as a hard case for even pluralist luck egalitarianism: Bullet Bob is a Russian-roulette player who accepts responsibility for the risk of severe injury in the event that he loses the game, which he knows is a real possibility. The problem this case illustrates is that while pluralist luck egalitarianism provides a moral reason to help Bullet Bob if he becomes severely injured as a result of playing Russian roulette, “it does allow bad outcomes as an implication of luck egalitarian justice” (Albertsen & Nielsen, 2020: 431–32).Footnote 1 On the other hand, Albertsen & Nielsen (2020: 432) admit that the pluralist strategy is not fundamentally challenged by this problem because, as seen above, other moral considerations can supplement luck egalitarian justice. Albertsen and Nielsen want to redefine the limitation of the pluralist strategy to take account of the incoherence objection.

To do so, Albertsen & Nielsen (2020: 437) refer to Lippert-Rasmussen’s argument that Anderson’s relational egalitarianism is subject to the harshness objection (2016: 190–91): Anderson’s relational egalitarianism endorses the ideal of relating to one another as equals; that is, it endorses a social order in which no social hierarchy that would treat specific people disrespectfully exists. Consider a society in which a global pandemic spreads in such a way that it affects all citizens negatively and equally—nobody is treated as unequal and all are very badly-off. Suppose a philanthropist offers everyone safe vaccines for free, which improves their lot. Then suppose that, while the prudent majority accept the offer, for religious reasons, a minority do not. According to Anderson’s relational egalitarianism, the resulting inequality between them would be unjust if it created a hierarchical social relation under which the minority would remain under worse-off conditions as compared with the prudent majority. Consequently, the vaccination of the prudent majority would be condemned under the scheme of Anderson’s relational egalitarianism, which has the same implication as luck egalitarianism does in terms of harshness.

Albertsen & Nielsen (2020: 438) thereby argue that this counterargument holds against the badness objection, but not against the incoherence objection, to luck egalitarianism, for it only explicates that Anderson’s relational egalitarianism also permits bad outcomes: in the pandemic case, creating a hierarchical inequality between the majority and the minority is worse than the outcome that all remain under the harsher conditions. However, in this argument, Albertsen and Nielsen presume that the badness of the outcome can be evaluated by Anderson’s relational egalitarianism independently of the ideal of democratic equality, just as it can be evaluated independently of intuitive judgments about the particular cases. This presumption is dubious because Anderson’s theory makes an evaluation in light of the democratic ideal of equal treatment.

This doubt is reinforced by Anderson’s support of the second-person justification for just decisions (Anderson 2010a: 16–17): relational egalitarianism is committed to the interpersonal standpoint from which a just policy ought to be rendered, such that any possibly affected agent holds the policy-makers accountable for their claim about whether the policy is reasonably acceptable to everyone or not. This view is called “deontologist” in that justice is intrinsically associated with the normativity of (the conduct of) agents, which is contrasted with a “consequentialist” view in which justice applies directly to a state of affairs.

Lippert-Rasmussen’s argument can then be construed as either (1) a counterargument against Anderson’s relational egalitarianism based on the ideal of basic moral equality that implies equal treatment or (2) not posing any criticism about the outcome (as a state of affairs) that only the minority are under dire conditions. If (1) is true—which I take to be the case because (2) seems not to be what Anderson would take seriously—Lippert-Rasmussen’s argument should be understood as demonstrating an incoherence between the ideal of basic moral equality and Anderson’s relational egalitarianism. But how incoherent are they? I propose that an important moral property inherent to the wrongful view to render the improved situation of the vaccinated majority unjust in light of the ideal of democratic equality is that this view disrespects their decision to get vaccinated, or more simply, disrespects the prudent majority. Although I do not contend that this is the only available view, this is plausible enough to constitute the counterargument against Anderson’s relational egalitarianism based on the ideal of democratic equality.

If my understanding of Lippert-Rasmussen’s counterargument is on the right track, this may well necessitate a re-examination of the badness objection against luck egalitarianism. This is because Albertsen and Nielsen distinguish the badness objection from the incoherence objection, and the above-mentioned argument highlights that the badness objection must differ from the incoherence objection. Now we should note that the incoherence objection is based on the ideal of basic moral equality in condemning the luck egalitarian treatment of the imprudently badly-off: the responsibly poor are treated with a lack of respect, and this is incoherent with the ideal of basic moral equality. Given this, the incoherence objection is posed only against the deontological form of luck egalitarianism, not its consequentialist form. Recall that consequentialism aims to realize the best possible state of affairs. In its consequentialist form, the goodness (badness) of states of affairs depends on the moral values that luck egalitarianism espouses, not directly on the ideal of treating people with equal respect and concern. The consequentialist version of luck egalitarianism fits well with Anderson’s meta-ethical view that luck egalitarians act on the third-person justification by which a just policy can be fully derived from a set of relevant premises—normative or factual. However, if luck egalitarianism is construed as the consequentialist version, we cannot appropriately estimate the importance of the incoherence objection that differs fundamentally from the badness objection.Footnote 2

Hence, the incoherence objection should be advanced as follows: First, targeted luck egalitarianism constitutes a deontological—hereafter, deontic—conception of justice. Second, the luck-egalitarian conception of justice is more incoherent with the ideal of basic moral equality (in terms of disrespectfulness) than Anderson’s relational egalitarian conception of justice. We should then see whether the incoherence objection obtains or not.

3 What Makes Luck Egalitarianism Deontic

Given that a deontic justification of claims appeals only to the normativity of (the conduct of) agents, there seem to be no difficulties in construing luck egalitarianism in a deontic manner. This is because luck egalitarianism is a theory of responsibility-sensitive egalitarianism. According to this format of luck egalitarianism, choices whose unequal consequences are just, are those for which agents are responsible (Arneson, 1989; Cohen, 1989, 1993; Barry, 2008; Stemplowska, 2013; Knight 2009, 2013; Lippert-Rasmussen, 2016; Albertsen, 2020). Obviously enough, this deontic notion of choice presumes (1) rational agency and (2) the availability of reasonable option(s): Unless agents are expected to have capacities for self-control in a rational manner, they are not responsible for choices; nor are choices deemed responsible if they are subject to manipulative processes that thwart the exercise of such capacities. Even if people can exercise the capacities, responsibility cannot ensue if there is no sufficiently good alternative that they can reasonably expect to obtain.Footnote 3 Note that there may be other conditions for responsibility-ascription, for example, certain epistemic conditions. Nor do I deny that some people (permanently) lack rational capacities—e.g., people with (congenital) cognitive disabilities. Suffice it to say that, for luck egalitarianism, the rational agency constraint and the reasonable option constraint are necessary conditions for holding people responsible for choices.

As construed here, luck egalitarianism can be seen as a deontic form in that the normativity of (choices of) agents are essential for making out which distributions are just. We should then ask if the luck egalitarian conception of justice is more incoherent with the ideal of basic moral equality than Anderson’s egalitarian conception of justice. Note that the incoherence objection derives its force from two kinds of disrespectful treatments of people: disrespecting the imprudently badly-off and treating the prudent disrespectfully. How can we compare the degree of incoherence inherent in the two egalitarian views? Both egalitarian theories share the ideal of basic moral equality, whereas they differ at the level of equal treatment. In other words, the luck-egalitarian conception of equal treatment is different from Anderson’s relational egalitarian conception of equal treatment. Given the difference between them, to assess the degree of incoherence, we should examine whether each view engenders either or both of the two forms of disrespect—that is, disrespect for the imprudently badly-off and disrespect for the prudent. Since Albertsen and Nielsen aim to show that luck egalitarianism would be incompatible with the ideal of basic moral equality, I will focus on whether the luck egalitarian conception of equal treatment would express either or both of the two forms of disrespect.

4 Does Luck Egalitarianism Disrespect the Imprudently Badly-Off?

The thrust of Albertsen and Nielsen’s argument for the incoherence objection is that luck egalitarianism is too disrespectful in its treatment of the imprudently badly-off. In examining this, let me address whether this criticism of luck egalitarian justice can be dislodged in a healthcare context.Footnote 4

In the context of healthcare, a typical manner in which the luck egalitarian conception of equal treatment might disrespect the imprudently badly-off is that it might license excessive responsibility-sensitive healthcare that would impose prohibitive costs on the imprudent. If this expectation proved true, it would be disrespectful to the imprudent. The point is that “a self-respecting person could not reasonably endorse such a policy when falling victim to its ruling herself” (Albertsen & Nielsen, 2020: 440). A question may be posed: Does deontic luck egalitarian justice consider this policy distributively just? Deontic luck egalitarians claim that the imprudent are responsible for the ruinous choice only if they are expected to exercise rational capacities, when presented with reasonable option(s) to choose. Can they hold that an option of reckless driving that may result in death, unless they have medical insurance, is such a reasonable option? The answer seems to be no; no rational agent would find it reasonable to opt for that choice.

This conforms to the manner in which existing luck egalitarians pursue a just policy under ordinary circumstances. They employ the reasonable person standard.Footnote 5 Since narrowing the scope of choice, while relevantly sensitive to responsibility, is of great significance for luck egalitarians, they hold people responsible for choices that a reasonable person would be expected to make, if given the opportunity to choose. This originates from Ronald Dworkin’s canonical claim:

I make no assumption that people choose their convictions or preferences, or their personality more generally, any more than they choose their race or physical or mental ability. But I do assume an ethic which supposes—as almost all of us in our own lives do suppose—that we are responsible for the consequences of the choices we make out of those convictions or preferences or personality (Dworkin, 2000: 7).

Indeed, his adapting the idea of the average insurer in a community to a choice of insurance schemes covering inborn disadvantages is a typical instance of the reasonable person standard (Dworkin, 2000: 77–78; Rakowski 1991: 97–106). A similar proposal is made by Roemer (1998: 8), who suggests an explanation of circumstances based on reasonable consensus among people in society. A similar route is followed by Segall (2010: 20), who posits that the distinction between choice and luck is based on what society can reasonably expect of people: the “unreasonableness criterion shifts the focus of attention from the individual to the society.” Luck egalitarianism appeals to a socially accepted understanding of choice and luck to establish a choice-sensitive and luck-insensitive egalitarian scheme (Tan, 2012: 94–97).Footnote 6

The reasonable person standard does not forbid paternalistic interventions. A consideration of this is important because, when Anderson (1999: 301–2) suggests that seat-belt legislation is no insult to people as long as “[s]elf-respecting persons can endorse some paternalistic laws as simply protecting themselves from their own thoughtlessness,” she suggests that luck egalitarians cannot provide “a more dignified explanation than that Big Brother knows better.” Albertsen & Nielsen (2020: 440) also believe that this is the heart of the incoherence objection; luck egalitarian justice cannot reasonably capture that “most would, out of respect for themselves, accept [the seat-belt] restriction on their own opportunity space.” However, by appealing to the reasonable person standard, luck egalitarians can reasonably countenance the seat-belt legislation. This is because, given our social conventions, driving without seat-belts does not reflect an important dimension of choice to which the luck egalitarian scheme should be sensitive. Note that this is not true in cases where the option of driving with seat-belts is available only to some, and not to others: the seat-belt restriction applies to all people and, thus, can be endorsed according to the reasonable person standard. This implies that paternalistic restrictions are unacceptable from the luck egalitarian conception of equal treatment if they are implemented discriminately, such that some people would be secured against their injury resulting from negligence while others would not. In this case, deontic luck egalitarianism does not permit the differential treatment of imprudent choices. This way, the reasonable person standard can function in determining what kind of restriction is acceptable from a luck egalitarian point of view.Footnote 7

Note that this use of the reasonable person standard is not in opposition to the choice-sensitive scheme, because the notion of choice-sensitivity is defined in ways that are sensitive to how reasonably people can act under ordinary circumstances. Admittedly, this is not based on the bare notion of choice. However, as Cohen (1989: 934) famously claimed, “there is no aspect of a person’s situation which is wholly due to genuine choice.” Even if there is such genuine choice, the metaphysical debate over the conditions under which a genuine choice can be made is of little relevance to the harshness objection. As Anderson’s case of the reckless driver illustrates, the harshness objection presumes that people can make an imprudent choice under ordinary circumstances.Footnote 8

Hence, we can claim that the reasonable person standard does not allow luck egalitarians to harbor a harsh policy that obliges the imprudent to incur tremendous costs for medical treatment. This is because the harsh healthcare scheme could not reasonably reflect the social practice of risk-taking in driving. The luck egalitarian choice-sensitivity does not pave the way for excessive responsibility-sensitive healthcare that contradicts the ideal of basic moral equality. This guarantees that deontic luck egalitarian justice does not disrespect the imprudently badly-off.

One might object that the reasonable person standard is too dependent on the social context, and thus cannot firmly prevent the incoherence objection; luck egalitarianism construed in this manner lacks distinctness as a normative theory that can provide an independent principle of egalitarian justice. However, if luck egalitarianism is not a theoretical view due to context-sensitivity, neither is Anderson’s relational egalitarianism, for the latter depends more crucially on the social context.Footnote 9 Specifically, Anderson’s argument lies in the various activities of people in civil society, including participation as equals in governmental affairs and economy (Anderson, 1999: 317). In addition, this does not expunge responsibility-sensitivity. In Anderson’s relational egalitarianism, the actual achievement of basic capabilities is conditional on the socially embedded productive system that demands citizens to fulfil their roles in a responsible manner (Anderson, 1999: 321). Hence, the context-dependency of the reasonable person standard for responsibility-ascription does not undermine my argument here.

5 Does Luck Egalitarianism Disrespect the Prudent?

Next, I will turn to whether deontic luck egalitarian justice treats the prudent majority disrespectfully in the pandemic situation. My answer is simply no, because deontic luck egalitarianism assigns importance to the voluntary decisions of individuals: the choice of the majority ought not to be renounced, even though this gives rise to an inequality between the vaccinated majority and the unvaccinated minority. This would be regarded as just only insofar as the rational agency constraint and the reasonable option constraint were met. We can reasonably presume that the two constraints are met in this imaginary situation. In this argument, one might challenge my view that the wrongful rendering of the improvement of the vaccinated majority as unjust functions to disrespect their prudent decision to get vaccinated. This can indeed be questioned, but the burden of proof should be shouldered by the objector to my view. Indeed, at the very least, my view has the advantage of shifting the burden of proof to the objector.

I believe that the two arguments for my positive answers to the two questions disempower the incoherence objection. It seems reasonable to claim that the luck egalitarian conception of equal treatment is less incoherent with the ideal of basic moral equality than Anderson’s relational egalitarian conception of equal treatment; the former can respect the prudent as well as the imprudently badly-off, whereas the latter cannot respect the prudent. Therefore, my argument demonstrates that luck egalitarianism need not be greatly concerned with the harshness objection, because the incoherence objection cannot undermine the luck egalitarian project for a plausible conception of justice.

On a final note, let us consider how deontic luck egalitarianism treats the case of Bullet Bob. According to the luck egalitarian argument, Bullet Bob would not be treated beyond his risk-stake proportion when severely injured. This may appear unduly harsh. As a consequence, one might claim, there remains an incoherence between deontic luck egalitarian justice and the ideal of basic moral equality. However, I deny this claim along two lines of argument. First, the option of playing Russian roulette, with the inherent risk of severe injury, seems not to befit the reasonable person standard under ordinary circumstances. It thus seems permissible for society to ban a game of Russian roulette under the luck egalitarian scheme, just as the seat-belt legislation is supportable in society. This way, society pre-empts the occurrence of the Bullet Bob case under the luck egalitarian scheme.

Second, society might permit a certain form of Russian roulette play, on grounds that this option for Bullet Bob-like people may offer them the invaluable opportunity to be acknowledged by society, as long as the requirement for the rational agency constraint and the reasonable option constraint are fulfilled. Bullet Bob would then not be medically treated if he were severely injured. However, in this case, the option that may result in Bullet Bob being severely injured without access to treatment may also be authentically preferred by him, and therefore, his opportunity to choose it should be respected, even though this would not be a reasonable option in society. One might consider that this argument begs the question, for leaving his severe injury untreated can be regarded as unduly harsh, independently of any conception of egalitarian justice. However, the harshness under discussion here concerns not disrespect for the imprudent, but rather disrespect for the prudent. Since we are now talking about a Bullet Bob scenario in which the option of playing Russian roulette is invaluable for, and thus authentically preferred by, Bullet Bob-like people, my way of dealing with the case of Bullet Bob does not beg the question.Footnote 10 I thus conclude that luck egalitarianism can deal with the case of Bullet Bob.

6 Conclusions

Anderson’s harshness objection, when construed as the incoherence objection, is not as strong as Albertsen and Nielsen believe, for two reasons. First, the luck egalitarian conception of equal treatment reasonably endorses the avoidance of excessively responsibility-sensitive healthcare, just as Anderson’s relational egalitarian conception of equal treatment does. Second, deontic luck egalitarianism is not troubled by cases in which the prudent are disrespected, which vexes Anderson’s relational egalitarianism that Lippert-Rasmussen’s argument highlights. I thus claim that the harshness objection is not a truly decisive objection against the luck egalitarian project. This is important because Anderson’s harshness objection has strongly influenced the debate over egalitarianism; hence, luck egalitarians have taken it seriously and examined how to cope with it. My argument shows that luck egalitarians need not bother with the harshness objection.