Luck egalitarianism and the harshness objection

Luck egalitarians refer at least to this very generic principle of luck egalitarian justice:

While luck egalitarians have various ways of formulating the above, fairness pertains to the elimination of influences for which people are not responsible. This opens the door for distributions reflecting people's exercises of responsibility. This particular element of luck egalitarianism gives rise to the harshness objection. Generally speaking, luck egalitarianism allows for seeing no injustice in leaving people alone to face even fatal consequences that are not due to bad luck. That this would be too harsh on the imprudent is the core claim of the harshness objection. Anderson provides the following formulation of this objection to luck egalitarianism.

Fleurbaey presents another famous formulation in an earlier text involving a motorcyclist, Bert.

The clear message of these stories is that, in some morally relevant respect, luck egalitarianism fails to provide an adequate answer when tasked to consider whether – as a matter of justice – the injured person is owed assistance or compensation for the hardships endured. Specifically, the claim put forward is that luck egalitarianism is too harsh in its answer. Luck egalitarians choose various routes when responding to the harshness objection. In the growing literature on luck egalitarianism, we can identify three luck egalitarian responses relevant to our discussion. First, the revisionist strategy redefines the principle of luck egalitarian justice so that it allows for less harshness (Knight, Reference Knight2015). Second, the pluralist strategy appeals to value pluralism in order to allow other values to wash away the harsh mark on luck egalitarian justice (Cohen, Reference Cohen2008; Segall, Reference Segall2010; Temkin, Reference Temkin2017). Third, the rejection strategy bites the bullet and denies that the alleged consequences of luck egalitarianism are unfair (Voigt, Reference Voigt2007; Eyal, Reference Eyal, Liao and O'Neil2017; Stemplowska, Reference Stemplowska, Liao and O'Neil2017). Whether these responses succeed and the extent to which the harshness objection has any real bite depends on what is, effectively, the underlying moral driver of the counterexamples.Footnote ⁵ In other words, it depends on the point of the harshness objection. The statements from Fleurbaey and Anderson underscore the difficulty in identifying what makes these difficult cases for luck egalitarianism. More precisely, they do not tell us what the harshness objection objects to. Several features from the cases could be put forward as possible interpretations of what harshness is. Is it the bad consequences, the readily available insurance, the triviality of the preference, the unlikelihood of the event turning out as bad as it does, or is something entirely different motivating the complaint? This ambiguity in what constitutes harshness renders it difficult to say anything regarding the characteristics of situations where harshness becomes relevant as well as the empirical frequency of such cases. Furthermore, if it is unclear exactly how luck egalitarianism fails, it is also harder to assess the adequacy of a luck egalitarian reply. Together, these concerns reveal the need for an elaborated understanding of possible ways to take harshness as an objection to luck egalitarianism. The next sections consider in turn four different important versions of the harshness objection. These are, respectively, that the failings of luck egalitarianism pertain to counterintuitive implications, the badness of the outcomes, disproportionality between responsibility and consequence, and inconsistency.

Harshness as a counterintuitive implication

Some push the harshness objection simply on an intuitive level. Indeed, many critics who provide other objections simultaneously hold luck egalitarianism to be simply counterintuitive, so the counterintuitive implication version of the harshness objection need not – and rarely does – stand alone. In fact, this intuitive critique is often employed as the entry point for discussing other versions of the harshness objection. However, if we are to understand what the point of the harshness objection is, we need to elaborate on the substance of the counterintuitive implication version in its own right. Once we have done so, we shall see that it is different in important ways from other forms of the harshness concern.

Daniels’ argument that luck egalitarianism is too narrow provides an example of where the counterintuitive implication version of the harshness objection stands alone. In Daniels’ account of health justice, luck egalitarianism is too narrow because it will abandon the sick or injured if their condition is their own responsibility, which Daniels claims to be less than what health justice intuitively demands. Daniels unfolds this argument as follows. On the one side, ‘They [luck egalitarians] conclude that we have claims on others for medical assistance only if we are sick through no fault or choice of our own’ (Daniels, Reference Daniels2008: 76). On the other side, ‘Others reject this conclusion. They find abandoning sick people to their chosen fate morally unacceptable, concluding that we have an obligation to fix the leg broken in a skiing accident simply because the skier cannot function normally if we do not’ (Daniels, Reference Daniels2008: 76). Daniels then finds luck egalitarianism too narrow to fit our intuitions by concluding that ‘even if considering responsibility has intuitive appeal, it is not so central or important that it should override other considerations involved in justice’ (Daniels, Reference Daniels2008: 76).Footnote ⁶

Daniels’ skier example serves here as a harshness case meant to show that luck egalitarianism misses the mark regarding what justice should intuitively demand, without further argument. Hence, on its own, it claims that harshness provides an intuitive argument against luck egalitarianism. Therefore, according to Daniels, on the grounds of reflective equilibrium, we should revise or reject the principle of luck egalitarian justice. As we will see, this criticism is compatible with a stronger internal critique explaining why luck egalitarians themselves should also reject the implication in the harshness case for reasons of consistency. But this is not Daniels’ argument. His point is much weaker: namely, that many would find these implications of the principle quite unattractive. As we will also see below, however, the intuition on which Daniels grounds his external critique is incompatible with the principled reasons of proportionality that motivate the disproportional version of the harshness objection. Thus, importantly, these separate critiques within the conception of the harshness objection must be disentangled.

How can luck egalitarians defend themselves against Daniels’ intuitive strike? One strategy is simply to reject the intuitive claim. This strategy is often pursued by reshaping the harshness case to involve not only a choice between treating the reckless driver or not, but also a choice between rescuing the reckless driver or rescuing another innocent person (pedestrian or passenger).Footnote ⁷ If it feels intuitively wrong to justify towards the reckless motorcyclist that we will not treat him because his injury is his own fault, it seems even more intuitively wrong to justify towards an innocent pedestrian that he will not be treated even though he was not at fault, while the reckless driver will be offered the needed treatment instead. Thus, as far as harshness goes and assuming that we cannot help everyone, intuition seems to push towards the view that disregarding the individual exercise of responsibility is less intuitively attractive than suggested (Albertsen and Knight, Reference Albertsen and Knight2015; Knight, Reference Knight2015; Lippert-Rasmussen, Reference Lippert-Rasmussen2016: 190).

The first limitation of understanding the harshness objection as merely an intuitive test is that it seems to leave ample room for luck egalitarians to reply by simply questioning the driving intuition, but the intuitive version is limited in yet another way. The most important limitation is that it provides us with little in terms of identifying the moral driver of our intuitive reaction. It fails to provide us with an adequate description of what the failure of luck egalitarianism consists of. Thus, while the counterintuitive version can hardly be the point of the harshness objection, it may serve as an entrance for further and more substantial harshness critiques, which unfold more principled reasons for skepticism about luck egalitarians. We now turn to these.

Harshness as bad consequences

A recurrent theme in the canonical descriptions of the harshness objection concerns the bad consequences suffered by the person responsible for their own situation. Knight formulates the harshness objection with reference to the claim that ‘no plausible theory of justice will abandon individuals to severely harmful outcomes’ (Knight, Reference Knight2015). Voigt remarks that ‘Perhaps, the strongest intuition underlining the importance of the Harshness Objection is the desire, or perhaps even duty, we feel to relieve (especially extreme) suffering’ (Voigt, Reference Voigt2007: 394). On this interpretation, harshness pertains to the very low level of advantage experienced by the individual responsible for being at this level.

The bad consequences version of the harshness objection is often employed in concert with the counterintuitive version elaborated above, adding to it that our strongest principled reason for dissatisfaction with luck egalitarianism is that it allows people to face severely bad consequences, which is what drives us to condemn luck egalitarianism as too harsh. Hence, while these two versions of the harshness objection – the counterintuitive and bad consequences versions – pull the same intuitive string, Daniels’ intuitive critique does not explain the driving intuition on principled grounds of the severely bad outcome, as in the examples mentioned by Knight and Voigt. Thus, we are better served by distinguishing the two criticisms and evaluating the bad consequences version of the harshness on separate grounds. Interpreting the harshness objection as stemming from the severely harmful outcomes lends itself to particular kinds of luck egalitarian responses, because if the harshness stems from bad consequences, then luck egalitarianism should seek to find ways of addressing that particular aspect of the situation. Luck egalitarians typically employ a pluralist or revisionist strategy to cope with bad consequences.

There are several pluralist responses in the luck egalitarian literature. In these, combining luck egalitarianism and another value (distributive or non-distributive) provides reason to compensate those who are faring very poorly. Such strategies include the luck prioritarianism developed by Arneson, which combines responsibility-sensitivity with a moral concern for the worst off (Arneson, Reference Arneson2000). Knight and Segall, respectively, have explored similar avenues (Knight, Reference Knight2009; Segall, Reference Segall, Knight and Stemplowska2011). Another pluralist strategy combines luck egalitarianism with sufficientarian concerns or concerns for people's basic needs (Segall, Reference Segall2010; Pedersen and Midtgaard, Reference Pedersen and Midtgaard2018).

We can also identify revisions of luck egalitarianism that explicitly address this interpretation of the harshness case. The so-called all-luck egalitarian position, recently proposed by Knight, holds that people are owed the expected outcome of their choices (Knight, Reference Knight2013). Knight highlights one advantage of this position as being how it allows for redistribution among those who take similar risks, which therefore allows us to improve the situation of those whose risk-taking puts them in dire straits.

There is an important difference between the luck egalitarian replies surveyed above, which becomes apparent once we consider the case of Bullet Bob. Bob is fully responsible for pursuing a career as a Russian-roulette player. He participates in so-called ‘thin-bullet’ tournaments. Thin bullets are extremely thin and therefore more likely to pass through the head, and participants who lose the game are thus more likely to be severely injured than dead. Predictably, Bob loses at some point and ends up very badly off.

A revisionist strategy, such as the all-luck egalitarian position, would provide people with the expected values of the risk they take. As the case of Bullet Bob demonstrates, however, this will not eliminate all of the cases in which people end up very badly off.Footnote ⁸ Thus, if bad consequences are driving harshness, the revisionist strategy cannot fully rescue luck egalitarianism. A pluralist luck egalitarian response would hold that although Bob's outcome is not in principle unjust since it perfectly reflects the exercise of responsibility, the severity of the situation provides another moral (prioritarian or sufficientarian) reason to help him. However, while this strategy serves to supplement luck egalitarianism to avoid bad outcomes, it does allow bad outcomes as an implication of luck egalitarian justice. This need not be an embarrassment for a luck egalitarian, and the pluralist defence is therefore widely acknowledged. As we shall see in the following sections, however, there are two other distinct versions of the harshness objection to which the pluralist strategy will not suffice. In the next section, we discuss harshness as disproportionality before turning to the inconsistency concern.

Harshness as disproportionality

In this section, we introduce another version of the harshness objection, which revolves around the claim that luck egalitarianism allows for disproportionally harsh outcomes for the imprudent, considering the content of their imprudence. Fleurbaey offers the most prominent version of this in his example of the motorcyclist, Bert (see above), intended as a critique of Arneson's equal opportunity for welfare (Fleurbaey, Reference Fleurbaey1995). Voigt also lists ‘the relationship between the probability that a certain outcome will occur and the severity of that outcome’ as one of the reasons why harshness arises (Voigt, Reference Voigt2007: 392). In this version of the harshness objection, the wrongness pertains to a disproportional relationship between the imprudence of the act and the consequences suffered.Footnote ⁹ It matters for our evaluation that we shift the focus from the badness of the consequences to whether the consequences reflect the risk taken. Had the imprudent suffered a loss of welfare proportional to their mistake, we would not find the outcome too harsh. ‘Our society’, Fleurbaey observes, ‘cures its criminals when they are injured, because this kind of physical penalty is not considered appropriate. In light of this example, the equal opportunity principle looks rather primitive’ (Fleurbaey, Reference Fleurbaey1995: 41).Footnote ¹⁰ On the disproportional version, then, luck egalitarianism is too harsh because it allows outcomes to be considered just, where the imprudent suffers worse circumstances than would be a proportional effect of their fault or choice.

Considering harshness as disproportionality is markedly different from the two versions of the harshness objection already presented. Unlike the counterintuitive implication version, the disproportional version gives a clear statement about what luck egalitarianism fails to do. Unlike the bad consequences version, it is not the suffering as such that matters, but that luck egalitarianism allows suffering that is disproportional to the actions of the disadvantaged person.Footnote ¹¹

Because of these differences, luck egalitarians who wish to defend against this version cannot simply appeal to intuition to support responsibility-sensitivity as they did in response to the counterintuitive implication version. This is because the disproportional version already accepts responsibility-sensitivity. If the problem with harshness cases (e.g. Fleurbaey's motorcyclist, Bert) is that the outcome suffered by the driver is disproportional to the mistake he made, this must imply that responsibility-sensitivity in some proportion is appropriate for distributive justice. It also implies that we can imagine situations involving extremely bad consequences for the imprudent as a result of a proportionally bad exercise of responsibility which therefore does not disturb justice from the perspective of proportionality. Recall the example above of Bullet Bob as a case where there is no disproportion between the risks of the acts undertaken and the very bad outcome. If what is bad in the harshness case is about disproportion, then Bullet Bob is not a genuine harshness case. It follows that the disproportionality concern is importantly distinct from the intuitive dissatisfaction with responsibility-sensitive narrowness and the badness of consequences that mobilized the two above-mentioned versions of the harshness objection. In fact, when Daniels objects that most of us would think that we would ‘have an obligation to fix the leg broken in a skiing accident simply because the skier cannot function normally if we do not’ (Daniels, Reference Daniels2008), he is making an assumption about justice that is fully incompatible with the responsibility-outcome proportionality to which Fleurbaey is appealing. Hence, although the two often go together as a package of criticism against luck egalitarianism under the heading of harshness, they are expressions of distinctive – even incompatible – concerns.

Moreover, as many contemporary luck egalitarians have argued, the principle of luck egalitarian justice is compatible with internalizing costs on risk-takers to make the consequences of imprudence fit the particular act of the imprudent (Williams, Reference Williams, Dryzek, Honig and Phillips2006). As a response to the disproportional concern, this strategy reminds us to distinguish between the principle of luck egalitarian justice and the political arrangements that this principle would recommend. One option would be to tax imprudent behaviour, such as the levying of fines for reckless driving to fund compensation and treatment for the bad outcomes resulting from people's exercise of responsibility; we can pool the costs of imprudence and divide them fairly among the imprudent (Lippert-Rasmussen, Reference Lippert-Rasmussen2001). Another kind of revision,Footnote ¹² which allows luck egalitarians to endorse a variety of institutional measures to hold people responsible, arises from Olsaretti's argument that a complete luck egalitarian theory needs a principle of stakes.Footnote ¹³ According to Olsaretti's account, luck egalitarianism needs a theory of responsibility, which, roughly speaking, provides the grounds for assessing whether people are responsible for their situations. Olsaretti notes, however, that the luck egalitarian ‘commitment to holding people responsible if their choices meet certain conditions does not settle the question of what costs they should be held responsible for’ (Olsaretti, Reference Olsaretti2009: 171).Footnote ¹⁴ The task of assigning the institutional response is carried out by a principle of stakes. A principle of stakes assigns costs to those deemed responsible. While both luck egalitarians and their critics talk as if there is a natural, deterministic relation between people's actions and the outcomes for which they are to be held responsible, Olsaretti argues that conducting the discussion in this manner loses sight of the fact that having a responsibility-sensitive view on justice is compatible with assigning various consequences to acts for which people are responsible.

In her discussion of this topic, Stemplowska offers the example of a policy that confers some disadvantage on those who show up late for work (Stemplowska, Reference Stemplowska2013: 404–405). Assuming we excuse those who are not responsible for being late, what kind of disadvantage should then befall those who are responsible for being late? A wide range of responsibility-sensitive principles could kick in. People could lose their jobs, have to work twice the minutes they were late, or be docked some of their pay. Their fines could be equivalent to the number of minutes they were late, or they could be attached to predetermined intervals (e.g. with the fine increasing every tenth minute a person is late). These principles of stakes could be evaluated, and they would probably be deemed appropriate in different circumstances. However, the important point is that they are all responsibility-sensitive, and they are all attempts at matching people's exercises of responsibility to their disadvantage.Footnote ¹⁵ A proper concept of stakes would be able to fend off the proportionality version of harshness, as critics cannot then invoke harshness against luck egalitarianism in the form of failing to proportionally balance outcome with responsibility. The principle of stakes would ensure proportionality. The ability of luck egalitarianism to avoid harshness once a proper principle of stakes is specified is something several authors have noticed as an advantage of understanding luck egalitarianism as requiring a principle of stakes (Dekker, Reference Dekker2009; Olsaretti, Reference Olsaretti2009; Stemplowska, Reference Stemplowska2009).Footnote ¹⁶ Olsaretti highlights how there is in fact no natural way to determine how to hold Bert responsible after his crash, even when we agree that he is responsible for the situation (Olsaretti, Reference Olsaretti2009: 172).

This section has elaborated on the version of the harshness objection that takes disproportionality between outcome and exercise of responsibility to be the main driver of the objection. We have given examples of the objection and discussed possible luck egalitarian responses. Our analysis revealed that the disproportionality version involves a distinct concern that is incompatible with the intuitive drive of the counterintuitive implication version. The next section discusses the inconsistency version of the harshness objection, which we argue the literature has unfortunately neglected.

Harshness as inconsistency

Luck egalitarians often respond to Anderson's concern in a manner that (implicitly) assumes it to revolve around bad consequences, disproportionality, or as if it is merely an intuitive test of luck egalitarianism's egalitarian credentials. While all of these interpretations are interesting and rightfully influential, they are not exhaustive in terms of what the point of the harshness objection could be taken to be.

Another plausible version of the harshness objection is overlooked, however, and we wish to bring it to the fore. We call this version the inconsistency version and identify it as a central aspect of Elizabeth Anderson's well-known critique of luck egalitarianism, which is the critique that luck egalitarians have been most occupied with arguing against. The central sentiment of this critique is that the harshness objection shows luck egalitarianism to contradict the idea that people should be treated as moral equals, an idea affirmed by luck egalitarians of all stripes. In this section, we flesh out Anderson's concern and discuss how luck egalitarianism can provide a reply. Note that even if some may disagree with the notion that the inconsistency version is important to Anderson, the significance of this interpretation does not hinge on its connection to Anderson's work.

As demonstrated in the discussions above, luck egalitarians have developed various reasonable responses to the harshness objection in its other versions. However, when discussing only these, luck egalitarians overlook the attack on luck egalitarianism for being inconsistent. Luck egalitarians are hardly to blame for missing the mark on this, because Anderson's criticism takes the form of a giant pool of mixed-together objections, and it is difficult to disentangle the relevant concerns. However, we must stress that it is important to understand every single version of the harshness objection in its own right if we are to evaluate the strength of the objection as a criticism of luck egalitarianism and the strength of the various responses to it.

Anderson begins the section in which she unfolds the problem of the abandonment of negligent victims by referring to the background assumption for Dworkin's contract theory; that is, that the state must treat every citizen with equal respect and concern. Dworkin not only assumes this to be so but grounds it on a generic metatheory of ethical individualism, implying that every person is of equal moral importance (Dworkin, Reference Dworkin2000: 5, Reference Dworkin2005: 272–73). Anderson then notes that this background assumption is not only Dworkin's; it would be accepted by all egalitarians (Anderson, Reference Anderson1999: 295). We believe that few luck egalitarians would deny this and therefore call this the general moral equality assumption. Anderson explicates that, according to her version of the harshness objection, ‘the reasons luck egalitarians offer for refusing to come to the aid of the victims of bad option luck express a failure to treat these unfortunates with equal respect and concern’ (Anderson, Reference Anderson1999: 295). This formulation is critical in developing the inconsistency version of the harshness objection. This version not only implies that luck egalitarianism allows harsh outcomes for the imprudent; broadly understood, it implies that luck egalitarianism is inconsistent with the background assumption – credited to Dworkin – of treating everyone with equal respect and concern. This inconsistency arises in one of the following ways. Either the failure pertains to luck egalitarianism failing to provide help for those responsible in harsh circumstances, for reasons that fail to show equal respect and concern. Alternatively, the failure pertains to luck egalitarianism providing such help, but similarly for reasons or in ways that fail to show equal respect and concern. In a nutshell, according to Anderson's objection, if you are an egalitarian, you should be sceptical about luck egalitarianism based on its political implications. Thus, Anderson's distinctive challenge is that one cannot consistently accept the egalitarian background assumption of equal concern and respect while at the same time accepting the distributive implications of luck egalitarianism. Anderson underscores this inconsistency as follows.

This passage illustrates how, in Anderson's writings, the various versions of the harshness objection are not disentangled from each other. It can be read as merely the counterintuitive version. It clearly includes the other versions of the harshness objections. The badness of consequences (in the concern that members of society are ‘sinking to such depths’) and concerns about disproportionality (in the notation that ‘even the imprudent don't deserve such fates’) are employed to strengthen the critique. In the end, however, the claimed inconsistency is indeed driving this version of the harshness objection. Hence, it is the reference to the lack of respect which seems to capture the other concerns as well in the broader inconsistency claim that luck egalitarians fail to treat people as moral equals. While these different versions can be difficult to disentangle, important differences remain. Below we demonstrate how each of the versions discussed thus far differs from the inconsistency version. While this version shares traits with the already presented versions of the harshness objection, it remains importantly distinct from them.

Consider, first, the counterintuitive version. The inconsistency version differs from the counterintuitive implication version in that it attacks the coherence of the luck egalitarian framework not merely how intuitively appealing its implications are. The inconsistency version also differs from the counterintuitive version because the latter refers to and relies on an assumed intuitive reaction to cases of harshness, whereas the latter points out that these are incompatible with fundamental sentiments of luck egalitarian distributive justice. If the inconsistency version was merely a more elaborate form of the counterintuitive version, we could easily reject this concern for the reasons examined earlier. However, the inconsistency version of the harshness objection invokes a completely different challenge. To illustrate, recall the family of replies of the rejection strategy, which merely denied the intuition behind the counterintuitive version. Such a strategy comes with a much higher cost if what the luck egalitarians accept is in fundamental tension with the kind of egalitarianism upon which luck egalitarianism is founded. According to the inconsistency version, the flaw in luck egalitarianism is not that it allows harshness in particular cases, but that it cannot treat people with equal concern and respect while insisting on a specific relation between the exercise of responsibility and distributive outcome. Hence, the inconsistency version voices a more fundamental critique of luck egalitarianism than the counterintuitive version.

We can see why it is also important to keep the inconsistency concern apart from the badness of consequences version of the harshness objection if we look at how luck egalitarians defend against Anderson's critique. If the driving concern of Anderson's objection was bad consequences alone, Lippert-Rasmussen offers a brilliant counterargument (Lippert-Rasmussen, Reference Lippert-Rasmussen2016: 191–92). Recall that Anderson's own favoured alternative is a principle, or ideal, of democratic equality whereby individuals stand in relations of equality to each other – there is no social hierarchy that constitutes a basis for disrespect. Now, Lippert-Rasmussen asks us to imagine a republic in which all citizens live under the same harsh conditions and, consequently, stand in a relationship to each other of perfect equality. Then imagine that a philanthropist offers all of the citizens a one-off vaccination programme that will make them relevantly better off. Imagine also that the majority accepts the invitation, while a small racial minority declines, let us assume, for religious reasons. Lippert-Rasmussen also asks us to imagine that the segregation between the vaccinated and non-vaccinated will result in differential status and social stigma. It then seems as though Anderson's ideal of democratic equality, Lippert-Rasmussen argues, ‘implies that it would be unjust if the majority were to get themselves vaccinated and thereby escape their harsh conditions’ (Lippert-Rasmussen, Reference Lippert-Rasmussen2016: 190). Consequently, Lippert-Rasmussen concludes, Anderson's attack on luck egalitarianism cannot be based on the implication of harsh outcome alone, because the same implication can be derived from her own alternative.

But if the problem identified by Anderson's harshness objection is, as we have suggested, not harshness per se but rather inconsistency, then Lippert-Rasmussen's counterexample does little to get luck egalitarianism out of the woods. The attack on luck egalitarianism points not to the occurrence of bad consequences but to the disrespect in terms of refusing to help people when such help is easily available, and to the resulting incoherence expressed by this refusal when the distributive principle is allegedly founded on values of equal concern and respect. Furthermore, the difference between understanding harshness as bad consequences and harshness as an expression of inconsistency also illuminates why and how Anderson rejects a number of proposals, which would, to varying degrees, remove the bad consequences. While mandatory insurance or paternalistic policies may eliminate the bad consequences, they do not remove the stain of treating people with a lack of respect, which is inconsistent with moral equality (or so the argument from inconsistency would claim).

The inconsistency version is also different from the disproportionality version. Whereas the disproportionality version of harshness identifies a problematic relation between the exercise of responsibility and the suffered consequences, the inconsistency interpretation arguably allows for harshness even when such disproportionality is not present. Imagine that Bullet-Bob has a cousin, Bullet-Bill. Bill is as careful as Russian-roulette players come. Every fifth year he enters a rubber-bullet contest, which, due to the use of rubber bullets, has the expected value of resulting in a severe headache with every sixth pull of the trigger. Nevertheless, Bill is unlucky. The frequent interaction with the rubber bullets and the gun blasts saddles him with severe tinnitus together with a very severe rubber allergy, which makes many everyday endeavours (e.g. driving) very uncomfortable, due to the frequent rubber-to-hand contact. Bullet-Bob is surely disproportionally worse off. Assume that some policy could restore proportionality either by paternalistically preventing the outcome, by ensuring that nobody enters such tournaments uninsured, or simply by providing these services at a reduced, responsibility-sensitive, cost, which is higher than that which would be charged to those who are not responsible for having such diseases. On the disproportionality interpretation of harshness, nothing can be wrong with how Bill is treated, just as Bullet-Bob's much worse fate cannot be considered problematic as forms of harshness. The inconsistency version of the harshness objection differs from this interpretation. Even if people end up with amounts of disadvantage adequately reflecting their exercise of responsibility, there is very possibly a range of societal responses to the sufferings of Bill and Bob that do not adequately treat them as morally equal to others.

Formulated more formally, the inconsistency version of the harshness objection consists of the following claims:

1. (i) Reasonable egalitarianism relies on basic moral equality – that every person's life is of equal moral worth and importance.

2. (ii) Basic moral equality implies that the state treats every citizen with equal respect and concern.

3. (iii) Abandoning people under harsh circumstances (even if their own fault) is a case of failure to treat people with equal respect and concern.

4. (iv) Thus (from i–iii), reasonable egalitarianism will not allow abondoning people under harsh circumstances.

5. (v) Luck egalitarianism allows for abandoning people under harsh circumstances or alleviates the suffering in ways that are in conflict with basic moral equality.

6. (vi) Thus (from iv–v), luck egalitarianism is inconsistent with reasonable egalitarianism.

In summary, then, the inconsistency version of the harshness objection concludes that a reasonable egalitarian cannot be a luck egalitarian. This clearly pushes a strong challenge against luck egalitarianism. Let us consider how prominent luck egalitarian responses would fare in relation to this version of the objection.

No egalitarian would deny (i). Even utilitarianism accepts equality in this form. We also think we should accept (ii). This is Dworkin's starting point, and, as mentioned, Anderson believes that it applies to all egalitarians. The last premise (v) is given by the harshness case, so we cannot dismiss it. This leaves us with (iii) and (iv). However, since (iv) follows logically from the preceding premises, it all essentially comes down to (iii). Thus, we need to pursue the following question: should we accept the third premise; that is, that allowing the abandonment of people under harsh circumstances is a failure to treat them with equal concern and respect, even when they are themselves responsible for their suffering?

First, it seems as though luck egalitarians cannot here simply accept premise (iii) – and wash off the intuition with other moral concerns as they could with the counterintuitive version. This is so, because biting the bullet of premise (iii) leads directly to either contradiction or rejection of reasonable egalitarianism. Similarly, they cannot simply change the setting of the harshness case into a priority situation between the reckless and the innocent, because the inconsistency version also demands a response to the original harshness case. That is, although it may influence our intuition in the harshness case to see that we would probably give priority to the innocent over the reckless, the case remains that luck egalitarianism should treat everyone as moral equals – even the reckless driver in the original case.

The pluralist strategy also seems inadequate due to the nature of the inconsistency version of the harshness objection. While pluralism might rescue luck egalitarianism from counterintuitive harshness and bad consequences, in the sense that it explains why luck egalitarian justice might at times be overridden by other moral concerns, such a defence is useless against an attack directly on the inherent consistency of the luck egalitarian framework. If the problem for luck egalitarianism is simply that we should avoid policies that do not offer anything to those responsible for being badly off, then surely it is an acceptable solution to refer to charity as a reason for helping those in such dire straits. However, this cannot be an attractive route if the problem facing luck egalitarianism is that it allows for policies in conflict with the basic requirement of equal respect and concern. While bringing in another value to justify a policy helping people in dire circumstances may readily alleviate the concern that luck egalitarianism allows people to suffer bad consequences, it does seemingly confirm that luck egalitarianism in itself does not treat people as equals in the relevant manner.

The revisionist strategy is more promising; at least some versions of it fare better. First, consider the idea of all-luck egalitarianism. If people are owed the expected value of their choices, we have provided a luck egalitarian reason to compensate those whose risky choices turn out bad. However, for those where the expected value of their choices is a very low level of disadvantage, this strategy offers little. The only way out would be to argue that Bullet Bob and his like have lost their claim to be treated for their disadvantage; that treating them as equals means not helping them. The already mentioned stakes strategy might provide a halfway house between pluralism and revisionism. The luck egalitarian principle is not altered; it is supplemented. Then again, the supplement is not a principle asking us to compromise on responsibility-sensitivity. Instead, the stake principle helps us to choose among various ways of treating people according to their exercise of responsibility. Some of these may directly contradict treating them as equals, but not all. It seems, then, that luck egalitarians must reject premise (iii). The most promising way for them to argue against the inconsistency version is to argue that rescuing the imprudent from harsh circumstances is not a requirement of basic moral equality if the harshness is a foreseeable result of their own exercise of responsibility, and therefore it is indeed not a sign of disrespect or unequal concern if the state refuses to assist the imprudent in such cases. While many luck egalitarians would plausibly in fact believe this, they rarely provide any argument in its support. This underscores our suspicion that the inconsistency version is very often overlooked in the literature.

To provide an argument of this kind in defence of luck egalitarianism changes the site of the disagreement between luck egalitarianism and its critics, because the question is now not when some distributive result is too harsh but what it means to treat others with equal concern and respect. Consequently, the strength of the inconsistency challenge now depends on what equal concern and respect comes down to. Anderson is not perfectly clear on this question, but she does provide some thoughts that might prove helpful. At one point, she writes about how ‘it is no great insult for a state to pass laws requiring the use of seat belts, so long as the law is democratically passed. Self-respecting people can endorse some paternalistic laws as simply protecting themselves from their own thoughtlessness’ (Anderson, Reference Anderson1999: 301). Apparently, it is not disrespectful to prevent people from driving without seat belts, because most would, out of respect for themselves, accept such a restriction on their own opportunity space. What seems core here to Anderson's account of respect is some golden rule-ish type of reciprocity. Whether I am acting disrespectfully towards others is determined by whether I could, as a reasonable and self-respecting person, endorse that others acted in a similar way towards me. This interpretation sits well with Anderson's ideal of democratic equality, one important aspect of which is the shift in moral justification from preference-based consequentialism to deontic moral dignity, ‘basing inalienable rights on what others are obligated to do rather than on the rights bearer's own subjective interests’ (Anderson, Reference Anderson1999: 319).

From Anderson's thoughts on paternalism, we learn that the problem with luck egalitarianism expressed in premise (iii) lies not in the responsibility-sensitivity itself. It would be strange to say that no disrespect is shown in passing paternalistic laws on seatbelt use, while insisting that it is disrespectful to democratically pass a law obliging people to pay their own healthcare costs should they sustain injury from driving without seatbelts despite being warned against doing so. According to Anderson, the problem occurs because there are situations such as the harshness caseFootnote ¹⁷ – that is, abandoning the imprudent reckless driver – where a policy based purely on luck egalitarian responsibility-sensitivity would be disrespectful. This is so, we must assume, because a self-respecting person could not reasonably endorse such a policy when falling victim to its ruling herself. Anderson's version of the harshness objection would therefore not appear to be targeted against luck egalitarians’ taste for a fit between the exercise of responsibility and distributive outcome, but rather against the tendency within luck egalitarianism to be obsessed with this distributive scheme, thereby forgetting that the grounding moral ideal should be equal concern and respect and that any distributive rule should be modelled by this ideal, never stand on its own.

Thus, the point of premise (iii) can be recapped as that the harshness case is an example of an outcome allowed by the luck egalitarian distributive principle, which can never be consistent with equal concern and respect (i.e. basic moral equality). Luck egalitarianism must therefore either be revised so as to cohere with basic moral equality or be rejected. Understood in this manner, the inconsistency version works to remind luck egalitarians to hold consistency with the core ideal of moral equality. It, therefore, has its strongest application when directed against luck egalitarian theories that build on responsibility-sensitive distributive fairness as a freestanding idealFootnote ¹⁸ rather than as an implication of basic moral equality.

Reflecting on the debate thus far, we would add the following to these considerations. Luck egalitarianism requires a principle of stakes (or a reformulation) which makes it clear that it does not abandon the imprudent. The stakes principle must treat people as equals. Then again, this does not necessarily mean treating everyone in exactly the same way. However, it does at least involve giving concern to all suffering, even that for which people are responsible; to provide help to all who are suffering, even those who are responsible for doing so; and to alleviate people from the suffering in the most respectful manner possible. The adjusted harshness case in which a higher number of individuals are treated also tells us something very interesting; it tells us that there may be opportunity costs associated with following a policy of saving all.Footnote ¹⁹