Good life egalitarianism

1. Luck egalitarianism

Luck egalitarianism is now a large body of work (Arneson 1989; Cohen 1989; Dworkin 1981a, 1981b; Knight 2009; Lippert-Rasmussen 2015; Roemer 1993; Tan 2011). Despite its breadth, I think we can safely identify its core as the belief that equality requires us to rectify differences that are due to bad brute luck, while permitting differences that result from the outcome of free choice. In Dworkin’s (1981a, 1981b) initial formulation, the distribution of resources should be ambition-sensitive, but endowment insensitive. In Cohen’s well-known words, ‘I believe that the primary egalitarian impulse is to extinguish the influence on distribution of…brute luck (things one didn’t choose)’ (1989, 908). Indeed, what we conventionally call ‘luck egalitarianism’ should more appropriately be labelled ‘luck-vs-choice egalitarianism’, because the doctrine is constituted by this critical distinction. So, for example, Eric Rakowski argues that

undeserved, unwagered, unchosen inequalities warrant redress…[Whereas] people who leap from airplanes, scale cliffs, or whirl around racetracks knowingly take their lives in their own hands and cannot expect others to foot their hospital bills or aid their dependents if fortune is uncharitable (1991, 76, 79).

Before critiquing luck egalitarianism, it is important to be clear about its strengths. What is, in my view, correct and deeply valuable about luck egalitarianism is the intuition that one should not be disadvantaged compared to others due to factors that are outside of one’s control. In Larry Temkin’s powerful formulation, ‘it is bad – unjust and unfair – for some to be worse off than others through no fault of their own’ (quoted in Richard Arneson 2007, 263). Some people have blonde hair while some have brown, some are born with skin that is characterized as ‘white’ and others ‘black,’ some are born with attributes that are assigned male, others female, or intersex, some develop severe physiological impairments which others do not, some are born into rich families and countries, while most others are not. What is surely correct about the luck egalitarian view, and what should remain a central pillar of egalitarian thought, is that such factors are ‘arbitrary from a moral point of view,’ ¹ and so must not lead to significant advantages or disadvantages in people’s lives. Since all human beings are of equal moral worth and dignity, it is just flat out wrong for some to suffer worse lives than others for no good reason (presuming, of course, that such inequalities are actually rectifiable). Indeed, allowing some to live worse lives than others for arbitrary reasons is one very precise and very common mechanism for treating them as beings of lesser worth and dignity. If this train of thought is correct, as I believe it is, then the applicability of luck egalitarianism is vast, since much of human history as well as contemporary social life is about the institutionalization of forms of privilege, oppression and domination on the basis of arbitrary characteristics of gender, race, class, geography, ability and so on. (I hasten to add that just because luck egalitarianism is capable of conceptualizing these inequalities, it does not follow that it is the only or best approach for doing so.)

So while luck egalitarianism is right to draw our attention to the moral arbitrariness of many social outcomes, where it goes wrong is in legitimating the inequalities that arise from choice. The view that unequal outcomes are legitimate when they arise from individual choice is of course very common. Indeed, the emphasis on personal choice and deservingness has long been central to conservative and neoliberal thought. Since the 1980s, the neoliberal worldview has become increasingly prevalent, whereby what one ends up with in life is seen as the result of one’s personal choices. If one works hard and succeeds then one’s riches are well deserved, whereas if one ends up poor, or in jail, one has only oneself to blame. Hence when Republicans passed ‘three-strikes laws’, and Clinton slashed welfare to the poor, the primary justification was that of personal responsibility (indeed, Clinton’s welfare reform bill was called the ‘Personal Reasonability and Work Opportunity Act’). Likewise, Republican intellectuals, like Greg Mankiw (2013), routinely defend inequality on the grounds that multimillionaires morally deserve their income. ² Consider a few typical quotations from prominent conservatives. Kathy Miller (a campaign manager for the Republican Party in Ohio) states that ‘If you’re black and you haven’t been successful in the last fifty years, it’s your own fault’ (Lewis and Silverstone 2016); Presidential candidate Herman Cain tells people: ‘Don’t blame Wall Street. Don’t blame the big banks. If you don’t have a job and you’re not rich, blame yourself’ (Martinez 2016, 18); Vice Presidential nominee Sarah Palin claims that ‘acts of monstrous criminality stand on their own. They begin and end with the criminals who commit them’ (Taslitz 2011, 104); likewise British Prime Minister John Major called for harsher criminal punishments, especially against youth, on the grounds that ‘society needs to condemn a little more and understand a little less’ (Waller 2011, 283). Now most luck egalitarians would no doubt reject such extreme positions. Nevertheless, the uncomfortable fact that luck egalitarians must face up to is that in embracing the principle of choice and personal responsibility, they find themselves claiming allegiance to the same basic principle (if not the rhetoric or radicalism) which animates their conservative rivals. ³

When luck egalitarians claim that inequalities resulting from choice are legitimate, what exactly do they mean? Although the argument is rarely spelt out, the intuition is clear enough: Competent adults are in control of their choices, so the choices they make are their own ‘responsibility’. What exactly does that mean? Responsibility is of course a notoriously complex subject. We can sidestep the deep debates about moral responsibility and free will by understanding the luck egalitarian position in straightforward desertist terms: To say that competent adults bear responsibility for their choices simply means that they rightly deserve praise or blame, reward or punishment, for the choices they make. Some people choose to work hard, whereas others choose to surf all day. Insofar as the resultant differences in income are the result of such personal choices, they are ethically deserved.

What is required for individuals to be deserving of (or responsible for) the outcomes of their choices? ⁴ For someone to deserve something morally (as opposed to simply being institutionally ‘entitled’ to it), the crucial requirement is that the desert basis must be under the individual’s control. This is the control principle, which Arneson formulates as the principle that ‘one should not be held responsible for what lies beyond one’s power to control’ (2007, 269). ⁵ The rationale for the control principle is that it is unjust and unfair for people to get rewards or punishments for things that they are outside of one’s control. For example, we commonly recognize that it is unacceptable to base income on the possession of blue eyes or blond hair or familial bloodline because these are outside of one’s control and so morally arbitrary. ⁶

With this background, we can now state the central argument, which is that we should be sceptical of the extent to which inequalities based on choice are legitimate. This is because our abilities to make ‘good choices’ (to study diligently, work hard, reason well, avoid conflict, engage in healthy activities, or whatever the case may be) are due to a range of social and historical factors largely outside our control – due, in other words, to luck. The evidence that people have markedly different choice-making capacities, and arbitrarily so, is growing every day. Indeed, part of my broader claim here is that having better awareness of the relevant psychological and sociological literatures importantly enhances our philosophical understandings. Let us consider a small handful of the psychological factors which impact all of our abilities to make so-called good choices and engage in good actions.

Thus we see that at as science and social science advance our understanding of human behaviour, the realm of ‘free choice’ shrinks ever smaller. What we used to see as a simple, straightforward ‘choice’ of a student to submit their papers late, we now recognize the role of ADHD; what we used to see as the ‘choice’ of individuals to become addicted to drugs, we now better understand the salience of past trauma; what we used to see as the ‘choice’ of a person to become suddenly enraged, we are now better attuned to the ways in which repeated damaging relationships with early attachment figures can become neurologically encoded and then subconsciously activated in stressful situations later in life.

The second thing to notice about this list is that these factors are not binary things – they are spectra. All of us fall somewhere on the normal distribution curves for them all; for all of us, it is not a case of yes or no, but more or less. The most crucial point is that, due to luck (genetic, familial and/or circumstantial), human beings have very unequal amounts of mental resources. And so to reward individuals for making good choices is to reward them for arbitrarily possessing particular mental resources – and that is morally equivalent to rewarding people for possessing white skin. It is wrong to base rewards or punishments on luck, since doing so is arbitrary, unfair, and fails to treat people with equal moral worth and dignity. Of course, rewarding people for the lucky possession of mental faculties is not exactly the same as rewarding people for skin colour, since there may well be instrumental and pragmatic reasons for why society would want to reward, say, effort or productivity, in ways that do not apply in the case of skin colour. Nevertheless, the point remains that from a non-instrumental, purely moral point of view, distributing significant resources, opportunities, or life prospects on the basis of luck – whether of skin tone or mental faculties – is equally egregious.

For example, Sam grew up in a loving, stable, family, where he was exposed to books, ideas, conversation, debate; he was loved and nurtured through secure attachments; he attended schools where teachers fostered his confidence, taught him how to diligently complete tasks and praised him for doing so. He thus entered adulthood as an enthusiastic and optimistic young man, with significant self-efficacy, a developed internal locus of control, strong curiosity, diligence and cognitive fortitude. He has no trouble obeying the law or de-escalating conflicts, and eventually becomes a successful, rich lawyer. On the other side of town, Farah grows up in a household that is just as materially well-off, but significantly worse off in terms of emotional and psychological support. Her single mother is frequently absent, short-tempered and emotionally detached. As a child she experienced profound trauma of sexual abuse from an uncle. She attended a school, where her teachers ignored her, belittled her race and assumed she was stupid due to a learning disability. She entered adulthood pessimistic, with periods of severe, debilitating depression, with little self-efficacy, significant learned helplessness, possessing a short-attention span, gripped by uncontrollable bursts of anger, difficulty in concentrating on a task, finds internal deliberation difficult and uncomfortable, feels constant anxiety and fear, and suffers from chronic back pain. She frequently finds herself in conflict, occasionally leading to fights and arrests. She finds it hard to hold down a job. As a young adult, she finds herself poor, pregnant and living on welfare.

Both Sam and Farah are competent adults with agency. Both have the basic ability to reflect, deliberate and make ‘choices’. However, what they do not have is equal ability in any of these respects. Sam finds it easy to make ‘good choices’ and undertake reflective, ethical action, whereas Farah does not. But note that neither of them chose to have the personalities and proclivities that they do (Strawson 1994). And therefore, neither deserves praise or blame for the outcome, because neither is responsible, in the sense of being in control of their differing choice-making abilities. If the story had been different so that Sam’s family was rich and Farah’s poor, the luck egalitarian critique would be familiar. The story we have told, however, is intentionally different. The issue of material luck has been bracketed to enable us to see clearly the issue of luck-based differences in mental resources. Indeed, it is probably fair to say that whereas luck egalitarians have been astute in their critique of the unfairness of material luck, they have been less so in terms of their critique of psychological luck. This is a problem because psychological resources are just as important as material ones for influencing life prospects. ¹⁰

The most common belief about desert or responsibility (which I suspect is held implicitly by many luck egalitarians) is the conventional legal status quo that every adult who passes the threshold of minimal competency (i.e. is neither a child nor insane) is fully responsible for their actions, and therefore morally deserves the money they receive for work, or the punishment they receive for crime. In Daniel Dennett’s formulation,

Moral development is not a race at all, with a single winner and everyone else ranked behind, but a process that apparently brings people sooner or later to a sort of plateau of development…everyone comes out more or less in the same league. (2015, 104) ¹¹

Although this perspective is very common – indeed, it forms the foundation of our legal system – it is actually quite strange. One issue is that two people can both pass the threshold and yet have substantially different choice-making capacities (i.e. very different levels of will power, competence, effortability, deliberative capacity and so on), leading to very different kinds of choices. So to say that they are both equally morally responsible is to say that they equally deserve praise/blame for their outcomes, which is unfair, because the outcomes are due to their arbitrarily different capacities. Both Sam and Farah have passed the threshold of minimal competence – both have a minimal capacity to reflect, make an effort, and so on. But to say that they are thereby equally morally responsible is bizarre and blind – a kind of false, formal egalitarianism. It ignores all the evidence of the very real, substantive differences between them. Sam has significant (unchosen) advantages in terms of his mental resources, while Farah has all kinds of (unchosen) disadvantages. How then is it possibly fair to hold them to the same level of responsibility? Of course, Sam and Farah made choices and continue to make them – they are not robots – but they do so with vastly different resources at their disposal. ¹²

Another problem with the threshold view is that it makes no sense why two people with quite similar capacities who fall just on opposite sides of the threshold are viewed as being in fundamentally different camps, whereas two people who both surpass the threshold, yet are much further apart in terms of their capacities, are judged to be in the same camp – fully deserving of praise or blame. Given that, empirically, our capacities exist on spectra, the belief that there is one and only one specific threshold that matters for responsibility makes no logical sense. For all these reasons we should prefer what we might call the ‘spectrum view of human abilities’ as opposed to the standard binary threshold view.

Now it might be objected that our ability to make choices is not due entirely to luck because it is within our control, at least partially, to choose to educate and train ourselves to have better and stronger capacities; in other words, one might argue that we can always choose to become better at making choices. Hence the common (though I think incorrect) clichés that ‘winners make their own luck’ and ‘it’s not the cards you’re dealt that matters but how you play your hand’.

That response fails for two reasons. First, even if we grant that some of our choice-making ability comes from previous choices, this is only true at the margin. Granted, there is some room for conscious change, nevertheless most of us are indelibly shaped by the personalities and proclivities that we inherited from our emergence out of adolescence. We are not creators of ourselves (the exuberance of certain existentialists notwithstanding). We cannot shape our own personalities to be any particular way that we might want nor modify the deep features of our psychology as if we were sculpting a statue. For instance, we do not, generally speaking, choose our level of implicit bias, optimism, self-confidence, enjoyment of deliberation or brain chemistry. Such things are largely out of our hands, resulting from random chance or the impress of surrounding social hierarchies on our minds. ¹³ Unless one believes, per impossibile, that we are fully self-creators, the responsibility-advocate who thinks we are somewhat able to change ourselves must also accept that we are somewhat unable to do so – our personalities are due to a mix of autonomous choice and luck. But this places the responsibility-advocate in a bind. Either they simply ignore the problem and say that people should be rewarded for all their choices even though these are partially constituted by luck (which is akin to saying that rewarding people on the basis of luck – such as height or skin colour – is acceptable, if only partially so); or the responsibility-advocate must specify some way of being able to analytically separate the degree to which ones capabilities are due to autonomous choice from the degree to which they are due to luck. Yet that is a hopeless task. Luck and choice are intertwined through and through, like the way that sunlight is incorporated into the carbon structure of the wood in a tree. Or, for another analogy, consider the famous nature/nurture dichotomy. That dichotomy is now widely recognized as invalid because there is no meaning to human ‘nature’ outside of an environmental context; nurture and nature are irreducibly co-constitutive. That is how we should understand choice; it has no existence apart from luck; ‘choice’ and ‘luck’ are similarly irreducibly co-constitutive of human abilities.

A second, and even more fundamental problem is that even granting that we can choose to improve our choice-making capabilities to some (I think limited) degree, this does not rescue desert, it merely pushes the problem back a stage, since different people have unchosen and very different abilities to engage in this kind of higher-level self-improvement. So the question now becomes: Why should some receive praise over others due to their greater luck in possessing superior higher-order capacities? Some people are blessed with significant higher-order abilities to reflect on their condition and seek to alter their personality, with the focus and determination to see such changes through – whereas others lack such abilities. But possession of this kind of ability is just as arbitrary as any other. Surely one cannot claim that they chose to have the capacity to increase their choice-making abilities. It cannot be ‘choice’ all the way down. As naturalistic creatures, we do not will ourselves into being. We do not (to use Nietzsche’s evocative phrase) pull ourselves ‘up into existence by the hair, out of the swamps of nothingness’ (quoted in Martinez 2016, 13). At some point – most conspicuously, in the years leading up to adulthood – our differential abilities to engage in higher order self-improvement, like every other ability, simply emerge. They develop from the normal human processes of learning and development; they are manifestly not the result of autonomous choice. Effortability and self-improvement ability are themselves simply skills – no different from throwing a football or singing in perfect pitch. We differ in these meta-abilities just as deeply, and just as arbitrarily, as any other skill (how could it possibly be any different for empirical beings without divine or transcendental qualities?)

Consider again the case of Sam and Farah. The hardcore libertarian will want to insist that Farah has only herself to blame for how her life turned out, since regardless of her life history or social circumstance, she can always (supposedly) choose to rise above her circumstance by practicing to improve her work ethic or deliberative capacities. But that is ludicrous; she did not choose to have depression, learned helplessness, low self-efficacy and chronic pain – all of which severely limit her abilities to self-improve. Those attributes were due to bad luck. They were thrust upon her by circumstance. So to blame her for such luck (or to praise Sam for the lucky possession of his higher-order faculties) is morally equivalent to blaming her for being female or black.

Perhaps, the most sophisticated objection to my critique of choice and desert comes from John Roemer (1993, 1998, 2003). His position is complex, but the essential points are as follows: Although individuals do differ in the total amount of psychological resources they possess (e.g. in terms of the total amount of effort, or will-power, or conscientious striving, that one is capable of exerting), as autonomous adults they nevertheless all share the capacity to exert more or less effort. In other words, everyone is in control of the degree of effort that they exert. One useful metaphor for this idea comes from David Alm (2011): we can think of our abilities as a fuel tank; although we all have different size tanks, we nevertheless all have the same autonomous ability (supposedly) to use, say, a quarter, half or all of our tank. Along these lines, Roemer tells us to conceive of effort – or for our purposes, ‘choice-making ability’ – not in absolute terms (which are out of our control and therefore unsuitable for desert), but in proportional terms (which he believes are in our control, and therefore suitable for desert).

This is an ingenious attempt to ground desert claims. Unfortunately, it is not successful. The problem is that individuals are not in fact in control of the degree of effort that they exert. This is a sleight of hand; since it is true that individuals can all exert more or less effort, it is tempting to conclude that we all do so equally autonomously. But that is false. The decision to exert some, most, or all of one’s energy is itself a decision that will be largely influenced by one’s individual personality – one’s proclivities, desires, pain thresholds, temperament and so on. Does Roemer really believe that someone who is clinically depressed can choose to exert the same degree of effort as a healthy and happy person? In fact, different people have, pace Roemer, different abilities to exert not just an amount of effort but a degree of effort too. (It’s a sleight of hand because Roemer plasters over the differential ability to exert degrees of effort that each individual possesses by simply labelling all such differences with the same label of ‘autonomy’.)

If it were true that everyone, universally, shares the same ability to autonomously choose their degree of effort, then everyone must be able to exert, say, 10% of their total effort one hour, and then 20% of their remaining effort for the second hour, then 30% of their remaining effort for the third hour, and so on in this careful, disciplined, meticulous, way. But that is clearly not equally possible for everyone. Being able to choose the degree of effort that one exerts on a particular activity, or for a particular amount of time, is itself a kind of skill that different people will possess to different degrees. This proves that even Roemer’s ‘propensity to exert degree of effort’ is not a pure matter of autonomy, but is itself a matter of luck, and therefore cannot serve as a basis for moral desert.

We can see the problem by re-examining Alm’s metaphor of the fuel tank. Alm (and Roemer) conceive of autonomy as the supposedly universal ability of each of us to use more or less of our ‘fuel tank’. But this raises the question: Who is driving the car?! Empirically variable human beings or transcendent god-like beings? The purpose of the gas tank metaphor is that it is supposed to give a naturalistic account of our empirically differing capacities. But if everyone, universally, is equally capable of pushing the accelerator/brake of the car to the exact same degree, with the exact same precision as everyone else (as the metaphor implies), this would seem to bring a ghost in to the machine through the back door. ¹⁴

The bottom line is that Roemer cannot escape the fundamental fact that we cannot separate our autonomous choice-making abilities from the unchosen, arbitrary features of our selves. And so to reward or punish someone due to their ‘propensity to exert effort’ is still to reward or punish them for arbitrary features like skin colour.

At this point, the sceptical reader will surely balk: Can we really not hold people differentially deserving? If two brothers are raised in the same family, yet one chooses to work hard – getting up at 6 am every morning, working until late at night – while the other works far less, choosing to spend much of his time in bed, can we not say that the first morally deserves more?

The answer is no. The reason that that they behave differently is that there are myriad unchosen aspects of their personality which influence their behaviour. One brother has high levels of effortability and self-confidence, while the other has low self-efficacy and chronic pain. Such differences – as in the case of Sam and Farah – are morally arbitrary because they were not chosen. Neither brother chose to have the personality he has. Both have some abilities to change, but neither of them chose the amount of meta-ability they have to self-improve. Therefore, praising the first and blaming the latter is just as morally unfair as punishing someone for, say, developing a brain tumour which impacts behaviour. We are not in charge of the elements that make us us. The implication of all this is that the goal of distributive justice should not be to establish just deserts. (Below I argue that a better goal is to provide material and social conditions for all to acquire a good life, regardless of our choices.)

The most common objection to an account like mine is concern over what it means for criminal responsibility. Are we no longer able to blame and punish a criminal for, say, committing physical assault? Are they no longer responsible for their choices?

The response is that while it is easy to cavalierly blame, it is notoriously difficult to establish someone’s actual degree of criminal responsibility. Everyone recognizes that children and the insane are not criminally responsible. But once we drop the naive binary threshold view (which says that everyone else has full and exactly the same level of responsibility), we recognize that humans come in all shapes and sizes, that moral ability to make ethical decisions and follow the law, like any other human capability, exists on a spectrum, coming in vastly differing amounts, which different people possess arbitrarily. Some are mentally ill. Some find it very difficult to contain their anger. Some are less able to reason through the consequences of their actions. Some have been socialized into violence and rely on it habitually. Some have faced years of sexual or racial trauma that builds up to the point where one is liable to explode at the latest microaggression, and on and on. As soon as we start to inquire seriously into one’s moral deserts, we are sent down the rabbit hole of needing to separate the degree of one’s actions that one can be held responsible for, versus the degree to which one cannot. But that is a dead end. It is a fool’s errand to think that we can ever burrow deeply enough into people’s brains or backgrounds to establish clear and accurate measures of genuine culpability. We will always be at a loss to know how to weigh one’s true moral deserts vis-à-vis others, and therefore at a loss for attributing a fair, proportionate punishment. If we simply shrug our shoulders and continue to punish heterogenous people according to a homogenous standard, we will thereby be punishing them for things that are out of their control (we might as well punish people for the colour of their eyes or their sexual preferences).

None of this is to say that we should do nothing when crime occurs. The point is that we should change our focus. We should shift from seeing criminal justice as about establishing just deserts, to instead seeing the goal as preventing such harm from reoccurring, by inducing people to change their behaviour in the future. This could mean swapping judicial tools away from retributive ‘eye for eye’ punishment, towards rehabilitation, transforming the social conditions that cause crime in the first place (poverty, marginalization, childhood trauma, etc.), and if necessary, using judicial punishments not as retribution but as future deterrence.

Let us sum up. What is powerful in luck egalitarianism is its sensitivity to luck and arbitrariness. What is problematic is the belief that ‘choice’ provides legitimate grounds for inequality. This is wrong because choice-making capacities are largely outside of one’s control; they are largely due to luck. Therefore, any attempt to reward ‘effort’ or ‘reason’ or ‘deliberation’ or ‘conscientious striving’ (even if useful for pragmatic reasons, which we discuss below) is morally akin to rewarding individuals for possessing natural talent or a large inheritance or white skin. This is not to deny the existence of autonomy. Most people do have autonomous abilities. They can exert effort and make moral judgements. However, these are all empirical skills, meaning that individuals differ mightily in their possession – some have more, some have less. And most importantly, the possession of differential skill in all these areas is due to luck and so morally arbitrary. We can never separate a pure autonomous part of our self from unchosen luck. For such reasons, we should be immensely sceptical of the claim that choice legitimately grounds inequalities.

The upshot is that the luck egalitarian embrace of ‘choice’ and responsibility took egalitarianism in the wrong direction. Unsurprisingly, bringing in a central tenet of conservatism makes the theory more conservative – more harsh and uncompassionate, more individualistic and less sensitive to historical and social forces. Instead, I have argued that egalitarianism should move in the opposite direction: we should reduce the emphasis on ‘choice’ and responsibility by being more sensitive to the degree in which social circumstance, as well as chance and caprice, affects us all – our abilities, our accomplishments – in arbitrarily different ways. This shift in perspective allows us to be more caring and compassionate to those who ‘choose’ poorly and so end up with less of the good things in life.

Note that it is entirely possible for luck egalitarians to agree with this critique. They can admit that, ‘to the extent that choice-making capabilities really are due to luck, we agree that they do not legitimate inequalities’. (One can easily imagine GA Cohen, for instance, taking such a position.) Indeed, I believe that a consistent luck egalitarian must, in following the psychological and sociological evidence, follow this path. If I am right about this, then we see that luck egalitarianism is actually a much stranger doctrine than is typically realized, in that taking it seriously leads to it being undermined. Paradoxically, one cannot actually be a consistent luck egalitarian (at least in the conventional sense that presupposes a clear and crisp distinction between luck and choice).

The question then becomes, given that luck egalitarianism (or what I called ‘luck-vs-choice egalitarianism’) is based on the dichotomy of ‘luck’ versus ‘choice’, what is left of the theory when one of its two foundational categories is discarded? Once we have cleared away the debris, I am not sure whether a coherent theory remains (are all inequalities now rendered illegitimate, no matter what?). In the real world, of course, people will always make choices, and so ‘choice’ will invariably remain a central grammar of people’s lives. However, what is not necessary is for choice to be a guiding principle of egalitarian social policy. What is clearer is the direction that egalitarianism is pushed: the arguments of these pages naturally lead us to focus less on the ‘choices’ people make and more on the things that they need. Luck egalitarianism thus ends up pointing us towards a different kind of egalitarianism, one that we develop below.

2. Relational egalitarianism

Let us now briefly turn to relational egalitarianism. While this too is now a broad field, the central idea is that the focus of egalitarianism should not be luck and choice but the character of social relationships. The general goal being to replace hierarchical and disrespectful relationships with democratic and respectful ones (Anderson 1999; Fourie et al. 2015; Scheffler 2003, 2005; Schemmel 2011).

There is much to admire in the writings of relational egalitarians. For instance, Elizabeth Anderson (1999) and Samuel Scheffler (2005) have given us penetrating accounts of the problems that arise from leaning too heavily on the notion of individual choice – such as harshness and moralism. They are also surely right to emphasize the importance – indeed the centrality – of relationships for any viable theory of equality.

However, the shift away from luck and choice has not been an entirely positive development. I want to highlight two basic problems with the relational egalitarian framework.

The first problem is that some of the inequalities that we face are not primarily about relationships but are constituted by unfair distributions of opportunities to flourish and lead a good life. Opportunities are a fundamental dimension of equality, just as important as relationships, though relational egalitarians tend to ignore them – especially when the opportunities in question are not clearly connected to issues of disrespect or domination.

For instance, consider the policy goal of accessible (free or at least cheap) university education for life. The reason why many progressives support this goal is for the opportunities that it offers individuals to flourish – not only in terms of getting good jobs but in terms of expanding one’s horizons and developing one’s capabilities. Relational egalitarianism, on the other hand, has no clear basis for advocating for this policy, because it’s hard (though perhaps not impossible with sufficient stretching) to argue that a lack of opportunity to go to university necessarily leads to disrespect or domination, since surely many people are very happy to bypass university and do not suffer disrespect or domination because of it. So Anderson argues that relational egalitarianism requires only that citizens have ‘enough human capital to function as an equal in civil society’, which she says requires only ‘a high school diploma or its equivalent’ (2007, 615, 620). The problem with such an approach is that seeing access to university as solely about its impact on relational issues of democracy or domination, misses its importance in terms of individual opportunity. ¹⁵

Much the same could be said for the issue of free time. A long-standing goal of the socialist left has been to create the material conditions for everyone to possess free time (undergirded by economic security), so that regular people become increasingly free to devote themselves to whatever matters most in their lives – art, music, philosophy, family, athletics or whatever it may be. Here again, the heart of the aspiration for egalitarians of this sort is not avoiding disrespect or domination, it’s not relational at all, it’s rather about having the opportunities to flourish, to ‘live deeply and suck the marrow of life’, as Thoreau ([1854] 1959) would say. What relational egalitarianism misses, in other words, is that the goal for progressives cannot be reduced to the absence of domination; it is just as much the presence of opportunities to flourish – so that everyone, not just the privileged, can access beauty, joy, meaning and purpose, so that everyone can live freely and fully, sufficiently secure in the material foundations of their lives to be able to explore the inner and outer reaches of their souls.

The second and, for our purposes, the main problem with relational egalitarianism is that is permits significant inequalities between people on the basis of brute luck. Consider again the case of the wealthy Sam and the impoverished Farah. Relational Egalitarianism will want to aid Farah insofar as she is being disrespected or dominated. But that is as far as the doctrine can go. Insofar as she simply has a worse life, even a much worse life (in terms of resources, well-being, opportunities or other metrics that are not strictly relational) due to misfortune, relational egalitarianism has nothing to say.

The problem here is particularly acute in the case of individuals with different abilities and impairments. For example, Jack and John are friends of similar age and abilities, just about to graduate university and looking forward to full, flourishing lives, when their car is hit by a drunk driver. Jack suffers partial brain damage, leaving him with reduced IQ, reduced speaking abilities, stiff and awkward body movements, as well as chronic fatigue. John also hits his head, but the accident does no damage, and actually leaves him with a much-improved photographic memory ¹⁶ (in addition to his already high IQ, developed social skills and high levels of bodily fitness and energy). The consequence is that, in the years that follow, Jack fares much worse in the market economy than John. Jack ends up working part-time stacking shelves in a grocery store, whereas John becomes a corporate executive. John earns one-hundred times greater income and accumulates one-thousand times more wealth, resulting in a much happier, more flourishing life.

What can relational egalitarianism say about this inequality? If Jack is being disrespected or dominated, then relational egalitarianism has much to say. But it’s easy to imagine situations where he is not; perhaps he works at a minimum wage job where he is protected by a strong union and treated kindly and empathetically by his managers and co-workers. In such cases, where Jack is not dominated or disrespected, but simply much worse off, relational egalitarianism has nothing much to say. So it appears that relational egalitarianism has limited resources for objecting to a society where the ‘productive’ earn 10-times, 100-times or even 1000-times more than the ‘unproductive’ due to mere luck. Of course, at some point, they can say that very high levels of inequality are a risk to democracy and so should be limited, but that is weak tea indeed, and scarce comfort to the relatively impoverished Jacks of the world.

The root problem here is that relational egalitarianism has lost the essential insight of luck egalitarianism: that it is unfair and wrong, deeply wrong, for individuals to suffer significantly worse lives than others through no fault of their own. Relational egalitarianism has no resources for saying the obvious thing: that significant differences in life opportunities resulting from brute luck are offensive. To shrug at the fact that Jack leads a significantly worse life than John due to differences in brute luck is to fail to treat them with equal moral worth and dignity. It is morally equivalent to rewarding and punishing them for differences in the colour of their skin. A just society would therefore differ significantly from relational egalitarianism in this regard. It would aim to provide Jack and John with as similar resources and life opportunities as possible. Granted, it will likely not be possible to do this entirely within the framework of a market economy (we discuss this issue further below), but it is surely possible to dramatically shrink the inequalities that do exist, increasing the entitlements provided for Jack, paid for by redistributing the income of John. Perhaps, differences of the order of 5:1 or 10:1 are necessary for markets to function, but those of the order of 100:1 and 1000:1 are not. ¹⁷ In contrast to relational egalitarianism, a just society would aim to reduce such luck-based inequalities as much as feasible.

3. Good life egalitarianism

Given what we have seen of luck egalitarianism and relational egalitarianism, the question becomes: How can we retain the strengths of these theories (in particular, the emphasis on the importance of luck-based arbitrariness, as well as the importance of social relationships in impacting good lives), while avoiding their weaknesses (in particular, the focus on choice and responsibility)? My answer is that we should conceptualize equality in terms of what I call GLE: given the arbitrary differences in individual capacities – including choice-making capacities – equality should not depend on ‘good choices’, but instead should aim to guarantee to all a robust and roughly equal distribution of the essential goods necessary for living a good life.

What are these essential goods? That is a very difficult question, based as it is on one’s perspective of what constitutes good and valuable human lives. Here I will only suggest (but not defend) the proposal that we should recognize two main categories of essential goods. First, everyone should be assured good opportunity rights, that is, the means to self-determine, so that they can live the lives that they autonomously value. The essential prerequisites for this include things like physical security, civil rights, education, economic security and free time. The level at which these things can be provided will depend on context, but in the rich countries of the Global North, ¹⁸ they should include, above and beyond the standard package of liberal rights, things like universal healthcare, access to free (or heavily subsidized) post-secondary education, income security (such as through a Basic Income and extensive public services such as affordable housing, public transit, universal childcare, and pensions) and rights to quality, part-time work (Malleson, forthcoming), all of which must be accessible to differently abled bodies. Second, everyone should be assured good relationship rights, that is, the goods necessary for ensuring egalitarian social relationships (so that workplaces, families and the state treat their members with equal respect, equal status and, where appropriate, equal democratic rights). ¹⁹ The goal of GLE, in other words, is to move much closer towards a society of equal outcomes, by raising the bottom through increased provision of the essential goods, paid for by raising taxes to lower the top. Such a society might be envisioned as a market socialist society, with high levels of redistributive taxation (so that income differentials are compressed to an order of perhaps 5 to 1 or 10 to 1), significant market regulations, extensive public services, a Basic Income, widespread workplace democracy, with typical employment being that of flexible, quality part-time work of, say, 15–30 hours per week. The details are beyond the scope of this article, but the important point here is that GLE envisions a robust social minimum, accessible to all. Not merely a meagre safety net, but a much more significant guarantee of the conditions for a comfortable, relatively affluent, flourishing life for all.

The foundation of GLE is the belief that the distribution of capacities to make good choices is deeply arbitrary. It follows from this that the distribution of essential goods should not be based on ‘choice’ but rather on need. In this way, GLE breaks apart the two dimensions of luck egalitarianism – it maintains the focus on luck and arbitrariness, while bracketing the role of choice and responsibility. This shift from ‘choice’ to ‘need’ is analogous to the shift that many progressives have long advocated vis-à-vis criminal justice, urging us to shift from backward-looking punishment to forward-looking rehabilitation. Just as an improved criminal system would, arguably, be one that focuses less on punishment and more on rehabilitation (Alexander 2010; Davis 2003; Rusche and Kirchheimer 2003), saying ‘It is not your past criminal choices that matter, but what you need to be able to live a crime-free life in the future’, so should a just distributive system be one that increasingly says, ‘It is not your past economic choices that matter, but what you need to be able to live a good, flourishing life in the future’. GLE is thus not a desert-based view of distributive justice but a needs-based one (where the essential needs are defined through democratic deliberation by society at large ²⁰ ).

What role, then, would choice actually play in such a society? People, of course, make countless choices all the time. We choose who to date and partner, how much schooling to undertake, how many hours to spend at the office, whether to have kids, whether to save or consume and on and on. These choices ineluctably lead to differences between people. The difficult question is deciding when such differences are problematic enough that we wish to call them ‘inequalities’, and aspire to mitigate them. Every egalitarian theory must provide an answer to this fundamental question, and the force of that answer will carry significant weight in terms of the persuasiveness of the theory as a whole. For GLE, the answer is that there are certain areas that are so fundamental for good lives that they must not be losable, regardless of ‘choice’. These are the essential goods (civil rights, income security, healthcare, anti-discrimination policy and so on). Essential because they are the foundation for good lives. They are the basic entitlements of everyone. Of course, no one is forced to avail themselves of such goods if they do not wish to, but access to them is guaranteed (with one caveat to be addressed shortly). This means that, in contrast with some of the more conservative luck egalitarians, under GLE individuals like Bert who ‘choose’ to not wear a motorcycle helmet and then get into an accident would still be guaranteed healthcare. ²¹ Individuals who have ‘chosen’ to drop out of university would always have the option of re-attending in the future for free. Those who ‘choose’ to quit their job would always be provided with basic economic security, and so on.

Does this mean that choice should have no impact in how people’s lives turn out? No. There are two vitally important countervailing issues to take heed of. First, there are personal preferences for non-essential goods. If X spends their disposable income on expensive cars, while Y prefers to take the bus so as to save more for retirement, X cannot then complain that they deserve compensation because they end up poorer than Y. As long as both possess basic economic security, no genuine ‘inequality’ results from individuals choosing to spend their disposable income differently. Likewise, if Z chooses to live an ascetic life in a monastery, their relative poverty is not something that society needs mitigate (as long as they always have secure access to jobs and educational opportunities should they desire them in the future). The underlying rationale in such cases is that it is important for society to provide ample space for the free play of personal preferences. We do not want bureaucrats meddling and interfering in our everyday choices. And since these preferences are playing out in the realm of non-essential goods, we need not be overly worried about the differences that accrue.

Second, there are many areas in which we have strong pragmatic and instrumental reasons for wanting choices to be consequential. For example, in order for a market system to function reasonably well, it makes pragmatic sense to allow people who ‘choose’ to work in areas of greater demand, or for those who ‘choose’ to work longer hours to receive somewhat more income than others, since encouraging such choices can lead to positive benefits for society as a whole. (Though not completely fair, this is an acceptable compromise with pragmatism, as long as the top income levels are not permitted to go too much higher than others). Likewise, if a sociopath ‘chooses’ to murder someone, it may well be pragmatically necessary for society to protect itself by confining that person (though hopefully in a hospital rather than a prison cage). This is the caveat mentioned above – in rare cases like this, an individual’s choices may need to lead to a reduction in essential goods (such as the right to live freely), because society needs to protect itself (though even here the losses should be as minimal as possible, the goal should be more rehabilitative than punitive, and the perpetrators’ essential rights should be restored as soon as possible). For both the economic and criminal cases, the rationale for permitting choice to lead to differences between us is instrumental – in that it is frequently necessary to provide overarching systems of incentives and disincentives to ensure the stability and quality of social life.

It is useful to emphasize an important point here, which is that the scale of inequalities between people cannot be totally extinguished, but they can be significantly compressed. In the case of distributive justice, the market requires some inequality to function. But it absolutely does not require the huge 200 or 300:1 pay differentials currently existing between American CEOs and their workers (Reich 2015). Similarly, for criminal justice, it may be that some punishment is required to disincentivize crime. But long prison sentences, dehumanizing conditions, solitary confinement and the death penalty are clearly not. The bottom line is that GLE recognizes – as any adequate theory of justice must – a necessary tension between the fact that some inequalities are useful for pragmatic reasons, with the truth that inequalities are usually morally undeserved. The distinctiveness of GLE is in arguing that that due to the arbitrariness of ‘choice’, we should compress these inequalities as far as possible. We cannot have total equality of outcome, but we can and should aim for something quite close to it.

We are now able to answer the question of when choices should legitimately ground differences between people (versus when egalitarian social policy should attempt to intervene). For GLE, the answer is demarcated by the category of essential goods. In other words, the fundamental distinction is not between choice and luck (since ‘choice’ is such a problematic category) but between essential and non-essential goods. Access to the former are (for the most part) insensitive to choice, whereas the latter are sensitive to it. As we have seen, the rationale for such a division is twofold. On the one hand, we want extensive space for the expression of personal preferences, while simultaneously recognizing that such preferences are deeply inflected by arbitrariness, and so mustn’t undermine one’s ability to access the essential goods in life. On the other, it is socially useful in many instances to incentivize certain kinds of choices through rewards or punishments, though here too such pragmatism must be circumscribed to not undermine access to essential goods.

In contrast to other theorists (such as Nussbaum (2003)), the reader will note that GLE is not a sufficientarian theory but an egalitarian one, in that the primary impulse is towards equality of outcomes (cf. Arneson 2005; Crisp 2003; Frankfurt 1987). Whereas sufficientarians are, by definition, unperturbed by the existence of inequalities above the sufficiency threshold, GLE sees those differences as arbitrary too. The wealth of John is just as morally arbitrary as the poverty of Jack. However, as we have noted, complete equality of outcome is neither possible nor desirable, and so the ultimate goal of GLE is not strict equal outcomes but reducing inequalities as much as practically feasible. In practical terms, this requires raising the bottom (by enhancing access to essential goods) and lowering the top (through redistributive taxation).

This said, although GLE is an egalitarian theory, that should not be interpreted to mean that equality is the only goal that matters for social justice. In many contexts, particularly poor countries, the primary aim may not be equality but growth to achieve a satisfactory level of sufficiency for all. In some contexts, the goals of equality and sufficiency might coincide, but in some cases they may diverge. For instance, if for some reason the only possible way for a society to achieve a basic level of sufficiency for all is by accepting more inequality (or, conversely, if the only way to increase equality in society is by levelling down in a way that reduces everyone’s income), then there is a difficult trade-off. On the one hand, a society with more equality is desirable on the grounds that we have been studying here (reducing the impact of bad brute luck on one’s life and reducing relational domination and disrespect). On the other hand, a society with improved income for the poor is also clearly desirable. I do not believe there is a simple a priori solution here. Deciding how to make this trade-off will depend on careful evaluation of many empirical, contingent factors: Is it really true that growth requires more inequality? What are the levels of inequality (Are they extreme? Are the poor dominated by the rich?) What are the levels of poverty (How poor are the poor? Are they starving or simply less well-off?). Although I cannot defend it here, it seems reasonable to suspect that when countries reach the levels of material prosperity existing in the Global North today (and particularly when we consider the constraints of climate change), then growth should increasingly take a backseat to concerns of equality.

Let us wrap up our exploration of GLE by taking note of a few illustrative contrasts between it and the now-dominant theories of egalitarianism. Consider all of those who, for whatever reason, suffer a difficult period in their lives. A ‘chooses’ to drop out of university due to depression. B, suffering from the trauma of residential schools and racial stigma, “chooses” to seek solace in drugs and alcohol, eventually ending up on the street unable to pay their rent. The logic of luck egalitarianism (though I suspect not the heart of many of the humane really existing luck egalitarians) is to simply say ‘too bad; you made your bed; you must sleep in it’. Relational egalitarianism does somewhat better, but perhaps not much. The logic of their position is that such circumstances are only a problem – and support should only be offered – to the extent that such people risk disrespect and domination. Do they? Perhaps, but perhaps not. (For instance, what Anderson (2007) says about education makes it seem that her version of relational egalitarianism would offer little support to A. And what she says elsewhere (e.g. 1999, 325) about poverty implies that as long as B can acquire ‘basic capabilities’ and is not a ‘peon’, then no additional support is required). ²² There are also cases – like where C suffers a bodily impairment and is forced to quit their job – where relational egalitarianism actually appears more cold-hearted than luck egalitarianism, as it is unattuned to misfortune, and so would seem to allow C to be live a life that is significantly worse-off than their luckier neighbours as long as they are not disrespected or dominated.

By contrast, GLE performs significantly better. In guaranteeing to such people perpetual access to the essential goods, regardless of their past ‘bad choices’, it insists that such choices should not define their future lives. It is thereby a theory of considerable warmth and compassion, even, if it is not too much to say, of (structural) love. As we have seen, all choice-based theories (be they luck egalitarian or neoliberal) tend to be moralistic and harsh, admonishing people for making mistakes and implying that their situation is their own fault. That is no accident; such moralism flows directly from the very DNA of a choice- and responsibility-based philosophy; indeed, it is the very point. The antidote to such moralism is compassion and love, since the primary impulse of compassion is its non-conditionality, its unconditional warmth, its desire to focus not on one’s faults but on one’s humanity, emphasizing the flickering all-too-brief existence of a human life and the overwhelming need for it to be as good as possible.

GLE is thus motivated at the deepest level by two primary impulses: an emotive sense of love and compassion (that so many people live lives of pain, suffering and deprivation), with an analytical scepticism of choice (that such suffering could be attributed to their own fault or responsibility). These two impulses merge in the desire for a robust safety net, which says to all, we have your back; if and when you fall in life, it’s ok, we will support you. Like a loving family to their children, GLE says to its residents,

No matter what choices you make in your life, no matter what accidents or misfortunes befall you, you will be (non-paternalistically) cared for and supported. Even if we dislike your choices, indeed, even if, in extremis, we must occasionally restrain your choices to prevent harm to others, you will still never be abandoned to the winds of fortune.

4. Conclusion

This article has sought to carve a new path between the two dominant wings of contemporary egalitarianism: Luck egalitarianism and relational egalitarianism. Whereas Cohen saw the introduction of choice and responsibility into egalitarianism as a ‘service’, it should, in fact, be more accurately seen as a Trojan Horse. The luck egalitarian emphasis on choice and personal responsibility is misplaced because individuals differ so deeply, and arbitrarily, in their choice-making capacities. Allowing inequalities to result from ‘choice’ is akin to allowing inequalities to stem from the possession of any other morally arbitrary factor – such as skin colour or gender. Just as progressives have rightly rejected these latter as appropriate bases for distributive justice, so should we be increasingly sceptical of the salience of the former. The move towards relational egalitarianism has been a case of two-steps forward, one-step back. While the (partial) shift away from the focus on choice is salutary, the concurrent rejection of luck is problematic, given the prevalence and importance of luck-based discrepancies in opportunities to lead a good life. GLE is one attempt to retain the strengths of these theories while avoiding their weaknesses. The guiding idea is that given the unavoidable arbitrariness of human capacities, the foundation for a good life should be assured for people regardless of the actual choices that they make. The essential goods necessary for leading a good life – such as the opportunities to self-determine and to enjoy non-dominating social relationships – should be guaranteed to all. Choice will necessarily play an important role in any free and healthy society, but its impact on distributive outcomes should be highly circumscribed.