LUCK AND EQUALITY

by Susan Hurley and Richard J. Arneson

IIRichard J. Arneson

ABSTRACT Does it make sense to hold that, if it is bad that some people areworse off than others, it is worse if those who are worse off come to be sothrough sheer bad luck that it is beyond their power to control? In her contri-bution to this symposium, Susan Hurley cautions against a closely related fal-lacy: from the fact that people have come to an unequal condition throughunchosen bad luck, it does not follow that, if we aim to undo the influence ofunchosen luck, we ought to institute equality of condition. Forswearing thefallacy that Hurley analyses is compatible with answering the question affirm-atively, and more generally with holding that principles of distributive justiceshould be sensitive to the distinction between chosen and unchosen bad luck.This essay explores how this might be done.

Writing about essays on equality by Ronald Dworkin, G. A.Cohen comments that it is Dworkins notable achievementto have rescued the notion of personal responsibility from thearsenal of the right wing of the political spectrum and to haverendered it serviceable for left-wingers.1 But personal responsi-bility is tricky. Is this a weapon that is always prone to misfire,or in some other way unsuitable for egalitarians?

In her contribution to this symposium Susan Hurley brilliantlydissects a pattern of erroneous reasoning into which the politicalphilosopher who invokes responsibility may fall.2

We may be tempted to notice that the life outcomes that indi-viduals reach are largely due to the operation of unchosen luck,affirm the moral desirability of undoing the influence ofunchosen luck, and infer that equality (or for that matter anyother distributive pattern) is morally desirable. As Hurleyexplains, the inference is fallacious. From the fact that a particu-lar unequal outcome pattern is due to luck it does not followthat adopting the aim of undoing the influence of luck shouldlead us to adopt the aim of establishing a distributive pattern inwhich everyones condition is relevantly the same. The aim of

1. Dworkin 1981, as cited in Cohen 1989.

2. Hurley, this volume.

SUSAN HURLEY AND RICHARD J. ARNESON74

eliminating the impact of luck cannot provide a basis or argu-ment for favouring any particular principle that specifies a justpattern of distribution.

Hurleys analysis also indicates that we should be cautiousabout the supposed moral imperative of undoing the effects ofunchosen luck. How we should regard luck as a determinant ofpeoples fate in a particular setting depends on the moral prin-ciples we accept, not the other way round.

All of this leaves it entirely open whether principles of distribu-tive justice should be sensitive to luck, and in particular to thedifference between chosen and unchosen good and bad fortune.3

Although it is easy to describe examples in which the judgmentthat some equally needy people deserve to be helped more thanothers may appear plausible, there are examples and examples,and it is far from clear that one can develop a principle ofresponsibility or deservingness that plausibly integrates such fac-tors into a theory of distributive justice.

It is important at the outset to distinguish an instrumental andan intrinsic significance that considerations of responsibilitymight be deemed to have. Social policies that attempt to identifythe deserving and undeserving poor and the deserving and unde-serving rich and to accord favourable treatment to the deservingmight be justified entirely by appeal to their expected good effectsthat are identified by principles that do not themselves containany notions of deservingness. Here norms of deservingness andresponsibility would be functioning in a purely instrumentalcapacity, as tools to bring about increased fulfilment of othervalues.

The crucial idea for the instrumental conception of responsi-bility is that of holding a person responsible for her action. Onan instrumental basis, I am responsible for an act if others wouldbe reasonable to hold me responsible for it, and holding a person

3. The core idea here is R. Dworkins distinction between option and brute luck inDworkin 1981. Dworkin stipulates that option luck is a matter of how deliberateand calculated gambles turn out-whether someone gains or loses through acceptingan isolated risk he or she should have anticipated and might have declined. Luckthat is not option luck is brute luck. So construed, the distinction is a matter ofdegree along at least three distinct dimensions. A given risk can be more or lessforeseeable, more or less avoidable by the agent if she chooses, and the risk can beavoidable by courses of action that are more or less advantageous in prospect for theagent. For this understanding of option luck I am indebted to Peter Vallentyne.

LUCK AND EQUALITY 75

responsible for an act is reasonable or unreasonable dependingon the consequences of this further act of holding. To hold some-one responsible for an act is to regard her as properly liable topraise or blame, reward or punishment, for its quality. The hold-ings, praisings, and blamings are justifiable just when they pro-duce better consequences than would occur in their absence.

Within a consequentialist framework, responsibility instru-mentally regarded is unproblematic. If using responsibility in theform of deployment of carrots and sticks induces improvedbehaviour over the long run, leading to better states of affairs,then responsibility notions are vindicated. Whether takingresponsibility to be intrinsically significant makes sense is, so far,still open.

I

The issue that Hurley artfully discusses has a recent history. In1971 John Rawls rejected the idea that whatever distribution ofgoods is brought about by the operation of a competitive marketeconomy is just. His objection is that in such a system distributiveoutcomes are improperly influenced by natural and social con-tingencies, factors that are arbitrary from a moral point ofview.4 The factors to which Rawls alludes are such matters asbeing born with a genetic disposition to be more or less talentedin various ways, inheriting wealth and social connections fromones parents, and being well or badly educated and nurtured inearly childhood. These matters are entirely beyond the individ-uals power to control. To the extent that they determine an indi-viduals life prospects, they do so as sheer brute luck, luck thatfalls on a person directly and independently of any choice shemight make.

In Anarchy, State, and Utopia, Robert Nozick accuses Rawlsof begging the question if he intends to argue from the arbitrar-iness of luck to the moral desirability of equality. Unless we havealready established a presumption in favour of equality, andNozick denies that we have, the fact that the luck of arbitrarycontingency plays a role in bringing about an unequal outcomedoes not per se undermine whatever moral legitimacy the out-come enjoys. Nozick also points out that under the operation of

4. Rawls 1999, p. 63.

SUSAN HURLEY AND RICHARD J. ARNESON76

Rawlss favoured egalitarian principle of justice, the differenceprinciple, who ends up with more and who with less will dependon contingencies no less arbitrary than those that were claimedto taint the distributive outcomes brought about by the operationof a market economy unregulated by the difference principle.5

One might say that luck comes in different flavours. Moralprinciples will have a taste for some and not others. Whether anoutcome that is in some sense brought about by luck is morallysuspect is determined by the moral principles we accept. Toidentify a luck outcome as morally arbitrary presupposes that wealready are committed to moral principles that fix what is andwhat is not morally arbitrary. According to Rawls, if brute luckfactors contribute to bringing it about that the difference prin-ciple, fair equality of opportunity, and the equal liberty principleare more perfectly implemented, the contribution of luck to theoutcome does not morally taint it. According to Nozick, if bruteluck factors influence the endowments and possessions that indi-viduals come to have, then, provided that this distribution doesnot come about ia violation of anyones Lockean rights, thecontribution of luck to the outcome does not morally taint it.

It is a nice question whether or not Rawls makes the error thatNozick diagnoses. Nozick himself distinguishes a positive and anegative argument from the arbitrariness of contingency thatRawls might be making. The positive argument would arguefrom the arbitrariness of natural contingency to the desirabilityof erasing outcomes due to such contingency and implementinginstead some egalitarian principle. The negative argument startsfrom an already established egalitarian baseline and rebuts anobjection. The objection is that people deserve by their effortsand achievements the superior position that the operation of theegalitarian principle would diminish. The negative argumentrejoinder is that what an individual can take credit for by wayof effort and achievement is intertwined with influences that areentirely beyond her power to shape or control and for which shecould in no way claim credit. There is no feasible way to untanglethis mixture of influences so as to sort out what people trulydeserve in a reliable way that could guide social policy. As Rawls

5. Nozick 1974, chap. 7, section II, esp. p. 219.

LUCK AND EQUALITY 77

states, The idea of rewarding desert is impracticable. 6 Whetheror not Rawls actually intends to assert what Nozick classifies asthe positive argument, he clearly intends the negative argument.So perhaps Rawls does not after all make the mistake that Noz-ick alleges and that Hurley carefully analyses.

The negative argument may mingle with the positive argumentat one point in Rawlss discussion. Arguing that the primary sub-ject of justice is the basic structure of society, the way that majorinstitutions combine to influence initial life prospects, Rawls dis-tinguishes deep and shallow inequalities among individuals.7

Shallow inequalities arise as adults interact, and conceivablymight be justified by norms of individual desert or responsibility.Deep inequalities are thrust on individuals at birth or early child-hood, so they cannot conceivably lie within the individualspower to control, and cannot possibly be justified by claimsabout individual responsibility and desert. Rawls also charac-terizes deep inequalities as those that are formed by the basicstructure of society as it shapes initial prospects. These twocharacterizations of deep inequalities do not coincide, but let thatpass. If Rawls supposes deep inequalities need justificationwhereas deep unchosen equalities would not, he may just beassuming equality as the moral baseline, or he might conceivablybe making the fallacious inference from is caused by luck toshould be altered to equality.

My hunch is that Rawls is not guilty of the charge that hemakes an inference from the fact that inequalities in peoplescondition are due to luck to the conclusion that the luck out-comes should be eliminated and equality of condition estab-lished. My further hunch is that, despite misleading rhetoric,egalitarians such as Ronald Dworkin, G. A. Cohen, and JohnRoemer are also innocent of this confusion. This assertion is notoffered by way of criticism of Hurley, whose interest lies in ana-lysing a pattern of argument, not ascribing it to anyone inparticular.

The pattern of argument that does recur in the literature iscloser to what Nozick calls the negative argument. Against thebackground of a presumption in favour of equality of condition

6. Rawls 1999, p. 274.

7. Rawls 1999, p. 7.

SUSAN HURLEY AND RICHARD J. ARNESON78

(or some other norm of justice that favours egalitarian distri-bution), the objection is made that many of the individuals whowould be required to relinquish resources they currently hold ifegalitarian justice were implemented deserve or merit their pre-sent advantages. The flip side of this same objection is that manyof the individuals who would be benefited if egalitarian justicewere implemented are negatively deserving or are at any rateresponsible by their choices for their present unfortunate state.Considerations of individual responsibility and deservingnessmilitate against acceptance of egalitarian justiceso goes theobjection. Against this objection the negative argument answersthat, if individual responsibility and deservingness are rightlyunderstood, they will be seen not to militate against egalitariandistribution as the objection supposes. Individuals to a largeextent become better off than others and secure advantages forthemselves in modern societies by processes that reflect the influ-ence of luck for which no individual can rightly claim credit. Theachievement of advantages that is claimed to render the achievermeritorious turns out under examination to be mainly the resultof luckcontingencies that are arbitrary from the moral pointof view.

Different versions of the negative argument vary in the extentof their negativity. At one extreme, one might claim that no oneever truly deserves anything. Nor is any one ever truly respon-sible for anything. No objection against egalitarian justice fromthe standpoint of deservingness or responsibility could thenstand. More concessionary responses to the objection are alsopossible. The advocate of equality of condition might qualify hercommitment to equality by holding that it is bad if some areworse off than others through no fault or choice of their own-but that it is not bad if some are worse off than others throughtheir own fault or choice. The prioritarian of well-being mighthold that moral value should be maximized and that the moralvalue of obtaining a gain (preventing a loss) for a person isgreater (1) the lower the persons well-being would be withoutthis benefit, and (2) the more deserving or virtuous the person is.The extreme version of the negative argument denies any role fordesert and responsibility at the level of fundamental principles ofjustice, and the moderate version includes desert or responsibilityin just this role.

LUCK AND EQUALITY 79

Notice that the negative argument can unfold along twodimensions. Against the claim that well-off people, being respon-sible for their advantages, or virtuous, should not be deprived ofthem for the sake of equality, one might either (1) deny that thecriteria of responsibility or virtue implicit in this claim are coher-ent and morally acceptable, or (2) deny that well-off peoplesatisfy these criteria to a greater extent than badly-off people.One might also do both, in varying degrees.8

Whether the negative argument is sound or unsound, it evi-dently does not make the mistake that Hurley is concerned todisplay.

Another reason to doubt that egalitarian theorists rely on thereasoning Hurley rightly rejects is that writers like Cohen andRoemer take Ronald Dworkins framework as the starting pointfor their reflections.9 But Dworkin explicitly does not aim to pro-vide any arguments for distributive equality besides providingthe best interpretation of the ideal, the one that is normativelymost attractive. Since he does not provide any arguments forequality, a fortiori he does not offer the one Hurley criticizes. Hedoes not infer from the claim that the inequalities we see arebrought about by sheer luck to the conclusion that they shouldbe eliminated. His concerns rather lie in the neighbourhood ofthe negative argument. He takes it to be obvious that people whomake choices and act on them from a baseline of equality shouldnot be restored to a position of literal equality of condition-doingso would in fact offend against the ideal of equality itself prop-erly conceived. He is then concerned to elaborate a moderateversion of the negative argument.

II

Rawls considers what he calls the system of natural liberty, inwhich there is a market economy that is laissez-faire except thatthe norm of careers open to talents is enforced. Of this systemhe comments that, since there is no attempt to preserve equalityof social conditions, the initial distribution of assets for any per-iod of time is strongly influenced by natural and social contin-gencies. The existing distribution of income and wealth, say, is

8. Thanks to G. A. Cohen for suggesting the content of this paragraph.

9. Cohen 1989; Roemer 1996; Dworkin 1981a and 1981b.

SUSAN HURLEY AND RICHARD J. ARNESON80

the cumulative effect of prior distributions of natural assetsthat is, natural talents and abilitiesas these have been devel-oped or left unrealized, and their use favoured or disfavouredover time by social circumstances and such chance contingenciesas accident and good fortune. Intuitively, the most obvious injus-tice of the system of natural liberty is that it permits distributiveshares to be improperly influenced by these factors so arbitraryfrom a moral point of view. 10

The last sentence might seem to suggest the fallacious inferencefrom is caused by luck to should be altered to equality thatHurley exposes. But I see here another formulation of the nega-tive argument. Against the backdrop of a presumption in favourof initial equality, we see the cumulative results of peoples mar-ket choices tending to deviate from equality. This might be mor-ally acceptable if the process that creates inequality has acharacter that confers legitimacy on the outcome. But accordingto Rawls that is not so in this case. Instead, chance contingencies,unchosen luck factors, strongly influence the distributive out-comes as they unfold over time.

The observation that outcomes are determined or partly deter-mined by luck that simply falls on people in ways that are notmediated by their choices can hardly be the determining factorin assessing the outcomes from the standpoint of justice. Afterall, it could happen by sheer chance that the operation of such amarket economy over time yields outcomes that satisfy whateveregalitarian or other principles of distributive justice are deemedmorally required. Hence, when the operation of a market econ-omy yields outcomes that fail to satisfy these justice norms, whatis wrong is not that chance factors are bringing about the out-come. What is wrong is that justice norms are not satisfied.

The question that is not touched by any of the above and isstill entirely open is whether or not the distinction betweenchosen and unchosen luck can carry a heavy load of justificationof distributive outcomes. Also, what principles best articulate themoral considerations that should be moving us when we appealto this distinction?

The lightly regulated market economy as described by Rawlsproduces inequalities in peoples initial life prospects. We might

10. Rawls 1999, p. 63.

LUCK AND EQUALITY 81

imagine these as life prospects calculated for individuals as new-borns, before they could possibly make responsible or irrespon-sible choices or act in deserving or undeserving ways. People facethese unequal initial life prospects on the basis of unearned con-tingencies such as possession of inherited talent, expected favour-able upbringing, family social connections, and so on. The issuehere is not merely that a norm of equality is violated, but thatthe luck of the market operating over time might well frustratea broad range of principles specifying what might be thought tobe a morally desirable pattern of distribution.

At any rate, that people might face unequal initial life pros-pects does not seem to capture the core of what is problematicabout the operation of luck within the system of natural liberty.For we can imagine that this feature of the scenario is eliminated,yet moral criticisms still apply. To rehearse a suggestion of BrianBarry, suppose we discover that nurses have been switching theidentity tags on newborn babies in a complicated lottery schemesuch that at birth all individuals have equal life prospects.11

Otherwise the system of natural liberty operates as previouslydescribed. Shortly after birth, after the nurses lottery is completeand one has been placed with a set of parents according to theoutcome of the lottery, individuals face vastly unequal life pros-pects through no fault or choice of their own. If the originalscheme was morally objectionable, the system of natural libertyamended by the nurses lottery still seems objectionablesomehow.

The amended system of natural liberty does bring about equal-ity of initial life prospects across individuals. (How to measurelife prospects is for the moment left open; we can suppose thatwhatever conception of life prospects favoured by the theory ofjustice is employed.) This achievement is morally non-trivial.Under this system each individual has a real chance of enjoyinga good quality of life, unlike what occurs in the unamended sys-tem of natural liberty. Nonetheless the nurses lottery hardlyinstitutes an egalitarian utopia.

Taking equality of condition to be the baseline state of affairsfavoured by justice, we might hold that it is bad if some are worseoff than others through no fault or choice of their own, and find

11. Barry 1988.

SUSAN HURLEY AND RICHARD J. ARNESON82

fault with the amended system of natural liberty for failure tosatisfy this norm. This objection can be stated in more generalterms. The thought that morality should make it a priority tooffset uncourted disadvantage need not be yoked to any affir-mation that equality of condition is per se desirable. The justicestandard might insist that we should equalize, or equalize at thehighest feasible level, or provide everyone at least a level deemedsufficient, or maximize the total sum, or maximize the average,or prioritize, and so on. Any such doctrine can be qualified so asto be responsibility-catering.

In what follows I shall suppose that life prospects should beassessed in terms of well-being (or welfare or utility). In assessingan individuals condition for purposes of distributive justice,what matters is the ensemble of circumstances she faces as theyaffect some function of her well-being over her life course. Thenotion of well-being in play here is left unspecified, though myhunch is that an Objective List conception of well-being yieldsthe most plausible version of welfarist distributive justice. Thechoice of a welfarist framework is controversial, but cannot bedefended here.12 I further assume that well-being is cardinallyinterpersonally comparable.

Consider then the proposal that distributive justice requiresthat each individual should have a fair opportunity for welfare.Fairness, by whatever standard it is specified, regulates oppor-tunity for welfare. The rough idea is that if I have an opportunityto get X amount of welfare, then there is a prudent course ofaction available to me that will yield me X if I take it, and, if Ido not, the responsibility for my lower welfare level lies with me.But the rough idea should be qualified, because identifying a fairopportunity with what one could get by perfect prudence is toodemanding, and itself unfair.13

A persons opportunity for welfare is the level of welfare thatshe would gain if she conducted herself throughout her life asprudently as could reasonably be expected. Reasonably expect-able prudence is distinguished from perfect prudence. A personwho conducts herself in a perfectly prudent manner always takesthe most efficient means to advance her own welfare. Doing this

12. Arneson 2000b.

13. Arneson 1997.

LUCK AND EQUALITY 83

might be either painful or difficult for a person to do. Makingthe prudent choice might require solving a mathematical equa-tion, which one person could do easily, another, if at all, onlywith great difficulty. Carrying out the prudent choice mightrequire skill, which some can muster easily, others, if at all, onlywith great difficulty. Making and executing the prudent choicemight be pleasant or at least painless for some, but painful forothers (at the limit, so painful that being prudent would be physi-cally impossible for the person). On the proposal under review,it is deemed that merely bringing it about that if one behavedperfectly prudently throughout ones life one would gain a levelof welfare that is fair would not sufficiently cater to the individ-uals moral interest in having access to welfare. The standard ofindividual conduct is relaxed by identifying a degree of prudencethat is as much as one can reasonably expect an individual toexhibit in each setting of action. Distributive justice then requiresthat society be arranged so that each individual can get a levelof welfare deemed fair provided that she behaves prudently to areasonably expectable degree throughout her life.

The ordinary connotations of the word opportunity do notapply in this specification of opportunity for welfare. One ordi-narily would say that when a person has an opportunity for somegood, she has both the option to choose and get this good andalso the option to decline it. An opportunity opens a doorthrough which the individual may or may not choose to walk.Having opportunity implies having freedom, open options.

But, on the suggested construal of opportunity for welfare, achange that lessens an individuals freedom may give him moreopportunity for welfare. With more money and more freedom,one might chooseimprudently but not below the threshold ofreasonably expectable prudenceto take heroin, and ones wel-fare would diminish, compared to the state of affairs one wouldreach if one behaved throughout ones life in a reasonably pru-dent way, faced with less money and freedom. In the same vein,one might have greater opportunity for welfare in given circum-stances if a single change is made, lessening ones freedom bylegally prohibiting heroin consumption, than if the change werenot made.

Opportunity for welfare in a fair opportunity norm might bedefined in at least two different ways: either the opportunity is

SUSAN HURLEY AND RICHARD J. ARNESON84

that if one behaves reasonably prudently over the course of oneslife one will get (1) a certain welfare level or (2) a certain expectedlevel of welfare. In either of the versions mentioned, fair initialopportunity for welfare is problematic as a (partial) conceptionof distributive justice for at least three reasons.

First, it is too unforgiving. The ideal of fair opportunity is thatjustice requires that a path be provided to each individual suchthat, if the individual stays on the path throughout her life, theoutcome she reaches (or the expected outcome) would be fair. Ihave softened the norm of responsibility informing this ideal by,in effect, widening the path, so that it is not too painful or diffi-cult for the individual to avoid straying from it. But what hap-pens to the individual if she strays off the path even by a slightamount is a dont care from the standpoint of this conceptionof justice. This is an ineliminable and implausible feature of theideal. A young adult may behave in an irresponsibly careless waythat fails to meet the soft standard of responsibility, but just by asmidgeon, then encounter incredibly bad luck, and end up facinghorribly grim life prospects that we could alleviate by furtherprovision of resources to her at modest cost. The fair equality ofopportunity of welfare account responds to such a case byinsisting that justice demands no transfers of resources to allevi-ate the errant individuals plight because any such transfer woulddiminish the fair equality of opportunity for welfare to whichothers are entitled.

A second problematic feature of the account is that by designit does not discriminate between virtuous and non-virtuousimprudence. Virtuous imprudence is failure to pursue ones ownself-interest to the maximum (within the constraints of requiredduties) in order reasonably to benefit other persons or some wor-thy cause. Think of Mother Theresa devoting herself selflessly tothe poor of Calcutta. Non-virtuous imprudence is failure to pur-sue ones own self-interest to the maximum either from careless-ness in identifying ones interest, or fecklessness in its pursuit, orunreasonable pursuit of small gains for other persons or for agood cause at excessive cost to oneself. The idea behind fairequality for welfare is that if society (all of us regarded collec-tively) provides the individual with a path she could take thatwould yield a fair level of welfare (or expected welfare) if shetook it, then the individual, and not society, is responsible for

LUCK AND EQUALITY 85

any welfare deficit for her that arises from proceeding alonganother path. It may be perfectly reasonable and even admirableto sacrifice oneself for some good cause in a supererogatory man-ner, but doing so does not generate duties of other persons toboost ones welfare back toward what it would have been hadone not veered from the prudent path. So holds fair equality ofopportunity for welfare. This is not on its face a crazy view, butit is controversial, very far from self-evidently correct.

A third problematic feature of the account is that it embodieswhat may seem on reflection a weird causal requirement. Com-pare a person, Smith, who behaves wildly imprudently on manyoccasions, but is very lucky, and suffers no misfortune. (He playsRussian roulette on himself with a loaded gun repeatedly, justfor fun.) He then suffers a brute luck misfortunean utterlyunpredictable earthquake destroys his house. Under a regime inwhich fair opportunity for welfare holds an individual respon-sible for outcomes that befall him due to option luck but notbrute luck, he is eligible for compensation for his brute luck mis-fortune. In contrast, consider his neighbour, Jones, who behavesalmost all of the time with impeccable prudence but on oneoccasion behaves very slightly imprudently and then has terriblebad luck, so the slight imprudence gives rise to personal disaster.Jones suffers misfortune through her fault or choice; so she isresponsible for the consequences that fall on herself. Smithmakes many more faulty choices, but the actual misfortune hesuffers does not come about through his fault or choice; so he isnot responsible for the bad consequences that fall on himself.

In principle, one could devise a responsibility account thatavoids these difficulties, but at a price. In what follows, variouspossible revisions of responsibility are described that graduallyreduce the difficulties, incurring various moral costs.

One possible move would be to give up the idea that distribu-tive justice requires compensation for misfortune that individualssuffer only if it is the result of brute and not option luck. Insteadthe soft account of personal responsibility is applied over theindividuals entire life, the dimensions of responsibility areassigned weights, and an overall score measures the degree towhich an individual is responsible for her current condition. Thelower the degree of responsibility that an individual bears for hercurrent condition, the better it is to secure improvements for her

SUSAN HURLEY AND RICHARD J. ARNESON86

if it is bad, and the less bad it is to worsen her condition if it isgood in order to improve the condition of others. This proposalaccommodates the too unforgiing concern but not weird causalrequirement or failure to distinguish irtuous and non-irtuousimprudence.

Another possible responsibility norm holds that, other thingsbeing equal, two individuals who conform over their entire livesto the same degree to the standard of behaving reasonably pru-dently should enjoy the same level of well-being. This proposalaccommodates two of the three mentioned difficulties. It failsonly with respect to distinguishing irtuous and non-irtuousimprudence.

A third possible norm holds that, other things being equal, itis morally more valuable to gain an increment of well-being fora person, the greater her lifetime virtue. A persons virtue isgreater, the greater the extent to which her will is disposedtoward morally good conduct, at each moment of her life.

These proposals move in the direction of abandoning any ver-sion of fair opportunity for welfare in favour of a norm thatgood fortune, other things being equal, should be made pro-portional to the degree of responsibility one exhibits over thecourse of ones life. There is also the further issue whetherresponsibility or virtue should be the basis for differential treat-ment of persons on some egalitarian conception of justice. (Thatis, should virtuous and nonvirtuous imprudence be distinguished,as the virtue accounts affirm and the responsibility accountsdeny?) The moral considerations favouring each of these dis-tributive norms are quite different. The idea of fair opportunityis that the world should be arranged so that each person wouldgain a fair level of well-being if she chooses to be prudent andbehaves as prudently as it would be reasonable to expect givenher choicemaking talents and unchosen socialization experiencesthat fix the difficulty and unpleasantness of making and execut-ing prudent choices. A qualification to fair opportunity that iswithin the spirit of the proposal is that if a person deliberatelyand voluntarily chooses to gamble she should live with the resultseven if they are bad for her.

The idea of the proportionality norm is that some type of con-duct is picked out as meritorious, and justice requires that, otherthings being equal, each persons lifetime level of good fortune

LUCK AND EQUALITY 87

or well-being should be made proportional to her level of meritexhibited in her conduct over the course of her life. Those whoare deserving should be rewarded according to their deserv-ingness. A fair equality norm holds that good fortune and badfortune that accrues to individuals as a result of morally permiss-ible choices and hence as a result of option luck properly belongsto the responsible individuals. A proportionality norm smoothesout these outcomes due to option luck in order to render goodfortune proportional to responsibility or moral merit. The fairopportunity norm and the proportionality norm evidentlyembody different moral attitudes toward good and bad luckenjoyed by equally responsible individuals.

III

This critical discussion of how to integrate personal responsi-bility into distributive justice may appear to have driven into adead end.14 The proposal that, other things being equal, the moremorally deserving one is over the course of ones life, the betterones life should go, could be regarded as a reductio ad absurdumof the project of making justice responsive to responsibility. Itwould be a silly idea to try to run a monastic community of men,all of whom have dedicated their lives to a common religiousideal, according to the standard that good fortune should bemade proportionate to ones true moral virtue. If the abbot of amonastery could not base a sensible governance policy on thisnorm, then surely it is irrelevant to the governance of a largediverse democracy that aims to be just.

Two distinct forces combine to put pressure on the idea thatdistributive justice should be responsibility-catering. One is thatpeople differ in contingencies, beyond their power to control,that shape their ability to choose and act in a responsible anddeserving manner. Since these contingencies are subtle and vari-ous, and hard to observe, the task of separating what one canand cannot reasonably be held responsible for becomes more dif-ficult. Furthermore, the project of designing institutions and

14. For further grounds for doubting that the distinction between chosen andunchosen luck should play a significant role in fundamental social justice principles,see Anderson 1999. For response, see Arneson 2000a; also Arneson 1999.

SUSAN HURLEY AND RICHARD J. ARNESON88

practices that would carry out this task in the context of a mod-ern society looks increasingly utopian. A second source of press-ure is that one can be led to find it appropriate to judge theunit of deservingness or responsibility to be the conduct of anindividual over the course of her life. If a persons conduct isassessed on the basis of a single act or the quality of her acts oversome limited fraction of her life, the judgment that is reached canbe anomalous with respect to the comparable judgment abouther life as a whole. But the idea of a social agency that monitorsthe lifelong deservingness and responsibility scores of all of itsindividual members is a chimera.

Nevertheless, the proposal that considerations along the linesjust mentioned provide good reasons to exclude considerationsof responsibility and deservingness from playing a role in funda-mental principles of morality and justice strikes me as mistaken.We should keep separate practical concerns about the prospectsfor implementing this or that aspect of justice, and theoreticalconcerns about what justice, as best conceived, actually requires.One reason for doing so is simply that these questions just aredifferent. At some times and places, given human intransigenceand the power of entrenched law and custom, slavery might be,for now, ineliminable. But the fact that there is no way right nowto get rid of it does not imply that it is right now just. Even anaspect of current social systems that we suspect can never beeliminated might still be unjust, provided that we can coherentlydescribe a logically and physically possible world in which thismoral evil is eliminated.

A second reason to place in separate mental compartments theissue what justice in principle requires, and the issue how best ina given setting to fulfil this best conception to the greatest extentpossible, is that if we run the questions together we may tooquickly foreclose possibilities of implementation. Even if we can-not directly enact a policy that is explicitly aimed at bringingabout a certain aspect of justice, we might be able to devise prox-ies for approximating, to a greater or lesser extent, to the realiz-ation of the conception we cannot directly implement. In thisrespect the notion of responsibility and the notion of human well-being are alike. Once we engage in reflection about the nature ofwell-being, we find that the conceptions of it that make mostsense will render an individuals well-being level difficult to

LUCK AND EQUALITY 89

observe and standards of well-being controversial to apply. Butit would be a mistake to decide that well-being is morally unim-portant just on the ground that it is difficult to measure andto apply in public policy. The same goes for responsibility anddeservingness.

Another consideration to bear in mind in this connection isthat, if responsibility and deservingness as they are most plausi-bly conceived in the best theory of justice should turn out to bevery difficult to implement, and indirect strategies of implemen-tation by proxy are unfeasible, then perhaps responsibility anddeservingness, though morally significant in principle, might beirrelevant in practice. What we cannot implement reliably at allwe may have to ignore. To achieve this recognition may requirethat we have investigated the theory of responsibility and deserv-ingness and seized on their theoretically best interpretations.That responsibility wrongly conceived may be readilyimplementable may then be of little interest.

But it is not credible that responsibility and deservingness mat-ter only instrumentally and are not intrinsically morally signifi-cant. Imagine that the social science research of the next centurysurprisingly demonstrates that we can identify the truly deservingand responsible and that social justice norms other than responsi-bility-catering itself can be more efficiently achieved by arranginginstitutions and practices so that vice is rewarded and virtue pun-ished and, other things being equal, sinners have far better lifeprospects than saints. Imagining this possible world, I find myselfpreferring that other justice values be sacrificed to some extentin order to bring it about that those whom a fine-grained accountof responsibility deems more virtuous or deserving should otherthings being equal enjoy better life prospects than the vicious andundeserving. Of course this leaves it open that a fine-grained the-ory of responsibility would have it that all of us are equallydeserving and undeserving and that the appropriate response tothe irresponsible or vicious agent other than oneself is alwaysThere but for the grace of God go I.

REFERENCESAnderson, E., 1999, What Is the Point of Equality?, Ethics, 109.Arneson, R., 1997, Equality and Equal Opportunity for Welfare, and Post-

script 1995, in L. Pojman and R. Westmoreland, eds., Equality: SelectedReadings (Oxford: Oxford University Press).

SUSAN HURLEY AND RICHARD J. ARNESON90

Arneson, R., 1999, Egalitarianism and Responsibility, Journal of Ethics, 3.Arneson, 2000a, Luck Egalitarianism and Prioritarianism, Ethics, 110.Arneson, R., 2000b, Welfare Should Be the Currency of Justice, Canadian Jour-

nal of Philosophy, 30.Barry, B., 1988, Equal Opportunity and Moral Arbitrariness, in N. Bowie, ed.,

Equal Opportunity (Boulder: Westview Press).Cohen, G. A., 1989, On the Currency of Egalitarian Justice, Ethics, 99.Dworkin, R., 1981a, What Is Equality? Part 1: Equality of Welfare, Philosophy

and Public Affairs, 10.Dworkin, R., 1981b, What Is Equality?: Part 2: Equality of Resources, Philo-

sophy and Public Affairs, 10.Hurley, S., 2001, Luck and Equality, this volume.Nozick, R., 1974, Anarchy, State, and Utopia (New York: Basic Books).Rawls, J., 1999, A Theory of Justice (revised edition) (Cambridge: Harvard

University Press).Roemer, J., 1996, Theories of Distributie Justice (Cambridge: Harvard Univer-

sity Press).