University of Southern California Dissertations and Theses

Distributive equality and labor burdens

(USC Thesis Other)

Distributive equality and labor burdens

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

DISTRIBUTIVE EQUALITY AND LABOR BURDENS

by

Tiffany Chang

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PHILOSOPHY)

August 2014

Table of Contents

Acknowledgments

Introduction 1

1 Sufficientarianism 4

2 Responsibility-Sensitive Egalitarianism and Luck 34

3 Reciprocity-Based Egalitarianism and Incentives 63

4 Burdens-Sensitive Egalitarianism 97

5 The Burdens of Care and the Reach of Justice 120

Bibliography 155

Acknowledgments

I am fortunate to have known many great philosophers whose encouragement and
tutelage have enabled me to write this dissertation. First and foremost, I am grateful for my
chair, Sharon Lloyd. I first began to develop a serious interest in distributive justice in her
graduate seminar, and her guidance throughout the years provided me invaluable support as I
developed the core ideas for this dissertation. Her encouraging feedback gave me both the
much-needed confidence to continue in my studies and the critical insights that allowed me to
improve. I am also very thankful to Gary Watson for his generosity with his time and his
patience in helping me untangle my arguments over the course of long discussions. My thanks
also go to John Dreher, with whom discussion always prompted me to think in new ways about
the commitments of the various distributive theories I was engaging with.
I would like to thank Gregory Keating for so graciously showing willingness to
participate as an outside member of my dissertation committee on short notice. In addition, my
thanks go to Alison Renteln, whose comments and insights as an outside member of my
qualifying committee were so helpful and constructive.
This dissertation is dedicated to my parents, Richard and Maria Chang. With their many
acts of compassion and kindness to others, they were the first instill in me an interest in social
justice. None of my accomplishments would have been possible without the unwavering love
and support they’ve provided me.
My thanks also go to my friends and peers. One of my fondest memories of my time at
USC will be of the wonderful discussions I had about philosophy with Yingying Tang, Xiao
Wei, Hsin Wen Lee, and Brandon Johns. Lastly, I’d like to thank my closest and dearest friend,
Andrey Gordienko, for his gentle encouragement and advice.

1

Introduction

The background literature informing this dissertation comes largely from the aftermath of
Rawls’s A Theory of Justice (1971). In particular, I am concerned with the debate surrounding
the question “To what extent should distributive justice redress inequalities that arise from
differences in initial life prospects?” The Rawlsian answer states, roughly, that such inequalities
are permissible only under terms of reciprocity in which they redound to the greatest benefit of
the least advantaged. I call this reciprocity-based egalitarianism. Responsibility-sensitive
egalitarians, on the other hand, allow divergences from equality only when they reflect
individual’s responsible choices and not their circumstances. And Sufficientarians hold that, so
long as everyone meets a certain threshold, inequality as such is not of moral concern.
These three positions have dominated the debate about distributive justice for the past 40
years. In this dissertation, I will offer an alternative to responsibility-sensitive egalitarianism that
better captures the intuitions that originally motivated responsibility-sensitive principles. In
addition, I accept the positive sufficientarian thesis that no one should fall below a certain
threshold of decency, but I reject their negative thesis that above that threshold, inequality as
such is no longer a concern of justice. The view that I propose draws upon my conclusions about
responsibility-sensitive egalitarianism and sufficientarianism. The result is an extension of
reciprocity-based egalitarianism that allows inequalities only when they are necessary to
compensate for the burdens of socially necessary labor. I call this view burdens-sensitive
egalitarianism.
2

In Chapter 1, I present objections to three of the most influential sufficientarian views. In
so doing, I hope to vindicate the claim that even when everyone has enough, economic equality
is of moral importance. The remaining chapters are devoted to a discussion of liberal egalitarian
views.
Chapter 2 turns to Ronald Dworkin’s Equality of Resources theory, which models
distributive equality according to a hypothetical auction in which individuals are given an equal
opportunity to purchase resources. I argue that Dworkin’s proposed test for evaluating
distributions fails to cohere with the underlying motivation of his theory to neutralize the
disadvantages that would otherwise result from bad brute luck. This work can be better
achieved, I argue, by responding to the burdens of socially necessary labor rather than to facts
about individuals’ responsibility.
Chapter 3 contains a discussion of my alternate proposal: the Burdens Compensation
Principle. I discuss some of the core categories of labor burdens and demonstrate how the
Burdens Compensation Principle is equipped to address issues surrounding disability.
In Chapter 4, I offer a defense of Rawls’s Difference Principle against G.A. Cohen’s
internal critique of it. However, I conclude, Cohen’s objections nonetheless offer suggestive
external critiques of Rawls’s theory. In particular, Rawls’s Difference Principle fails to
accommodate the intuitively compelling idea that the burdens of socially necessary labor are
relevant, from a moral point of view, to the distribution of income and wealth. If we are to take
seriously the ideal of reciprocity, we must be prepared to compensate people in accordance with
the burdens that they undertake in contributing to the social good.
Lastly, in Chapter 5, I continue my discussion of Cohen’s criticism of Rawls by offering
a response to his charge that Rawls’s account of the basic structure is arbitrary. We can look to
3

the circumstances of justice to understand what the normative restraints on coercive power are.
These restraints establish the boundaries of the basic structure. Feminists may object to this
account of the basic structure on the ground that it allows oppressive gender norms to interfere
with women’s freedoms and opportunities as citizens. The Burdens Compensation principle
offers a solution that both the political liberal and the feminist will find appealing: because care-
givers engage in socially necessary labor, they are not to be regarded as disengaged from social
cooperation, and as such must be compensated for the burdens they undertake. By compensating
for household labor and care of dependent family members, the Burdens Compensation Principle
is better equipped to prevent the social and economic disenfranchisement of caregivers.
Moreover, the Burdens Compensation Principle is not committed to any illiberal claims that the
family be internally ordered according to substantive claims about the moral equality of men and
women. It justifies compensating privately provided care on the ground that individuals who
undertake socially necessary labor are engaged in social cooperation and as such, their labor is
governed by terms of reciprocity. These are grounds that all citizens can accept.
In sum, the Burdens Compensation Principle is compatible with political liberalism,
better captures the intuitions that motivates the responsibility-sensitive view without inheriting
the need for a substantive account of moral responsibility, and is equipped to address both
disability theorists and feminist concerns as primary concerns of distributive justice.

4

1 Sufficientarianism

The case for egalitarianism acquires a tone of moral urgency when we consider the abject
conditions of the worst-off under systems of vast and systematic inequalities. Societies that fail
to adequately uphold the ideal of equality often lead to outcomes in which individuals sink to
indecent levels. The solution, egalitarians argue, is to raise the levels of the worst-off by
implementing principles that seek to close the gap between the rich and poor. For
sufficientarians, however, our concern for the worst-off need not commit us to egalitarianism.
Being worse-off than others is not always the same as being badly off. Our sense of injustice
responds not to the relative position of the poor but rather to the unacceptable conditions that
they suffer. What is important is that we raise the levels of those who suffer under unacceptable
conditions to levels that are decent or “good-enough,” not that we close the gap between the
worse-off and the better-off. The primary aim of distributive justice is to ensure that everyone is
raised to some threshold level; once everyone is guaranteed a “good-enough” life, economic
inequalities are no longer a concern of justice. Sufficientarians argue that their view addresses
our concern for the worst-off without being vulnerable to the standard objections against
egalitarianism, such as the leveling-down problem.
1
Moreover, they show that our intuitions
regarding particular distributive cases align with their view. When confronted with inequalities
that obtain above the minimal threshold, we do not intuit any instances of injustice; thus, our
concern for the badly off need not commit us to egalitarianism but rather leads us to support
sufficientarianism instead.

1
The leveling down objection challenges egalitarian principles on the grounds that they would be committed to
lowering the levels of the better-off when nothing can be done to raise the levels of the worst-off. Robert Nozick,
for example, poses such a worry for Rawls’s discussion of the distribution of natural talents; see Robert Nozick,
5

In this chapter I discuss the theories of three of the most prominent sufficientarians:
Harry Frankfurt, Roger Crisp, and Elizabeth Anderson. I offer objections to Frankfurt and
Crisp’s arguments that inequalities above the threshold are no longer of moral concern, thereby
leaving open the possibility that economic inequality has intrinsic value. Next, I turn to
Anderson’s version of sufficientarianism, which states that all individuals should have the
capabilities necessary for democratic citizenship. While I agree with her claim that a society
should secure that minimum for its members, I reject the further claim that justice does not
require limits on economic inequalities above that threshold. By offering objections to each of
these three versions of sufficientarianism, I hope to provide motivation for the claim that a just
society is one in which economic inequalities, even when above some minimal threshold for a
good-enough life, are regulated by principles of justice. This will pave the way for discussion of
other theories of distributive justice, such as Ronald Dworkin’s theory of resource equality
(Chapter 2), John Rawls’s theory of justice as fairness (Chapter 5), and G.A. Cohen’s socialist
critique of Rawls’s difference principle (Chapter 3).
I SPECIFYING A THRESHOLD
Sufficientarian theories are challenged with two main tasks: to specify the goods that
everyone should have enough of and to provide non-arbitrary grounds for setting a threshold
level for those goods. The first task answers to the question “Enough of what?” and the second
to the question “How much is enough?” I will begin by turning to the ways in which Frankfurt
responds to these two challenges.
Harry Frankfurt’s influential 1987 article “Equality as a Moral Ideal,” contains one of the
first arguments for sufficientarianism and continues to exert much influence on the course of
6

sufficientarian thought today.
2
Frankfurt argued for the following claims, together comprising
the Doctrine of Sufficiency:
a) With respect to the distribution of economic assets, what is important from the point
of view of morality is not that everyone should have the same but that each should
have enough.
b) If everyone had enough, it would be of no moral consequence whether some had
more than others.
3

The Doctrine of Sufficiency claims that it is morally important that each have enough, and that
once everyone has enough, “economic equality is not, as such, of particular moral importance.”
4

Specifically, once everyone has enough income and wealth to be “content with what he has,” we
should not care that some have more than others.
5
An individual is content with what he has
when he does not have an active interest in getting more and regards “having more money as
inessential to his being satisfied with his life.”
6

We have here two preliminary answers to the two questions earlier stated. First,
everyone should have enough economic goods with which to live a good enough life. Second,
one has enough economic goods when one is content with what one has. Before continuing, we
should note a few things about the second response. Frankfurt carefully points out that “having
enough” does not equate to merely having just enough to get by.
7
Sufficientarianism responds to
the idea that it is morally important to ensure that everyone has enough to avoid economic
misery; those who have only enough to scrape by do not have enough by that standard.
8
In
addition, although Frankfurt argues for sufficiency with respect to economic goods, his theory

2
Robert Huseby, for example, defends a theory that continues along the same lines, though with added
qualifications, in Robert Huseby, "Sufficiency: Restated and Defended," Journal of Political Philosophy 18, no. 2
(2010).
3
Harry Frankfurt, "Equality as a Moral Ideal," Ethics 98, no. 1 (1987): 21.
4
Ibid.
5
Ibid., 39.
6
Ibid.
7
Ibid., 38.
8
Ibid.
7

does not preclude the possibility that a good enough life requires more than just an adequate
amount of economic goods. The possession of social and political goods may also be important
for attaining a sufficiently good life.
Although Frankfurt denies the intrinsic value of economic equality, he nonetheless
recognizes its instrumental value. Egalitarians often argue for economic equality on the grounds
that disparities in income and wealth are often accompanied by inequalities in other dimensions
that are of moral concern to us. Large economic inequalities can promote unjust disparities
between the political power and influence of the rich and poor, for example, or the occupational
and educational opportunities open to the two groups. But here the disvalue of economic
inequality is derivative on its impact on social and political inequalities, and it is the latter kind
of inequality that is of moral concern to us. While economic equality may be instrumentally
valuable insofar as it contributes to social and political equality, it is not clearly an ideal that has
intrinsic value.
Frankfurt provides several arguments to show that economic equality is not intrinsically
valuable. First, he argues that under conditions of scarcity, equality may not only fail to
maximize aggregate utility but instead minimize aggregate utility and lead to outcomes from
which we would recoil; he writes:
Thus, suppose that there is enough of a certain resource (e.g., food or medicine) to enable
some but not all members of a population to survive. Let us say that the size of a
population is ten, that a person needs at least five units of the resource in question to live,
and that forty units are available. If any members of this population are to survive, some
must have more than others. An equal distribution, which gives each person four units,
leads to the worst possible outcome, namely, everyone dies. Surely in this case it would
be morally grotesque to insist upon equality!
9

In cases such as these—that is, when the attainment of a certain threshold is what divides a good
outcome from a bad one—benefits to some may not in fact have any value at all if they are not

9
Ibid., 30.
8

enough to raise them to the threshold level. An equal distribution of the resource would fail to
bring anyone up to the threshold level and hence would waste the potential value the resource
could have, were it to be distributed along sufficientarian lines. Frankfurt takes this case to show
that under conditions of scarcity when not everyone can have enough, egalitarian distributions
can be morally unacceptable. However, he does not deny that there can be cases in which giving
some resources to someone below the threshold is beneficial, even if it fails to raise her to or
above that threshold.
10
His point, rather, is that there is not always a moral case for giving a
benefit to someone below the threshold when it is not enough to raise her to that threshold.
11

Rather, what should be done in these situations will vary on a case-by-case basis. This argument
attempts to attack not only the egalitarian position but the prioritarian one as well. Although
prioritarian theories tend to also deny the intrinsic value of equality, they nonetheless hold that
priority should be given to the worse off because the moral value of a benefit is greater, the less
well-off the recipient is.
12
Both egalitarianism and prioritarianism hold that there is always a
default case for giving to the worse or worst off.
However, Frankfurt’s argument that an egalitarian distribution can be morally
unacceptable under conditions of scarcity does not show that economic equality is of no intrinsic
value. It merely shows that when equality is just one among many competing values, it may be
assigned less priority. In this case, the value of saving the most possible lives, for example,

10
Ibid., 32.
11
For example, he writes, “And while a person with one unit of food may live a bit longer than someone with no
food whatsoever, perhaps it is worse to prolong the process of starvation for a short time than to terminate quickly
the agony of starving to death.” Ibid., 31.
12
For discussion of Prioritarian principles, see Richard Arneson, "Egalitarianism and Responsibility," The Journal
of Ethics 3, no. 3 (1999); Dale Dorsey, "Equality-Tempered Prioritarianism," Politics, Philosophy & Economics 13,
no. 1 (2014); Nils Holtug, "Prioritarianism," in Egalitarianism: New Essays on the Nature and Value of Equality ed.
Nils Holtug and Kaspar Lippert-Rasmussen(New York: Oxford University Press, 2006); Thomas Nagel, Equality
and Partiality (New York: Oxford University Press, 1991); Derek Parfit, "Equality and Priority," Ratio 10, no. 3
(1997); Thomas Porter, "In Defence of the Priority View," Utilitas 24, no. 3 (2012).
9

clearly outweighs the value of equality. At most, this example impugns the implausible view
that equality always outweighs other moral values.
To see this, consider a relevantly similar, though simplified, case. Suppose that we have
only one unit of medicine M but two ill patients. Patient A needs only one unit of M to be
restored to health, while patient B is much more ill than A and needs 4 units of M to be restored
to health. The simple egalitarian might recommend splitting the one unit of medicine in half and
distributing each half to A and B; the prioritarian would recommend giving the medicine to B, as
B is worse off than A; and the sufficientarian would prescribe giving the medicine to A rather
than B, for that is the distribution that brings the greatest number of people to a minimal
threshold of health. If the available resources are such that we must choose between saving a life
or saving no lives, then we must choose the distribution that saves a life. This seems to suggest
that, in this case, while there is moral value in distributing a resource with the aim of raising
individuals above a threshold level, there is no moral value in distributing a resource with the
aim of equalizing individuals or raising the worse-off.
However, consider that administering a unit of medicine is an instance of providing
medical attention. Even if a particular amount of medical attention is not enough to bring the
patient to health, it is not clear that there is no moral reason to provide that attention. In fact,
intuitively, quite the opposite seems to hold—so long as a patient is in need of medical attention,
there is a pro tanto moral reason for providing care.
13
In the case Frankfurt describes, that pro
tanto reason is outweighed by other reasons that appeal to the fact that an unequal distribution of
medicine would save more lives than an equal distribution would. It would indeed be morally

13
Frankfurt, too, would have to agree that so long as an individual needs medical care, there is a pro tanto moral
reason for providing it. He uses considerations such as these, in fact, to argue that what matters from a moral point
of view is the satisfaction of needs, not the relative position of individuals. Egalitarians take their concern for the
poor to be a concern in favor of equality; this is a mistake, Frankfurt says, because what is really morally important
is the satisfaction of their needs, not their relative position.
10

repugnant to insist upon equality when it produces the worst outcome, but this does not yet show
that equality has no intrinsic value or that there is no moral reason in favor of giving a resource
to the worse-off individual. His case only shows that the moral value of giving to the less well-
off can be outweighed by other competing values, not that there is no intrinsic value whatsoever
in benefiting the less well-off.
14

Frankfurt also regards the view that economic equality is not intrinsically valuable to be
supported by the suggestion that we can be satisfied or content with the economic goods we
have. What we are or can reasonably be content with sets the threshold level:
To say that a person has enough means that he is content, or that it is reasonable for him
to be content, with having no more money than he has. And to say this is, in turn, to say
something like the following: the person does not (or cannot reasonably) regard whatever
(if anything) is unsatisfying or distressing about his life as due to his having too little
money. In other words, if a person is (or ought reasonably to be) content with the amount
of money he has, then insofar as he is or has reason to be unhappy with the way his life is
going, he does not (or cannot reasonably) suppose that money would—either as a
sufficient or as a necessary condition—enable him to become (or have reason to be)
significantly less unhappy with it.
15

Roughly, a person is content (or should be content) with what she has when the addition of
further economic resources will not give her reason to regard her life as less dissatisfying.
Frankfurt acknowledges that content individuals may nonetheless welcome having more. But, he
argues, content individuals lack an “active interest in getting more.” That is to say, “a contented
person regards having more money as inessential to his being satisfied with his life.”
16
If it is
indeed reasonable to expect individuals to be content with a finite amount of economic
resources, then it is plausible to suppose that individuals are primarily concerned with attaining

14
Paula Casal, offers a different consideration to the same conclusion:
…suppose that the medicine to be distributed does not prevent death but merely makes death less painful.
It would be unfair if all the pain-relieving medicine was distributed to the fortunate few. The fact that such
medicine cannot secure sufficiency does not make its distribution morally unimportant.
Paula Casal, "Why Sufficiency Is Not Enough," Ethics 117, no. 2 (2007): 307.
15
Frankfurt, "Equality as a Moral Ideal," 37-38.
16
Ibid., 39.
11

good enough or satisfying lives, and not the best or most satisfying lives possible. Thus
Frankfurt attacks the difference principle on the grounds that it presumes that rational persons
seek to maximize the benefits they can obtain.
17
Instead, a reasonable person, Frankfurt argues,
“might believe that the unequal distribution provided him with quite enough, and he might
reasonably be unequivocally content with that, with no concern for the possibility that some
other arrangement would provide him with more.”
18
In maintaining that a rational person may
be content with what she has under a particular distributive scheme, Frankfurt supposes that what
an individual is content with is largely independent of her position relative to others and of the
possibility of attaining more. And if, in fact, she can be content with her levels despite
occupying a position that is lower with respect to others or lower with respect to her
counterfactual position under a different distributive scheme, then what matters to distributive
justice is not that everyone enjoys relatively equal shares but rather that everyone enjoys enough
to feel that they can lead a satisfying life. What matters from a moral point of view is not
equality but rather sufficiency.
However, the claim that reasonable individuals can be content with their shares
independently of how much others have is untenable. What and how much we are content with
will depend on various historical and social contingencies. For example, Sally, who is growing
up in a rural village in a developing country, might be content with having access to fresh water,
shelter, food, police protection, the opportunity to acquire a high school education, and the
opportunity to acquire a decent occupation. Sue, who is growing up in an affluent, developed
country, needs much more to be content—she might require the opportunity to obtain a college

17
Parties to the original position who select justice as fairness, including its difference principle, are assumed to
rationally prefer more rather than fewer primary social goods. See John Rawls, A Theory of Justice (Cambridge:
Belknap Press of Harvard University Press, 1971), 142. The difference principle states that social and economic
inequalities are to be “to the greatest benefit of the least advantaged.” See ibid., 302.
18
Frankfurt, "Equality as a Moral Ideal," 36.
12

education, a wider range of occupational opportunities, and access to culture and the arts. The
point is not merely that individuals vary in their attitudes and that some are more easily content
than others. Rather, what it is reasonable to expect individuals to be content with depends on
what resources are available to others. Under conditions of bounty, it is reasonable for an
individual to be dissatisfied with what she would otherwise be content with under conditions of
scarcity. So even if it is reasonable to expect individuals to be content with a certain amount of
goods, we have not yet shown that an individual’s economic position relative to others is
superfluous to justice.
Moreover, Frankfurt equivocates between specifying the threshold according to what
people are content with and what they can reasonably be expected to be content with. However,
he will need to settle on a specification of the threshold according to what people can be
reasonably expected to be content with, not what they are actually content with. Some
individuals have expensive tastes and need far more than others in order to be content. If we
demand that justice secures for everyone what she is or would be content with, then we might be
committed to undesirable outcomes in which those who are not satisfied unless they enjoy luxury
goods are guaranteed levels that are much higher than those who affirm less materialistic values.
Others, on the other hand, may be self-sacrificial and be content with far less than others. More
generally, people may sometimes be mistaken or unreasonable in their own assessment of the
quality of their lives. To be plausible, the threshold must be more or less standardized so that it
applies in a similar way across individuals according to some objective identification of what
people need in order to lead a good enough life. So, Frankfurt should approve only of
distributions in which everyone is at a level they should be content with or, alternately, levels
13

which reasonable people would be content with. Henceforth, let’s suppose that the threshold is
set according to what it is reasonable to expect people to be content with.
My objection that what it is reasonable to expect Sally and Sue to be content with will
depend on what relevant others have illuminates a deeper problem with Frankfurt’s argument.
He assumes that we can identify what a reasonable individual is content with independently of
facts about what others have. It might be true that we can identify a person as being content or
not independently of seeing whether others are content. A solitary individual in the state of
nature is either content or not. If content, she is clearly content independent of facts about what
others have. But in order to identify what a reasonable person ought to be content with, we must
presuppose some background distribution. Suppose I know that an individual has access to clean
water, food, shelter, and basic medical care. Without any further knowledge about what others
have, I am not warranted in pronouncing that it is reasonable for her to be content with what she
has. I have no reason to think that its reasonable to expect her to regard the possession of
additional resources—such as access to higher-education—as inessential to a good enough life.
Moreover, individuals vary in their own identification of what a good enough life involves. For
some, membership in a religious community is essential to such a life; for others, inessential.
Once again, the sufficientarian must settle on a specification of the threshold level (and with it,
an identification of what the good enough life requires) that can be standardly applied across
individuals; however, in order to do this, reference must be made to the total resources available
to others. Under conditions of scarcity, it is reasonable to expect individuals to be content with
less than what they’d be content with under conditions of plenitude.
Frankfurt may respond by saying that though a comparison of one’s position relative to
others may be illuminating in our evaluation of whether we have reason to be satisfied with what
14

we have, it is nonetheless not essential. What is relevant to an individual’s assessment of his life
is “how closely the course of his life suits his individual capacities, meets his particular needs,
fulfills his best potentialities, and provides him with what he himself cares about.”
19
And these
questions can be answered without reference to others’ circumstances.
The following problem will then arise for Frankfurt. As I argued earlier, an assessment
of whether an individual leads a life she should be content with must take into account what she
can legitimately expect. For example, consider the deliberation that a reasonable person may
engage in when she wonders whether the course of her life suits her individual capacities. The
answer to this question may run the gamut from “not at all” to “completely.” It’s plausible to say
that if her life does not at all suit her individual capacities, then it is reasonable for her not to be
content with it. And if her life completely suits her capacities, then she should be content with it.
But how are we to evaluate the answers that lie between these two extremes? To what degree
should an individual’s life suit her capacities in order for her to have reason to be content? We
cannot answer this question without fixing several background facts, such as whether resources
are scarce or bountiful and what the exercise of her capacities costs to others. If, for example, I
live in a society in which resources are scarce enough so that no one else lives a life that
completely suits her capacities, then it might be unreasonable for me to demand additional
resources (at a cost to others) so that I can live a life that completely suits my capacities. But if I
live in a society in which resources are bountiful and everyone else leads a life that completely
suits her capacities, then it is reasonable for me to be content only when I too lead a life that
completely suits my capacities.
What is required, then, for a full defense of the sufficientarian threshold is an account of
the extent to which an individual’s comparison of her goods and position relative to others

19
Harry Frankfurt, "Equality and Respect," Social Research 64, no. 1 (1997): 7.
15

affects whether she has reason to be content with what she has. Egalitarians may intervene here
by predicting that an individual will only have reason to be content if her possessions are
consistent with a proper interpretation of equality. The suggestion that it is reasonable to expect
individuals to be content with a certain amount of goods is not inconsistent with the egalitarian
claim that distributive justice is concerned with people’s relative positions.
Moreover, one might question whether contentment is a suitable notion upon which to
build a distributive theory. After all, what it is reasonable to expect someone to be content with
will depend on what they can legitimately expect. But what individuals can legitimately expect
will depend on the distributive schemes they fall under. So what individuals should be content
with will vary according to the distributive schemes under which they find themselves. For
example, the worst off members of a society governed by the difference principle can expect that
any benefits to the better off will work to their advantage. It is reasonable for them to be
dissatisfied, then, if they do not benefit from a gain to the most advantaged. But the
sufficientarian will expect that they be content with what they have even if the most advantaged
acquire even greater benefits. Whether it is truly reasonable to expect them to be content even
when others are far more advantaged depends on whether one already agrees with the
sufficientarian view. To build our theory upon the claim that it is reasonable to expect
individuals to be content with a certain amount of goods is like attempting to build a house by
starting with the roof. We cannot know whether an individual can be reasonably expected to be
content with what she has unless we also know facts about the resources available to others and
about what she can legitimately expect. But these facts obtain only against a background
distribution. The fact that reasonable individuals can be content with what they have, then,
16

cannot serve as a reason to accept that what matters is that an individual has enough,
independently of what others have.
Other sufficientarian theorists have suggested alternative ways of setting the threshold.
Roger Crisp, for example, propounds a principle of sufficiency founded on the ideal of
compassion. Using several cases that compare the welfare levels of different groups, Crisp
contends that our intuitions about these cases clash with the prescriptions of egalitarian and
prioritarian principles. A principle of sufficiency, however, not only accords with but also
explains the basis for our intuitions about these examples.
According to Crisp’s Compassion Principle,
Absolute priority is to be given to benefits to those below the threshold at which
compassion enters. Below the threshold, benefiting people matters more the worse off
those people are, the more of those people there are, and the greater the size of the benefit
in question. Above the threshold, or in cases concerning only trivial benefits below the
threshold, no priority is to be given.
20

One way in which Crisp’s theory provides a more sophisticated version of sufficientarianism
than Frankfurt’s is by applying prioritarian principles to inequalities below the threshold.
Sufficientarian theories such as Frankfurt’s, which assign absolute priority to raising individuals
above the threshold, are vulnerable to objections that the theory might be committed, in some
cases, to pulling down the worst off to even lower levels in order to raise the levels of some who
are close to the threshold, thereby minimizing the number of people below the threshold.
21
For
example, suppose we could raise three people who are just below the threshold to adequate levels
by transferring resources to them from one person who is far below the threshold. Suppose that
those just below the threshold are only mildly ill but the person far below the threshold is on the
brink of death. If the threshold is good health, then sufficientarians who do not differentially

20
Roger Crisp, "Equality, Priority, and Compassion," Ethics 113, no. 4 (2003): 758.
21
See, for example, Arneson, "Egalitarianism and Responsibility," 236.
17

weigh the importance of raising the levels of those below the threshold are committed to
accepting the transfer because the transfer will raise three people to the threshold. When they
fail to incorporate prioritarian principles into their theory, sufficientarian theories produce the
strange result that there is no difference, in terms of moral urgency, between the claims of those
far below the threshold and those who are just below it. Crisp’s theory, on the other hand,
accommodates the intuition that those far below the threshold have a greater claim for aid than
those just below it.
Crisp does not offer much detail about the exact conditions under which compassion
applies, but he is careful to distinguish his theoretical notion of compassion from our ordinary
notion. In his theory, the notion of compassion relies on the device of an impartial spectator: the
threshold level is set at the point beyond which the impartial spectator no longer feels
compassion:
The Sufficiency Principle
Compassion for any being B is appropriate up to the point at which B has a level of
welfare such that B can live a life which is sufficiently good.
22

The compassion principle, supplemented by the sufficiency principle, provides the foundation
for an account of distribution “which allows us to give priority to those who are worse off when,
and only when, these worse off are themselves badly off.
23

While Crisp acknowledges that it remains to be seen what exactly constitutes a
sufficiently good life, he nonetheless provides intuitive support for the compassion principle with
an array of cases. For example, he presents the following case, called Beverly Hills:
24

10 Rich 10,000 Super-rich
Status Quo 80 90
Lafite 1982 82 90

22
Crisp, "Equality, Priority, and Compassion," 762.
23
Ibid., 757.
24
Ibid., 755.
18

Latour 1982 80 92

We have here a comparison between three scenarios in which varying amounts of wine are
distributed to the rich and super-rich. Egalitarian and prioritarian principles would prefer the
Lafite over the Latour distribution. However, Crisp claims that we should intuitively favor the
Latour outcome, since it is the outcome that maximizes aggregate utility. He further argues that
even if “the benefits to each of the Rich and the Super-rich are identical and their numbers are
the same, there still seems to [him] nothing to be said for giving priority to the ‘worse off.’”
25
In
this case, he concludes, justice is neutral between the two options: we have no more reason to
give to the Rich than the Super-Rich.
Let us grant that Beverly Hills is a fixed case—that is to say, external factors that are not
mentioned in the description of the case should not be relevant to our considerations about it.
This defuses objections, such as one put forth by Temkin, that in a world in which people are
starving, neither Lafite nor Latour represent good or preferable outcomes.
26
So if our only
options are to either give cases of wine to one group but not to the other, or to give the cases of
wine to neither group, then it seems we should give the cases of wine to one of the groups, so as
to avoid an outcome in which a good is put to waste. In Crisp’s example, we have two
competing values to take into consideration: increasing aggregate welfare, on the one hand, and
promoting distributions reflecting priority or equality, on the other. Suppose that giving the
cases of wine to the 10,000 super-rich will result in a net increase of 20,000 welfare points
(based on a 2-point increase per super rich individual) while giving the cases of wine to the 10
merely rich individuals will result in a net increase of only 20 welfare points (again, based on a
2-point increase per rich individual). If we are to give priority to raising aggregate welfare over

25
Ibid.
26
Larry Temkin, "Egalitarianism Defended," Ethics 113, no. 4 (2003): 771.
19

increasing the levels of well-being of the worse off, then we should give the cases of wine to the
super-rich and not the rich. If, on the other hand, we are to give priority to benefiting the worse
off, then we should give the cases of wine to the merely rich.
The intuition Crisp wishes to elicit from the reader is two-fold: first, “there is no case
based in justice for giving priority to those who are very rich over those who are super-
abundantly rich”
27
and second, since giving to the super-rich will result in an outcome containing
greater aggregate welfare, we should give to the super-rich. I argue that the role that our
intuitions about this case can play in leading us to any conclusions about prioritarian or
sufficientarian principles is far more limited than Crisp takes it to be. In response to the first, I
contend that I simply don’t share the intuition; moreover, I think the claim is too deep and
theoretically robust to be the content of a strong and clear intuition. Indeed, it is difficult to see
how one can have the conviction that there is no case based on justice for giving priority to the
rich over the super-rich without already having strong views about the content of justice. But the
content of justice is precisely what we are attempting to illuminate. The second part of Crisp’s
intuition relies on the first, and to the extent that the first is weakened, so is the second. The
most our intuitions about Beverly Hills can reveal to us is whether we already think it is more
important to increase aggregate utility or to increase the well-being of the worse off. The
Beverly Hills case, then, can only convince those who are already disposed towards
sufficientarianism. Ultimately, Crisp’s intuition about this case fails to provide any support of its
own for affirming a sufficientarian principle and rejecting a prioritarian one. And since it is the
basis for Crisp’s main objection to prioritarianism, Crisp has failed to establish a strong
foundation on which to reject the theory.

27
Roger Crisp, "Egalitarianism and Compassion," Ethics 114, no. 1 (2003): 121.
20

Crisp might reply that I have failed to emphasize one aspect of his case that is crucial to
his conclusion that once everyone is sufficiently well off, there is no reason to give priority to the
worse off. I had stated earlier that in deliberating about Beverly Hills we either give priority to
increasing aggregate welfare or to raising the levels of the worse off; what I had failed to observe
is the crucial description of the first option as increasing aggregate welfare when the levels of
well-being of all individuals involved are sufficiently high. The choice, then, is not between
increasing aggregate welfare simpliciter and increasing the levels of the worse off; rather, it is
between increasing aggregate welfare when everyone is above the threshold and increasing the
levels of the worse off (who are already above that threshold). Posed in this way, Crisp would
maintain that most people’s intuitions would side with increasing aggregate welfare (on the
condition that all affected are already above the compassion threshold), and their intuition that
we should increase aggregate welfare when everyone is already above the threshold is much
stronger than any intuition we might have that we should increase aggregate welfare simpliciter.
However, even when we foreground the fact that the worse off group under consideration
is affluent and far from badly off, it is problematic to assume that we intuitively take justice to be
completely indifferent between the option of benefiting the worse off or increasing aggregate
welfare. Even if we had the intuition that the wine should go to the super-rich rather than the
merely rich, several alternative explanations for this intuition are available to us. We could point
to the huge disparity between the numbers of people we could benefit in each case, for example.
If we give to the merely rich, we only benefit 10 people, but if we give to the super-rich, we can
benefit 10,000! Here, utilitarian considerations may attract us to the option of benefiting more
people, overriding any prioritarian impulses we might otherwise have. But this does not show
that, all else equal, justice does not require priority to be given to the worse off group. Indeed, in
21

order to show this, Crisp would need to present a case in which even though the number of
people affected and the extent to which they are affected are the same in each group, we still
have the intuition that justice is indifferent toward benefiting the worse off group. Assuming that
we share it, the most that Crisp’s intuition in the Beverly Hills case can show is that utilitarian
considerations can sometimes (but not always) outweigh prioritarian ones. And this observation
is not inimical to the prioritarian position; indeed, for the prioritarian, priority should be given to
the worst or worse off because the moral value of a benefit is greater, the lower the recipient’s
level of benefits.
28
When we are presented with cases in which we can either give a small
amount of benefits to the worse off or a larger amount of benefits to the better off, a calculation
of the overall value of benefits in the latter case may reveal it to be greater than in the former
case. This calculation is consistent with the Prioritarian view that what matters is distributing
benefits in such a way that their moral value is maximized.
Indeed, utilitarian considerations may not always outweigh prioritarian ones, even when
all groups involved are well above the compassion threshold. Consider the following
modification of the Beverly Hills case:
10 Rich 11 Super-rich
Status quo 80 160
Lafite 1982 82 160
Latour 1982 80 162

Even if one has the intuition that we should prefer the Latour 1982 scenario in the original
Beverly Hills case, one might not have the same intuition in this modified case. It seems
plausible to defend the intuition that Lafite is to be preferred here. Moreover, it seems
unreasonable that an aggregate increase in 2 welfare points should justify giving the goods to a
group whose welfare level is double that of the worse-off group. Keep in mind that though they

28
See, for example, Arneson, "Egalitarianism and Responsibility," 237.
22

are labeled “Rich” and “Super-rich,” the two groups are not to be thought of as differing solely in
economic well-being.
29
Rather, the numbers are meant to represent some unit of overall well-
being which may correspond to welfare, opportunities for welfare, access to advantages or
capabilities, or any other preferred metric. Considering that the Super-rich enjoy, in this case,
double the amount of overall well-being that the Rich enjoy, we should prefer benefiting the
Rich, even if the Rich are already well off.
Up to this point, the success of my argument turns on whether the reader shares my
intuition. So let us turn to another line of argument. Perhaps what is driving the intuition in the
original case is not the status of each group but rather the fact that the benefit to be distributed is
not one that really matters or that is substantial. This might explain why even if one is inclined
to benefit the rich over the super-rich, one would not feel that any great injustice has been done if
the benefit ends up going to the super rich. Indeed, what harm has been done if the already rich
lose out on some cases of wine? But suppose that instead of giving out wine to the two groups,
we were distributing some substantial good—health care, opportunity for education, political
representation of one’s group interests, or opportunities in careers, for example. Goods, in other
words, that are standardly taken to be the subject of distributive justice. If we were to pick a
substantial good to be distributed and to emphasize that the two groups differ not only
economically but in many other dimensions that matter to us as well, we have a different
intuition about the relevancy of considerations based on equality or priority than the one that
sufficientarians would hope for. Indeed, while sufficientarian theorists such as Crisp are careful
to accommodate considerations concerning the size of the benefit and the number of recipients,
not enough attention has been paid to the nature of the benefit in question. Indeed, our intuitions

29
As he emphasizes in a response to Temkin about his original Beverly Hills case; see Crisp, "Egalitarianism and
Compassion," 120-121.
23

about whether priority should be given to one group over another may often turn on whether the
benefit is trivial or substantial.
II TRIVIAL VS SUBSTANTIAL GOODS
Let us first turn to some cases provided by Yitzhak Benbaji, who suggested a few
amendments to Crisp’s compassion principle. Benbaji suggests that sometimes the value of
giving a sizable benefit to an individual or group outweighs the value of distributing that much
benefit evenly among many more people. He provides the following case, which he calls
Michelangelo:
Here we might give a very good chocolate to each member who belongs to a large group
of the badly off (they will gain tiny but nontrivial pleasure from it), or give the whole
quantity to a great artist, whose life is already very good. Yet, for some reason this
chocolate would bring the artist to perfection: thanks to it, he would paint something like
the Sistine Chapel. Suppose, finally, that the amount of well-being the artist gains is
smaller than the aggregate utility gained by the other individuals.
30

Benbaji supposes that our intuition about this case is that it is preferable to give the chocolate to
Michelangelo, since doing so will enable him to paint the Sistine Chapel. This case is meant to
show that, in some cases, priority should be assigned to increasing the size of a benefit over
spreading that benefit over a larger number of recipients. This leads Benbaji to defend the
following:
The Multilevel Doctrine of Sufficiency
Benefiting people matters more, the more priority lines there are above the utility level at
which these people are, the more of those people there are, and the greater the size of the
benefit in question. But, the number of beneficiaries matters less than the size of the
benefit.
31

This version of sufficiency departs from the theories of Crisp and Frankfurt insofar as it does not
posit a single threshold level above which inequalities are of no moral concern. Rather,
Benbaji’s principle is open to the possibility that several thresholds, or utility levels, may be

30
See Yitzhak Benbaji, "Sufficiency or Priority?," European Journal of Philosophy 14, no. 3 (2006): 335.
31
Ibid., 343.
24

salient. As with Crisp, Benbaji also proposes a theory that incorporates prioritarian principles
according to how far below each of the threshold levels an individual is.
However, perhaps the success of Benbaji’s case rests on the fact that the benefit in
question is a trivial one. If spread over many recipients, the chocolate would not contribute
significantly to their welfare nor to their productivity. When amassed and bestowed upon one
individual, however, the chocolate enables the recipient to contribute significantly to society. In
the former case, the benefit is trivial; in the latter, however, it can plausibly be described as a
resource which enables the recipient to yield a greater return in benefits to society. Perhaps
Benbaji’s claim that the size of the benefit matters more than the number of beneficiaries is
correct with regard to trivial benefits. However, it is not so clear whether it applies to cases in
which the benefit to be distributed is nontrivial or substantial or in cases in which we must
choose between distributing a trivial benefit or a substantial one.
To help us think about the case, I will provide a different, but structurally similar, one.
Let’s call this case Raphael. Members of a group of individuals—say, 50 people—have limited
opportunities for education. Suppose that scholarships are given out according to a two-step,
partly random, process. First, individuals must apply to be within an applicant pool. Of the
thousands who apply, 100 are selected at random. Then the scholarship will go to one individual
out of the 100 who shows the most merit, need or whatever other trait is relevant to that
scholarship. Suppose that we could exempt those 50 under-privileged individuals from the first
stage of this process and directly enter each of them into an applicant pool. There are, then, 50
applicant pools containing at least one under-privileged individual. This will result in a small
rise in their educational opportunities. Now, instead of providing small increases in educational
opportunities for each of these underprivileged individuals, we could instead enter Raphael—
25

who is well above the “good enough” threshold with respect to educational opportunities—
directly into 50 scholarship applicant pools. Given Raphael’s extraordinary merits, there is a
high chance that he will end up receiving many of these scholarships and go on to accomplish
much with them.
This case is meant to be analogous to Michelangelo. In both, we can either give small
but nontrivial benefits to many people or give a large benefit to one individual. In Crisp’s case,
we can give small pieces of chocolate to each of the under-privileged individuals or an enormous
mass of chocolate to Michelangelo. In my case, we can give small increases in educational
opportunities to each of the underprivileged individuals or we can pile on all those educational
opportunities onto one already educationally advantaged individual. Do our intuitions about
what we should do stay the same when we move from Michelangelo to Raphael? While I do not
think the intuitions we have about either case are very strong, I do believe that our preference to
giving the benefit to the better-off individual weakens as we move from the Michelangelo to the
Raphael case.
Or consider another modification of Crisp’s Beverly Hills scenario. Suppose that instead
of distributing cases of wine, we are instead considering how to distribute access to health care:
10 Middle-class 10,000 Billionaires
Status quo 80 90
Plan A 82 90
Plan B 80 92

Suppose that members of the middle-class are above the “good enough” threshold. They can
afford their health care plans, and those plans provide an adequate amount of coverage, but the
provision of Plan A would provide them with even greater coverage, thus raising each middle-
class individual’s level by two points. In response to the original Beverly Hills case, even
readers who have the intuition that the wine should go to the merely rich, as opposed to the
26

Super-rich, will not be so concerned if the wine ends up going to the Super-rich. After all, it
does not seem to really matter if the already rich lose out on some cases of wine. However, the
suggestion that we should select Plan B in this case is far more unattractive. One might argue,
for example, that it is already much easier for the billionaires to gain access to greater health care
than the middle-class. They easily have the means to travel to hospitals that employ the most
talented doctors or that have the most advanced equipment, and their pursuit for superior, as
opposed to merely adequate, health care is feasible for them, given their plentiful economic
resources. In comparison, it may not be feasible for the middle class individual to pursue that
same superior level of health care. The pursuit of superior health care may be impossible or very
costly to her. While the middle-class individual can pursue superior health care only by directing
a substantial portion of her resources to that goal, thus sacrificing her pursuit of other
opportunities or projects which also may be dear to her, the billionaire has the purchasing power
to obtain the superior health care with relative ease and may not need to make much of a
sacrifice. In light of these considerations, I think there is a strong case for saying that members
of the middle-class group do have a greater claim to the extra resource than the billionaires.
These considerations show that the intuitive support that theorists such as Crisp and
Benbaji rely on succeeds only with respect to cases involving trivial benefits. When the benefits
under consideration are substantial, however, our intuitions may align with egalitarian or
prioritarian prescriptions. This poses a substantial problem for sufficientarianism since
distributive justice is primarily concerned with the distribution of substantial goods rather than
trivial ones.
Naturally, the next question to pursue is why exactly our intuitions about cases such as
those previously discussed should depend on the importance of the benefit being distributed.
27

Why might our intuitions agree with sufficientarian principles with respect to cases involving
trivial goods but align with egalitarian or prioritarian ones when we consider the distribution of
substantial goods? We may begin to answer this question by observing that in cases involving
substantial goods, considerations of utility may be countervailed by other competing values, such
as those of respect, community, and interpersonal justification.
III INTERPERSONAL JUSTIFICATION AND DEMOCRATIC COMMUNITY
Recall the Raphael case. Here, we deliberate about whether to bestow a wealth of
educational opportunities on Raphael or to distribute those opportunities evenly over a group of
50 individuals who are much less well-off than Raphael. My contention was that even if we do
not reject Benbaji’s intuition in Michelangelo that we should give all the chocolates to
Michelangelo, we might nonetheless reject the analogous suggestion to benefit Raphael. Even
though Michelangelo is structurally similar to Raphael, we have different intuitions about
whether the size of the benefit trumps the number of potential recipients. This difference shows
that sometimes the nature of the benefit in question matters as well. I think that one reason why
we would favor benefiting more people in Raphael because this option better represents a
distribution which reflects the idea that goods should be distributed in a way that supports the
attainment of a fully functional democratic community—that is, a community of equal citizens
who can justify to one another the terms upon which they enjoy their place in the social,
economic, and political scheme, thereby upholding the ideal of respect for one another.
The sufficientarian might wish to argue, however, that one might justify giving all the
scholarship opportunities to Raphael rather than distributing them out to the 50 others because
while Raphael can predictably accomplish great things with the scholarships he might receive, it
is less likely that any of the underprivileged individuals will be able to accomplish much even if
28

they received a scholarship. Moreover, perhaps Raphael alone can accomplish more than can be
accomplished by all 50 individuals combined! If we assume that Raphael’s accomplishments are
to the benefit of society, considerations of utility or efficiency lead us to prefer benefiting
Raphael over the 50 other individuals.
However, this line of response is not available to the sufficientarian. Suppose, to
strengthen my opponent’s case, that the 50 individuals already have enough educational
opportunities to enable them to participate in democratic processes. This means that they have
the opportunities to acquire a basic understanding of the issues surrounding the adoption or
repeal of policies and laws, acquire the knowledge necessary for them to make informed
decisions when voting, and to articulate and their reasons for their political views to others.
There are nonetheless considerations in favor of strengthening the educational opportunities of
the 50 individuals, especially if the opportunities open to them are not bountiful enough for them
to feasibly acquire education that is more advanced than that required for full democratic
participation. Firstly, it is often the case that when an individual enjoys greater educational
opportunities, the occupational opportunities available to her increase as well. In addition,
generally speaking, one is more likely to obtain one’s preferred occupation when one has a wide
range of occupational opportunities available. And the obtainment of one’s first choice
occupation is one of the most important factors determining one’s quality of life. It is rational,
then, for each individual to prefer to obtain goods that will widen her range of occupational
opportunities over maintaining her status quo, since this will increase the likelihood that she can
enjoy a fulfilling and meaningful life. To provide Raphael increased educational opportunities
when he already enjoys so many and is already more likely to obtain his preferred occupation
would be to encourage a distribution which may undermine any sense of fraternity between the
29

well-off and the less well-off. Unless the other 50 individuals have reason to regard the
increased benefit to Raphael as justified on the grounds that it would make them better off than
they would be were they to receive the extra opportunities, the increased inequality between
Raphael and the other 50 individuals may provide grounds for them to feel envy or resentment
toward Raphael and to regard the distributive scheme as insufficiently sensitive to their
interests.
32

Note, also, that the sufficientarian cannot say that benefits to Raphael are justified on the
grounds that they benefit the other 50 as well. Since everyone in this case is above the threshold
level of educational opportunities required for democratic participation, sufficientarian theories
would say that any inequality above that threshold is permissible. So even if the benefits to
Raphael do not redound to the benefit of the other 50, sufficientarianism is committed to
regarding that distributive scheme as permissible. I think that this undermines the likelihood of
attaining a political society of equal citizens who can justify to one another the terms upon which
they enjoy their position in the social, economic, and political scheme, thereby upholding the
ideal of respect for one another. Consider how difficult it might be for Raphael to justify to one
of the 50 individuals his enjoyment of additional educational opportunities at the expense of
increased opportunities that could have gone to them instead.
Perhaps this is not such a problem for the sufficientarian. She might point out that while
sufficientarian theories deny the intrinsic value of economic equality, they are not committed to
denying the value of social and political equality. Cases involving substantial goods, such as

32
Indeed, one may even point out, simply, that such a distribution seems intuitively unfair. See Harry Brighouse
and Adam Swift, "Educational Equality Versus Educational Adequacy: A Critique of Anderson and Satz," Journal
of Applied Philosophy 26, no. 2 (2009): 125. Additionally, Larry Temkin writes “As important as considerations of
sufficiency and compassion may be, fairness matters too. And considerations of fairness do not lose their force
simply because someone is sufficiently well off that he does not elicit our compassion” in Larry S. Temkin,
"Equality, Priority or What?," Economics and Philosophy 19, no. 1 (2003): 65.
30

educational opportunity or healthcare, do not provide counterexamples to sufficientarianism, for
those goods are to be distributed in accordance with social and political equality. For example,
unless everyone has access to her equal share of educational opportunities or health care, social
equality has not been adequately upheld.
This line of response, however, is not open to Frankfurt, Crisp, or Benbaji. I take it that
all three envision their distributive principles as appropriate for a democratic society and that
they affirm the ideals supporting the value of democracy. Economic sufficiency must be shown
to be compatible with, if not directly supporting, the aim of realizing a community of citizens
who can participate in democratic processes as equals. If economic inequalities beyond the
sufficientarian threshold can be shown to undermine the attainment of a fully functional
democratic community, then we have a compelling reason to reject the sufficientarian position.
We have, then, several challenges to the sufficientarian view. First, sufficientarians must
show that we can isolate the political arena from inequalities in social and economic status.
Generally speaking, those with great income and wealth can attain greater political influence, not
only by having the resources to encourage others to vote for their preferred policies but by
having greater resources for attaining political offices as well. Their theories must properly
regulate the extent to which greater income and wealth can bolster one’s political power and
influence. Second, economic inequalities must clearly be constrained by the requirements of
social and political equality. Sufficientarian theories must uphold the claim that economic
inequalities that support social inequalities (or vice versa) are impermissible.
In response to these challenges, Elizabeth Anderson articulates one of the most attractive
versions of sufficientarianism. She accommodates the motivations behind egalitarian views by
identifying the primary egalitarian aim as securing “for everyone the social conditions of their
31

freedom in terms of capabilities”
33
and arguing that achieving this aim requires that everyone has
effective access to the levels of functioning sufficient to stand as an equal in a democratic society
over the course of an entire life.
34
This means that each enjoys the capabilities necessary to
enable them to avoid or overcome oppression and to participate in democratic processes as an
equal citizen.
35
For Anderson, these necessary capabilities are informed by the different aspects
of individual functioning “as a human being, as a participant in a system of cooperative
production, and as a citizen in a democratic state.”
36
Democratic equality is achieved when each
citizen has access to the means that would enable them to fully function in these three ways.
Once democratic equality has been achieved, however, no further measures need be taken
to redress other kinds of inequalities, including economic inequality. In comparing her theory
with the theory of equality expressed by Rawls’ Difference Principle, Anderson notes that hers
requires a less demanding form of reciprocation and is more permissive of economic inequality:
“Once all citizens enjoy a decent set of freedoms, sufficient for functioning as an equal in
society, income inequalities beyond that point do not seem so troubling in themselves.”
37
On her
view, inequalities in wealth are only derivatively unjust when they proliferate inequalities in
social and political relations. In itself, however, economic inequality is not of primary
importance to justice. Once each citizen has secured the capabilities necessary for her to
participate in democratic processes and to engage in social relations—both public and private—
that are free from oppression, other forms of inequality, including differences in wealth, needn’t
be redressed.

33
Elizabeth Anderson, "What Is the Point of Equality?," Ethics 109, no. 2 (1999): 316.
34
Ibid., 319.
35
Ibid., 316.
36
Ibid., 317.
37
Ibid., 326.
32

I agree with Anderson that a democratic society should secure for its citizens the
capabilities needed to function as a human being, as a member of a cooperative system of
production, and as a member of a democratic society. However, I reject her negative thesis that
economic inequalities beyond that threshold are no longer a concern of justice. Again, in order
to attain a democratic community of equal citizens, members must be prepared to justify to one
another the terms by which they enjoy their place in the social, economic, and political scheme.
And I think that Anderson would be committed to holding this claim as well, since her threshold
capabilities are meant to capture, among other things, what is required for individuals to function
“as a participant in a system of cooperative production.”
38
A system is not cooperative unless its
participants can accept the terms of their interactions and commitments to one another.
Moreover, acceptance of those terms relies on whether or not those terms can be justified to one
another. Cooperative enterprises are aimed at mutual advantage; if one party does not stand to
gain anything while the other party does, the two parties are engaged not in cooperation but
rather in coordinated exploitation, benevolence, or some other scheme involving asymmetrical
expectations.
I agree with Anderson’s suggestion that no one should fall below the threshold level of
access to the capabilities necessary for full democratic participation and the avoidance of
oppressive social relations. But in order to attain a fully functional democratic society,
distributive justice must reflect the values of mutual respect, cooperation, and interpersonal
justification. The negative sufficientarian thesis that inequalities above the threshold level are no
longer the concern of distributive justice may lead to distributions that come into conflict with
those values.

38
Ibid.
33

I will introduce here a basic notion of interpersonal justification. Roughly, a policy P is
interpersonally justified if and only if reasons or arguments can be given for P which reasonable
addressees could accept in light of their common interests as democratic citizens.
Sufficientarian views do not reject distributions of great economic inequality, unless
those distributions fail to secure what reasonable citizens can be content with (Frankfurt),
benefits that can address the compassionate concern of an ideal spectator (Crisp), or the requisite
capabilities for full participation in a democracy (Anderson). It is commonly recognized that
income and wealth enable people to pursue their own plans of life. But it should also be
acknowledged that greater income and wealth confers greater freedom in pursuing that plan in
addition to allowing a greater range of different plans of life to be feasible. Distributions that
allow vast economic inequalities, then, are also distributions that allow inequalities in citizens’
freedoms to choose and pursue life plans. Given the fact that all citizens have a higher order
interest in that freedom, such distributions cannot be interpersonally justified.
In conclusion, by posing objections to each of the three sufficientarian views, I hope to
motivate the claim that economic equality is of moral importance. In the proceeding chapters, I
will discuss the theories of three prominent liberal egalitarians—Ronald Dworkin, John Rawls,
and G.A. Cohen—and provide considerations in favor of my proposed Burdens Compensation
Principle, which reflects what I take to be an intuitive conception of economic equality that
better addresses both luck egalitarian and feminist concerns than do these influential liberal
egalitarians.

34

2 Responsibility-Sensitive Egalitarianism and Luck

Responsibility-sensitive egalitarianism was first articulated in the work of Ronald Dworkin,
who viewed Rawls’s Difference Principle as “insufficiently sensitive to those with natural
handicaps, physical or mental, who do not themselves constitute a worst-off group…”
39
While
the Difference Principle regulates inequalities between groups
40
whose position of advantage is
measured by social primary goods,
41
Dworkin sought to provide a device for aiming at “equality
of resources person by person” [my emphasis].
42
On this view, a theory of distributive justice
should hold individuals responsible for their preferences but offer some form of compensation
for handicaps and other outcomes of natural misfortune. The core idea that distributive justice
should, in a certain sense, track responsibility—that is, by compensating individuals who suffer
through no fault or choice of their own as a result of factors beyond their control—became taken
up by many other political philosophers working in the wake of Rawls’s Theory of Justice.
43

39
Ronald Dworkin, "What Is Equality? Part 2: Equality of Resources," Philosophy & Public Affairs 10, no. 4
(1981): 339.
40
The “least well-off” group, for example, is taken to be either all those with the average income and wealth of the
unskilled worker, or “all persons with less than half of the median income and wealth.” See Rawls, 98.
41
Primary goods are “all-purpose means that are generally necessary to enable citizens to… pursue their determinate
conception of the good;” social primary goods in particular are the basic rights and liberties, freedom of movement
and free choice of occupation against a background of diverse opportunities, the powers and prerogatives of
positions of authority, income and wealth, and the social bases of self-respect. John Rawls, Justice as Fairness: A
Restatement (Cambridge: Harvard University Press, 2001), 57-58.
42
Dworkin, "What Is Equality? Part 2: Equality of Resources," 340.
43
See, for example, Richard Arneson, "Equality and Equal Opportunity for Welfare," Philosophical Studies 56, no.
1 (1989); Arneson, "Egalitarianism and Responsibility."; Richard Arneson, "Luck Egalitarianism and
Prioritarianism," Ethics 110, no. 2 (2000); Richard Arneson, "Luck and Equality," Aristotelian Society
Supplementary Volume 75, no. 1 (2001); Richard Arneson, "Luck Egalitarianism: An Interpretation and Defense,"
Philosophical Topics 32, no. 1 (2006); Nicholas Barry, "Defending Luck Egalitarianism," Journal of Applied
Philosophy 23, no. 1 (2006); Gerald A. Cohen, "On the Currency of Egalitarian Justice," Ethics 99, no. 4 (1989);
Alexander Kaufman, "Choice, Responsibility and Equality," Political Studies 52, no. 4 (2004); Carl Knight, "In
Defence of Luck Egalitarianism," Res Publica 11, no. 1 (2005); Carl Knight, Luck Egalitarianism: Equality,
Responsibility, and Justice (Edinburgh: Edinburgh University Press, 2009); Eric Rakowski, Equal Justice (Oxford:
Oxford University Press, 1991); John Roemer, Theories of Distributive Justice (Cambridge: Harvard University
Press, 1996); Kok-Chor Tan, "A Defense of Luck Egalitarianism," The Journal of Philosophy 105, no. 11 (2009);
Temkin; Temkin, "Egalitarianism Defended."; Larry Temkin, "Exploring the Roots of Egalitarian Concern," Theoria
69, no. 1-2 (2003).
35

This line of thought is commonly known as luck egalitarianism.
44
In this chapter, I will focus on
the seminal work that generated this literature: Dworkin’s theory of equality of resources, which
I will also call the Auction Theory.
The first section of this chapter will provide a brief overview of Dworkin’s theory.
According to this view, the metric of distributive justice should be resources, not welfare, and
equality of resources can be articulated through a progression of three hypothetical auctions:
first, an auction in which individuals are given an equally large amount of currency with which
to bid for their preferred bundle of resources, subject to a restriction that Dworkin calls the Envy
Test; second, an auction that provides individuals with the opportunity to insure themselves
against handicaps, disabilities, and other medical conditions; and third, an auction that provides
individuals with the opportunity to insure themselves against unemployment and
underemployment. The first auction establishes an initial equal distribution of resources while
the second and third insurance auctions are intended to correct for any divergences from equality
that arise from natural misfortunes such as handicaps and lack of natural talent. In the second
section, I discuss Dworkin’s proposal that an income tax could serve as a best approximation of
the two auctions. I then offer two tax schemes that seem to be the most plausible schemes that
could be consistent with the Envy Test. Both tax schemes, I argue, are indefensible. One allows
distributions of resources that are intuitively unfair and conflict with the underlying motivations
for Dworkin’s theory of equality while the other results in inefficient distributions. The failure
of these two tax schemes suggests that the Envy Test fails to provide a necessary condition for

44
What I call “responsibility-sensitive egalitarianism” is, roughly, luck egalitarianism; I use the former term in
reference to Dworkin, however, as he denies holding a luck egalitarian view. He explains that his theory does not
require that “people be fully compensated for any bad luck after it has occurred, but rather that people be made
equal, so far as this is possible, in their opportunity to insure or provide against bad luck before it has occurred, or, if
that is not possible, that people be awarded the compensation it is likely they would have insured to have if they had
had that opportunity.” See Ronald Dworkin, "Equality, Luck and Hierarchy," Philosophy & Public Affairs 31, no. 2
(2003): 91.
36

the equality of a distribution of resources. In the third section, I present and reply to the
objection that the Envy Test plays no role in measuring the equality of post-tax distributions. In
the fourth section, I briefly introduce the Burdens Compensation Principle and show how it can
better accommodate the various thought experiments that have been shown to challenge the
Auction Theory.
I EQUALITY OF RESOURCES AS MODELED BY A HYPOTHETICAL AUCTION
Liberal egalitarian theories typically begin with the supposition that a just distribution takes
equality to be a default starting position. The next steps for these theories are to specify what
exactly is to be distributed—whether welfare, resources, or capabilities—and to specify the
conditions under which divergences from equality are permissible. Rawls’s theory of justice as
fairness, for example, begins with a general conception of justice according to which resources
(namely, social primary goods) are to be distributed equally unless an unequal distribution would
be to the benefit of all.
45
Together with the premise that unequal distributions of wealth may
indeed be to the benefit of all (and, more importantly, maximally raise the levels of the least well
off), the general conception then gives rise to the Difference Principle: “Social and economic
inequalities … are to be to the greatest benefit of the least-advantaged members of society.”
46

The Difference Principle allows divergences from equality of income and wealth only when the
greater advantages of the better off redound to the greatest benefit of the least advantaged.
Dworkin’s Auction theory similarly begins with an equal division of resources. We are first
asked to consider a hypothetical scenario in which the need for distributive principles arises for
survivors stranded on an island. We assume that the survivors do not have any prior entitlements

45
Rawls, A Theory of Justice, 62.
46
Rawls, Justice as Fairness: A Restatement, 42-43.

37

to any of the resources on the island. The survivors also accept that the following test specifies
the conditions under which a distribution may be regarded as equal:
The Envy Test
No division of resources is an equal division if, once the division is complete, any immigrant
would prefer someone else’s bundle of resources to his own bundle.
47

Roughly, the Envy Test upholds the intuitively plausible idea that an equal division of resources
will be one in which people are satisfied with what they have in comparison to others. The next
task for the newly arrived islanders is to figure out a way to arrive at a division of resources that
would pass the Envy Test.
One way they could arrive at an equal distribution of resources is to have an individual tally
up all the resources of the island—such as land, clean water, vegetation, animals, shoreline, and
so on—and divide these resources so that everyone has an equal bundle of these resources.
However, we can easily see that there will be several problems with this method. Some of these
resources are not uniform in quality—for example, some of the land will be more arable and
fertile than the rest—and other resources may be such that it is impossible to equally divide
them.
48
How, for example, might the islanders divide 12 boars among 50 survivors? Even if a
strictly equal division of resources were achievable, such a division would, more likely than not,
fail to efficiently reflect the differences in people’s preferences or needs. The vegetarian
survivors might prefer a bundle that has more mangos in lieu of their share of the boar, for
example, while meat eaters may prefer the opposite. A natural solution to these difficulties is to
introduce the device of a market.
The divider is transformed into an auctioneer, and each survivor is given an equally large
number of clamshells with which to bid. Each item on the island is to be sold as a lot in the

47
Dworkin, "What Is Equality? Part 2: Equality of Resources," 285.
48
Ibid.

38

auction, unless a survivor expresses a desire to bid on a part of an item, in which case that part
will be listed as a distinct lot. An initial set of prices for the lots is then proposed, and the
auctioneer runs the auction to see whether all lots are sold with only one purchaser per lot. If
not, then the auctioneer must adjust the prices. The survivors may, in addition, change their bids
or propose different lots. We can surmise that this process will eventually come to an end, and
each will be satisfied with her lot. The goods will then be distributed accordingly. This
distribution passes the Envy Test because if anyone had preferred another’s bundle of resources,
she could have purchased the bundle or have proposed different lots, in which case a different set
of prices will have been cast and the auction restarted.
49

This hypothetical auction illustrates the appeal of a theory that identifies resources, as
opposed to welfare, as the metric of distributive equality. For while a theory of equality of
welfare will be required to provide plover’s eggs and pre-phylloxera claret for the bon vivant
who cannot reach a level of welfare comparable to that of others without such luxuries,
Dworkin’s resource theory—in which equality of resources is modeled after a hypothetical
auction—requires individuals to make their choices against a background of information about
the total stock of resources available, the competing preferences in play, and the impact of their
choices on others.
50
This informs each member of the community of what she can reasonably
expect as her share of the total stock of resources. For example, if the island happens to have
expensive luxuries, such as high desired plover’s eggs and pre-phylloxera claret, the individual
who insists on a bundle of resources that includes these goods must accept that she must make
sacrifices with respect to other goods. The bon vivant pays the price for her choice to indulge her
taste for luxury goods, and this price is determined by what others must give up. She enjoys the

49
Ibid., 287.
50
Ibid., 288. For a discussion of the problem of expensive tastes for theories of equality of welfare, see Ronald
Dworkin, "What Is Equality? Part 1: Equality of Welfare," Philosophy & Public Affairs 10, no. 3 (1981).
39

luxury goods but few other resources while others give up the luxuries but enjoy, instead, a
greater variety and volume of less “expensive” resources. The preceding examples illustrate how
the Auction Theory is sensitive to two considerations: (i) that equal distribution of resources will
involve different types and quantities of resources, in accordance with people’s differing
preferences, and (ii) the resources an individual can expect to have will at least partly depend on
the preferences of others and the total amount of resources available.
With an interpretation of the equal initial distribution in place, we must next consider how
and whether equality of resources can be maintained over time. Suppose an islander chooses a
lot that contains a large area of shoreline but few other resources with the expectation of
capitalizing on her expert spear-fishing abilities. However, she later develops cataracts and goes
blind. While her initial set of resources seem fair, given that they are the result of an auction that
passes the Envy Test, many of us would find that her stroke of bad luck introduces a morally
significant complication. It does not seem fair that, even though her lot is consistent with
equality of resources, she be left to suffer disadvantages as a result of her physical affliction. Or
suppose that one of the survivors arrives on the island with a pre-existing handicap. Although
she does not prefer anyone else’s bundle of resources, the individual with the handicap may
nonetheless have be reasonably dissatisfied with her lot, given her inability to utilize her
resources as efficiently as others.
The auction provides us a clever device for illustrating how equality of resources can be
sensitive to the responsibility that individuals should bear for their preferences and choices. But
it is not enough for equality of resources to hold people responsible for their preferences and
choices; it must also be sensitive to the distinction between option luck and brute luck: option
luck being “a matter of how deliberate and calculated gambles turn out” and brute luck being a
40

“matter of how risks fall out that are not in that sense deliberate gambles.”
51
The distinction
between these two outcomes of luck is significant insofar as it purports to distinguish two types
of disadvantages: those we should be held responsible for because they, in turn, reflect our
choices and preferences for which we are responsible, on the one hand, and those for which we
should not be held responsible because they reflect features of our unchosen circumstances or
other factors beyond our control, on the other. As is suggested by Dworkin’s description of
option luck, a paradigmatic example of such luck is the luck of the gambler at a casino: the
gambler makes a deliberate choice to accept the risks of the gamble in exchange for the
possibility of some monetary gain. Brute luck, on the other hand, is not the outcome of a
deliberate choice but rather is a matter of factors beyond one’s control. Someone is said to suffer
from bad brute luck if she is struck by natural misfortunate or otherwise disadvantaged through
no fault or choice of her own, such as when she loses her house in a hurricane, inherits a genetic
disorder, or is hit by a drunk driver.
If individuals are indeed responsible for their option luck, then it seems fair for distributive
theory to allow divergences from equality that reflect differences in option luck. But how should
an egalitarian theory deal with the problem of bad brute luck? One can attempt to equalize every
one’s bad brute luck by pushing down all those who enjoy natural advantages to the level of the
least lucky. This suggestion is clearly one we would not wish to take up. Another alternative,
which Dworkin proposes, is to neutralize the unfairness of bad brute luck by providing a way for
it to be transformed into option luck through the device of an insurance auction.
52
Once again,
each islander is given an equal number of clamshells with which to bid. In this second scenario,
however, we must impose a thin “veil of ignorance.” We suppose that no one has knowledge of

51
Dworkin, "What Is Equality? Part 2: Equality of Resources," 293.
52
ibid., 297.
41

their handicaps or other preexisting conditions, nor of their likelihood of developing or
contracting certain disorders or diseases. As in the first auction, the auctioneer begins with a set
of prices and a division of lots. In this case, the lots contain insurance for various medical
conditions, handicaps, disabilities, and other natural misfortunes.
If individuals are given an equal opportunity to insure themselves against bad brute luck,
then the compensation they would receive when they do turn out to suffer from bad brute luck
would help ameliorate the unfairness of their situation. The idea is that when individuals
purchase insurance against bad brute luck, the outcome that results becomes a matter of option
luck. Of course, although it is not up to them whether they are struck with bad luck, the amount
of compensation they receive in case they are is a matter of their prior choice to purchase
insurance. And while the unfairness of being struck with bad brute luck is not completely
eliminated, it is at least partially relieved by the opportunity to insure oneself against it. A
system of redistributive taxation that mimics this hypothetical insurance market would similarly
mitigate the unfairness of bad brute luck by compensating those individuals according to the
amount they would have received had their bad brute luck been transformed into option luck via
the use of an insurance market.
Though Dworkin himself never explicitly foregrounded the distinction between
responsibility and circumstance as the chief motivation behind his theory, one may readily see
that the distinction between option luck and brute luck in his theory gains its moral force from its
alignment with the distinction between the outcomes of responsible choice, on the one hand, and
circumstances on the other. Indeed, G.A. Cohen hailed Dworkin’s theory as having, “in effect,
42

performed for egalitarianism the considerable service of incorporating within it the most
powerful idea in the arsenal of the anti-egalitarian right: the idea of choice and responsibility.”
53

II TALENTS AND LUCK IN PRODUCTION AND TRADE
However, even if the two auctions establish a distribution of resources that appropriately
responds to individuals’ preferences and choices and provides them an equal opportunity to
insure themselves against future misfortune, the initially equal distribution will be compromised
with the introduction of trade and production. In particular, the natural distribution of
differential talents involved in production and trade will disrupt the equality of the distribution of
resources over time. To see this, we will contrast two cases. In the first case, we will look at the
preferences and ambitions of two individuals and see that the distribution that results from their
differing preferences and ambitions indeed passes the Envy Test. In the second case, the two
individuals share similar preferences and ambitions but differ in their native abilities. The
distribution that results from their unequal talents fails the Envy Test.
Let’s first consider the case in which trade and production result in distributions that are
consistent with equality. Suppose that everyone has equal talents. Sally forms the ambition to
amass a relatively large amount of resources by spending 12 hours per day spear-fishing and
trading her fish for other goods. Sue, on the other hand, prefers a more leisurely daily routine
and spends only three hours spear-fishing, which is just enough to allow her to spend five hours
napping in a hammock. Over time, Sally will amass far more resources than Sue; will Sue then
prefer Sally’s bundle of resources, thereby showing that the distribution fails the Envy Test?
Here, Dworkin argues that we should consider each individual’s bundle of resources as
containing more than just the resources themselves—they will also make an essential reference
to the means by which those resources were gained. Accordingly, Sally’s bundle of resources

53
Cohen, "On the Currency of Egalitarian Justice," 933.
43

consists not only of the goods that she has gained through trade but the twelve-hour workday,
with its attendant daily lost leisure time, required for her to have been able to amass that fortune.
This leads us to the observation that the Envy Test is to be applied diachronically. It considers
the bundle of resources an individual has over a lifetime and includes the means by which those
resources are acquired. Sue, then, will not prefer Sally’s bundle of resources because she would
not prefer to take on the twelve-hour workday, despite the larger amount of goods that she could
thereby gain through trade.
However, when the distribution of natural talents and abilities is unequal, the distribution
of resources that result from production and trade will reflect that inequality. Moreover, the
resulting distribution of resources will not pass the Envy Test. Suppose, for example, that both
Sally and Sue form the ambition to amass a great amount of wealth. They both spear-fish for
twelve hours a day. But while Sally’s keen eye, quick reflexes, and naturally lithe movements
make her particularly adept at spearfishing, Sue’s clumsy and blundering ways make her an
inefficient spear fisher. Sally is able to catch 54 fish out of the 60 that she attempts to spear; Sue,
on the other hand, can only catch a pitiful 16 out of her 60 attempts. In this case, it is reasonable
for Sue to prefer Sally’s bundle of resources. Not only would Sue prefer to have Sally’s
resources, she would prefer to have those resources even when they entail a twelve hour
workday; after all, this is already part of her much smaller bundle. This preference shows us that
not all distributions of resources resulting from production and trade will pass the Envy Test
when there is an unequal distribution of natural talents and abilities.
In order to maintain equality of resources over time, the Auction Theory must deal with
the unequalizing effects of the natural distribution of talents and abilities. Unequal talents
introduce a particularly difficult complication for equality of resources. On the one hand, we
44

have the idea that there is a difference in circumstance, rather than choice, between the
individual who is successful at production and trade as a result of certain native talents and the
individual who does not possess those same native talents but who nevertheless forms the same
ambitions and exerts as much effort as the first person. On the other hand, the talents, skills and
abilities involved in successful production and trade are not always a matter of genetic luck;
rather, such talents are often inextricably linked to people’s ambitions and choices. Typically,
individuals are not born with full-fledged abilities that automatically vault them into the
occupations that capitalize on those natural advantages, and natural talents do not, by themselves,
guarantee individuals the competitive edge they need to be successful in the market. Rather,
individuals must develop and hone those native talents into advanced skills and abilities. Some
lucky individuals may have a wide variety of natural talents from which to choose. They may
decide that they could do more social good in the future by being a doctor than by being
gymnast, for example, and so choose to develop the talents that will enable them to become good
doctors rather than the talents that will enable them to become good gymnasts. On the flipside,
genetically unlucky individuals may still work hard and succeed in developing the talents, skills
or abilities required for them to successfully pursue their ambitions. How would a market
determine whether the success of an individual is attributable to her native endowments or her
choices? Moreover, the connection between ambitions and talents can also run the opposite way:
it is plausible to say that individuals have a natural propensity to form their ambitions according
to the natural talents they have. The closely intertwined relation between talents and ambitions
means that no change in the division or variety of the stock of goods in the auction can yield
distributions that are sensitive to people’s ambitions but at the same time do not reflect
differential native endowments.
45

III THE INCOME TAX AS COMPROMISE
In response, Dworkin’s suggestion is to introduce a periodic redistribution of resources in
the form of an income tax.
54
Of course, an income tax is not a perfect device for maintaining
choice-sensitive and circumstance-insensitive distributions, for reasons noted above. But, he
argues, it comes as close as possible to allowing people freedom in developing their course of
life while at the same time recognizing the role that genetic luck plays in the development of that
life.
55
The result, Dworkin claims, is a theory that reaches a compromise between two competing
ideals within equality: the requirement that equality of resources should hold people responsible
for paying the cost of the life they choose to lead (measured in terms of the cost of that life on
others), and the idea that equality of resources should seek to maintain distributions that do not
reflect differential natural talents and other native endowments.
It seems that Dworkin’s auction theory already provides a natural way of arriving at an
appropriate tax scheme. Suppose that Sue is just as willing as Sally to work twelve hours a day
at spear fishing. Recall that while Sally’s natural grace makes her especially well suited to this
trade, Sue’s ambitions are thwarted by her insurmountable clumsiness. Given that the only
difference between the two individuals lies in their differential realized natural talents, it is
reasonable for Sue to prefer Sally’s bundle. Because the distribution fails the Envy Test, we may
then propose the following method for arriving at an appropriate tax scheme. First, impose a tax
on Sally that redistributes some of Sally’s bundle to Sue, and see whether Sue continues to prefer
Sally’s bundle. If so, then raise the tax. Once again, we subject the new distribution to the Envy
Test and see whether Sue still prefers Sally’s bundle. We will repeat this process until Sue no
longer prefers Sally’s bundle. The final distribution will then pass the Envy Test.

54
Dworkin, "What Is Equality? Part 2: Equality of Resources," 312.
55
Ibid., 313.
46

Dworkin concedes that an income tax is a less-than-perfect device for responding to the
reciprocal influence between ambitions and endowments but points out that at least it is a
compromise between two requirements of equality and not a compromise between equality and
some independent value, such as efficiency.
56
However, whether this is true will depend on the
terms of the income tax and how accurately it can yield a distribution that discounts for the
effects of the natural lottery on the market. What remains to be seen is whether a taxation
scheme driven by the Envy Test, as illustrated above, can achieve the desired level of accuracy.
In what follows, I will argue that attempting to use the Envy Test as a guide to an
adequate tax scheme will fail. I discuss two different reconstructions of Sally and Sue’s
preferences that could serve as justifications of two different tax schemes. Both of these tax
schemes, I will argue, are unsatisfying: the first because it sacrifices equality for the sake of
efficiency, and the second because it leads to an inefficient distribution of resources.
1 The Incentivizing Tax Scheme
In thinking about the redistribution of resources from Sally to Sue, we supposed that we
may arrive at an adequate taxation scheme through use of the Envy Test: tax rates are adequate
when they result in a distribution of resources that passes the Envy Test. When Sue, for
example, prefers Sally’s bundle of resources, an increase of the tax on Sally’s bundle is called for
until Sue no longer prefers it. But at what point exactly will Sue cease to prefer Sally’s bundle of
resources? Let us consider the following two scenarios. In the first, we suppose that Sue will be
satisfied when her bundle is the same as Sally’s. This suggests what I’ll call the Strictly
Equalizing tax scheme. In the second scenario, Sue will be satisfied with her bundle at the point

56
Ibid.
47

at which an additional increase in Sally’s income tax would result in a decrease in Sue’s
bundle.
57
This leads us to what I call the Incentivizing tax scheme.
Because the standard by which Dworkin judges the equality of a distribution is the Envy
Test, whether a post-tax distribution is consistent with equality, then, will depend on the agent’s
preferences. If a post-tax distribution fails the Envy Test, then the tax scheme must be amended
until it will yield a distribution that passes. Of course, in our hypothetical scenario, the
preferences that drive the Envy Test are not the preferences of any actual agents. Rather, they
are preferences that we suppose would be rational for the agents in our thought experiment to
have. In order to find an adequately equalizing tax scheme, then, we will need to reconstruct the
preferences of agents like Sally and Sue. It’s plausible, I think, to reconstruct their preferences
in two ways. The first way I discuss leads to what I call the Incentivizing Tax Scheme.
Suppose that the amount that Sally is taxed affects her preferences regarding her
occupation and level of productivity. For example, let’s say that Sally prefers to be a spear fisher
over a mango farmer because she is more talented at spear fishing and, as a result, can make
more money this way. But if she can make as much money mango farming as she can by spear
fishing, then she would prefer to be a mango farmer. Suppose that we raise her income tax to
40%. If, at this rate, her post-tax earnings will be the same as the earnings she would have as a
mango farmer, then she will decide to become a mango farmer instead. How will this affect
Sue? When Sally’s income tax is 30%, Sue prefers Sally’s bundle because Sally’s contains more
resources, and the only relevant differences between Sally and Sue lie in their native talents.
Because Sue prefers Sally’s bundle, the Envy Test proclaims the distribution to be unequal, and
we will then have to raise the tax on Sally so as to redistribute to Sue. If we raise the tax on
Sally by 1%, then Sue will still prefer Sally’s bundle because even though Sue has a little more

57
It is commonly thought that Rawls’s Difference Principle would result in a tax scheme that operates in this way.
48

than she had before, Sally’s bundle of resources is still greater than Sue’s and, as before, the only
relevant differences between the two lie in their native talents. So we will continue to raise the
tax by these small increments until we propose an income tax rate of 40% on Sally. Recall that
at this rate, Sally will decide to become a mango farmer. Now suppose that only Sally is able to
catch pompano fish, which happen to be Sue’s favorite fish to eat. In fact, Sue loves pompano
fish so much that it is her favorite meal. If Sally decides to become a mango farmer, this will
negatively impact Sue because then her bundle will no longer contain pompano fish. Sue will be
worse off when the tax on Sally is at 40% than when it is 39%.
We have seen how Sally’s preferences may be affected by the income tax rate. Now let
us turn our attention to Sue’s preferences. Compare Sue’s bundle when Sally’s income tax is
39% to Sue’s bundle when Sally’s income tax is raised to 40%. At the income tax rate of 39%,
Sally is still a spear fisher, and Sue’s bundle contains her favorite meal: pompano fish. But at the
rate of 40%, Sally decides to become a mango farmer instead, and Sue’s bundle no longer
contains pompano. Sue, then, is better off when the tax on Sally is 39% than when it is 40%.
Suppose that Sue can foresee that if the tax on Sally is raised to 40%, then Sue will end up with a
worse bundle than she currently has. Now, Sue prefers her current bundle over the one she
would have if Sally decided to become a mango farmer instead. Given this preference, perhaps it
might make sense to say that Sue will be satisfied with her bundle when the tax on Sally is 39%,
even though Sally’s bundle is still greater than Sue’s. If this is a plausible account of Sue’s
preferences, then we can provide a justification for the following tax scheme: income tax is to
allow the better off only the minimum incentive needed to retain their productivity. Let’s call
this the Incentivizing Tax Scheme. Recall that Dworkin’s test for equality in a distribution of
resources is the Envy Test. The Incentivizing Tax Scheme, then, is compatible with equality of
49

resources so long as its resulting distributions pass the Envy test. In our reconstruction of Sally
and Sue’s preferences, we see that Sue may cease to prefer Sally’s bundle of resources even
when Sally’s bundle is greater than hers. This will occur when any additional increase in the tax
on Sally will negatively impact Sue.
We see that we can justify the Incentivizing Tax Scheme only if Sally and Sue’s
preferences take shape in the manner described above. However, I think that we have several
reasons to reject this reconstruction of their preferences. First, one might think that it’s
implausible to think that people can shape or control their preferences based on counterfactual
considerations—for example, considerations regarding the consequences of having a certain
preference. The reconstruction above conflates the preference that Sue should have, if she is to
maximize her interests, and what she actually prefers. There is a sense in which she should not
prefer Sally’s bundle because doing so would result in a redistribution of resources that leaves
her worse off than she is now. It would be in her best interests not to prefer Sally’s bundle. But
even if she acknowledges this fact, does she then cease to actually prefer Sally’s bundle? I think
not. Consider how Sue would respond if she were given the option to “trade places” with Sally.
If she trades places with Sally, then she could enjoy all that Sally enjoys in the current
distribution of resources (while Sally would end up with Sue’s bundle of resources). I see no
reason why Sue would not want to trade places with Sally. After all, we’ve stipulated earlier that
Sue is just as willing to work hard as Sally; the only relevant differences between them lie in
their differential native talents and the advantages that result from the exercise of those talents in
production and trade. The consideration that Sue would wish to trade places with Sally is, I
think, a consideration that we should take to be relevant to our judgment of the equality of the
distribution. Either (a) this consideration shows that Sue does not actually cease to prefer Sally’s
50

bundle of resources, even when she acknowledges that the consequence of having that preference
is contrary to her best interests; or (b) even if Sue ceases to prefer Sally’s bundle after taking into
account this counterfactual consideration, it’s reasonable to suppose that she nonetheless wishes
she could trade places with Sally, indicating that Sue is not satisfied with her lot in the
distribution of resources. We may have reason, then, to pause and reconsider whether the
distribution is truly consistent with equality. For whatever it is that the Envy test was initially
proposed to capture—the idea that when people reasonably desire someone else’s bundle of
resources, their desire is a good indication that the distribution of resources is not equal, for
example—Sue’s wish to switch places with Sally may very well count as the kind of
consideration that the Envy Test was initially formulated to be sensitive to.
If Sue does not actually cease to prefer Sally’s bundle of resources, then the distribution
does not pass the Envy test; hence, the Incentivizing Tax Scheme will lack justification. But on
the supposition that Sue might cease to prefer Sally’s bundle, we may nonetheless regard Sue as
being reasonable in being dissatisfied with her lot and wanting to switch places with Sally.
Indeed, the luck egalitarian may have a ready diagnosis for why Sue may be thought of as
dissatisfied with her lot, even if she no longer “prefers” Sally’s bundle of resources. By insisting
that the distribution of resources should not reflect the unchosen features of people’s
circumstances, the luck egalitarian may view the distributions that result from the Incentivizing
Tax Scheme as reflecting a compromise between equality and efficiency.
Recall that Dworkin maintains that the income tax reflects a compromise between
conflicting demands within the ideal of equality rather than a compromise between equality and
independent values such as that of efficiency. We have seen that Dworkin envisions equality as
giving rise to two desiderata: (a) to hold people responsible for the costs of their plan of life,
51

measured in terms of the costs of that life to others, and (b) to mitigate, if not neutralize, the
disadvantages that result from the unchosen features of people’s circumstances. The auction
theory in conjunction with a taxation scheme is meant to fulfill both these desiderata. The first
auction—which results in a distribution of resources that passes the Envy Test—illustrates how
free markets can be justified as upholding the ideal of equality by satisfying (a). Meanwhile, the
second auction for insurance, along with a subsequent income tax, shows how an overall scheme
of taxation can be applied to the proceedings of the free market to result in a distribution of
resources that satisfies (b).
However, if Dworkin’s Envy Test allows the Incentivizing tax scheme, it will yield
distributions that fail to adequately achieve (b) because it sacrifices its pursuit of this desiderata
in exchange for gains in efficiency. As a result, luck egalitarians will have good reason to reject
the modified Envy Test as an appropriate measure for the justice of a distribution. As stated, (b)
does not require a complete elimination, or neutralization, of the unequalizing effects of the
natural distribution of talents. Indeed, this may be impossible, since many talents are such that
they enable their possessor to take up more attractive occupations, differences in remuneration
aside. But the luck egalitarian will claim that a distributive theory should seek to mitigate the
disadvantages of the unlucky as far as possible. On their view, a theory that fails to do enough
in pursuit of (b) will fail to yield a just distribution.
The Incentivizing tax scheme will result in a compromise between equality and efficiency
when it is reasonable for Sue to continue to prefer Sally’s bundle even when a further
redistribution would reduce Sue’s bundle. After all, Sally and Sue work equally as hard and
have the same ambitions; the only relevant differences between them result from their genetic
luck. Sue could arm herself with the luck egalitarian intuition that the inequality in the
52

distribution is unjust because it reflects differences in the unchosen features of their
circumstances. But while the further redistribution that Sue’s preference for Sally’s bundle
would bring about would work toward equality between Sally and Sue, it is not in Sue’s best
interest to prefer Sally’s bundle when doing so would result in a redistribution of resources that
would leave her with less than she currently has. There is a tension, then, between equality on
the one hand, and efficiency on the other. The Incentivizing scheme, which allows distributions
that reflect the unequal distribution of natural talents, sacrifices equality for the sake of
efficiency.
What’s missing from Dworkin’s picture is the requirement of interpersonal justification
between the members of the economy. If Dworkin indeed wishes to avoid distributions that
compromise equality for the sake of independent values such as efficiency, then he will need to
show how the Envy Test can be suitably responsive to the reasonable preferences people have,
where reasonable preferences are those preferences that can be justified to others as grounding a
claim to certain resources. As a first step toward this aim, we should attend to how we might
modify the Envy Test so that it would disallow distributions of resources under which some
individuals wish to switch places with others.
Suppose, then, that we modify the Envy Test in the following way:
The Positional Test
No division of resources is an equal division if, once the division is complete, any
immigrant would prefer to switch places with another in the distribution of resources.

What I call the Positional Test seems as plausible as the Envy Test. It seems to capture the same
initial motivation for the Envy Test: if people in an initial situation of equality find that they
prefer another’s lot, then this is at least prima facie evidence that the distribution of resources is
unequal. However, the Positional Test is not equivalent to the Envy Test. To see this, note that
53

the Positional Test would not allow a taxation scheme that yields a distribution that reflects
differences in people’s native endowments. For under such distributions, the unlucky would
prefer to switch places with the lucky. Not only do the lucky enjoy the advantages of their good
luck in the natural lottery, they are, as a result, able to enjoy greater resources as a result of their
utilization of these natural talents in production and trade. They are, in a sense, doubly
advantaged by their good genetic luck. Any rational individual, then, would prefer to be in the
situation of the lucky over being in that of the unlucky. Certainly, Sue would prefer Sally’s
place in the distribution of resources, given that they work the same number of hours in the same
occupation.
The aim of my discussion in this section thus far was to show that the Incentivizing Tax
Scheme will only pass the Envy test if we reconstruct Sally and Sue’s preferences under the
assumption that an agent’s preferences can be shaped by counterfactual considerations. I have
suggested that this assumption is implausible and that, as a result, the distributions that result
from the Incentivizing Tax Scheme will not pass the Envy test. But even if this assumption is
correct, we still have reason to reject the Incentivizing Tax Scheme. Even if Sue ceases to prefer
Sally’s bundle of resources, it’s reasonable to think that Sue nonetheless wishes to switch places
with Sally. It is a failure of the Envy test not to be sensitive to Sue’s dissatisfaction with her
bundle. And if we modify the Envy test to correct for this failure, the revised version (that is, the
Positional Test) will no longer allow the distributions that would result from the Incentivizing tax
scheme. In the ensuing discussion, I will see whether the Positional Test will suffice. As we
will see, even if we were to adopt the Positional Test in place of the Envy Test, we would still
find that the test fails to ensure the justice of post-tax distributions.
2 The Strictly Equalizing Tax Scheme
54

Let us now consider whether the strictly equalizing tax scheme is feasible. On this
scheme, income tax is to equalize the bundles of resources amongst individuals of the same
occupation who work for the same number of hours. The aim of the strictly equalizing tax
scheme is to discount the natural advantages enjoyed by the talented and to yield distributions
that reflect only the responsible choices of individuals and not the unchosen features of their
circumstances. This scheme can be shown to be congruent with Dworkin’s interpretation of
equality of resources if we can reconstruct Sally and Sue’s preferences in in such a way that the
distributions that result from the Strictly Equalizing Tax Scheme would pass the Envy test.
It is easy to see how Sally and Sue’s preferences should be reconstructed in order to bring
about the Strictly Equalizing tax scheme. If we suppose that Sue will continue to prefer Sally’s
bundle of resources so long as Sally has more, then Sue will be indifferent between her bundle
and Sally’s only once they are the same. Any distribution of resources under which Sally’s
bundle is greater than Sue’s will not pass the Envy test. Hence, we may propose the Strictly
Equalizing tax scheme, which brings about redistributions of resources so that individuals of the
same occupation who work similar hours enjoy similar bundles of resources.
It is also fairly easy to see how the Strictly Equalizing tax scheme can respond to the
differences in native endowments between two individuals who are otherwise relevantly similar.
Sally and Sue, for example, have the same occupation and work just as hard. But when we
expand our attention to include individuals who hold different occupations, it becomes difficult
to see how the scheme will work. Let’s introduce Tim, who is similarly ambitious to Sally and
Sue but works as a coconut harvester. How should we compare the bundle of resources that Tim
enjoys with those enjoyed by Sally and Sue? The Positional Test would require us to redistribute
if Sally and Sue prefers to switch places with Tim. But how are we to determine whether they
55

wish to do so? While the Strictly Equalizing tax scheme may adequately neutralize the unfair
differences between workers of the same occupation who only differ in their genetic luck, it
encounters difficulties in accounting for the differences between workers of different
occupations.
Moreover, the Positional Test may not prove sufficiently sensitive to the role that luck
plays in the market. This problem can be seen by the following thought experiment. Suppose
that all of the islanders, with the exception of Tim, suffer from an extreme fear of heights. They
are also crazy for coconuts. Given the high demand for coconuts, Tim decides to harvest
coconuts and charges a very steep price for his products. As a result, he enjoys a bundle of
resources that is much greater than others’ and can afford many luxury goods. Meanwhile, the
supply of mango farmers is high, and each mango farmer earns a minimum subsistence wage.
Clearly, this distribution would pass Dworkin’s Envy Test because, given their extreme fear of
heights, no one would prefer Tim’s bundle, since it involves an occupation that requires one to
climb very tall trees. Would the distribution, however, pass the Positional Test? The Positional
Test, you will recall, rejects distributions under which any individual would prefer to switch
places with another individual in the distribution of resources. Tim’s “location” in the
distribution of resources is determined by the stock of goods contained in his bundle of resources
as well as his occupation. Because the islanders have an extreme fear of heights, none of the
mango farmers would prefer to switch places with Tim because doing so would involve taking
up his tree-climbing occupation. The Positional Test then would allow Tim to enjoy a far
greater bundle of resources than the other islanders.
This will be a problem for the Positional Test if we can provide some motivation for
thinking that this distribution is intuitively unjust. The luck egalitarian may step in and claim
56

that although Tim’s wealth is in part a result of his choice to take up the occupation of coconut
picking and to supply a product that is highly desired by others, a good deal of his wealth is the
result of his good fortune in being the only non-acrophobic islander. His wealth, in other words,
results from unchosen features of his circumstances. While this is normally taken to show that
justice demands greater compensation to the unlucky person, in this case we find reason to deny
a lucky individual his undeserved advantages. If the stock of goods were plentiful and abundant,
we may have no reason to do this; however, assuming moderate scarcity, the undeserved wealth
that Tim enjoys comes at a cost to others. Hence, the luck egalitarian will find this distribution
to be intuitively unjust and the Positional Test inadequate in failing to disallow it.
Given the troubles facing the Positional Test, my opponent may wish to show that
Dworkin’s theory can nonetheless propose the Strictly Equalizing tax scheme by reverting to the
Envy test. After all, the Envy test was shown to be inadequate only under the assumption that
Sue’s preferences will be shaped by counterfactual considerations. But we are making no such
assumption here.
Again, however, I think that the Envy test can be shown to be inadequate as a guide to a
proper income tax scheme. Recall that Sally’s preferences might change as further resources are
distributed away from her. At some point, she may find that in the absence of the extra post-tax
income that she would otherwise have received, she prefers to be a mango farmer rather than a
spear fisher. The Strictly Equalizing tax income would require that resources are distributed
from Sally to Sue so long as Sally has more than Sue. This would result in distributions in which
Sally decides she no longer has incentive to remain in her profession. Consequently, Sue will
find herself worse off. For while Sue was able to enjoy pompano fish prior to the imposition of
the Strictly Equalizing tax, her bundle of resources under the post-tax distribution lacks the fish.
57

Because neither Sally nor Sue is made better off by the tax, we may reject the Strictly Equalizing
tax scheme on grounds of inefficiency.
At this point, a defender of the Auction Theory could point out that the Envy Test is not
meant to be applied to post-tax distributions in the first place. Rather, it is important to maintain
a distinction between the two stages of the theory: in the first stage, or the ideal stage, we clarify
the content of the ideal of equality through the various hypothetical auctions in which the Envy
Test plays a role; in the second, we consider how this theory of equality can have practical
application through the implementation of a tax system that best approximates the hypothetical
auctions. There is no income tax in our hypothetical scenario: the survivors merely have an
initial auction to start off with an equal distribution of resources, and the tendency toward
unequal distributions brought about by the commencement of production and trade is countered
by an additional auction for insurance against under and un-employment. My argument, which
purports to show that the Envy Test does not secure the justice of post-tax distributions, misses
the mark; the Auction Theory never claims that post-tax distributions should pass the Envy Test
because there simply is no taxation in the hypothetical scenarios of the first, ideal stage of the
theory.
In reply, I concede that Dworkin only mentions taxation as part of his discussion of the
second stage of his theory in which we are shown how the hypothetical island scenario can be
made relevant to us in the real world. Moreover, he only mentions the Envy Test in relation to
the auctions of the first stage of his theory. However, these considerations alone should not lead
us to conclude that the Envy Test has no place in our judgments of the justice of post tax
distributions in the second stage of the theory. The most pressing question when we move from
the first stage to the second stage of the theory is: How well does the system of taxation
58

approximate the hypothetical auctions of the ideal theory? Of course, not all schemes of taxation
will be just. Only those that are the best approximation of the hypothetical auctions will be just.
To determine this, we need to see what rates of un- or under-employment insurance people in the
first stage would be willing to purchase.
But how exactly shall we, in the real world, fill in the missing information about the
insurance that hypothetical agents would purchase? It seems to me that the only relevant data we
can use on which to base such speculations would be the behavior of economic agents under
various systems of taxation. But this gets things backwards. Real world market conditions are
significantly unlike the ideal conditions of the hypothetical auction. This is one reason why we
take the hypothetical auction as a model against which to judge actual market outcomes.
One way to test whether a tax scheme is a good approximation of the hypothetical
insurance auction is to consider whether that tax scheme would produce distributions of
resources that conserve the initial equality of resources set in place by the first auction for
resources. After all, the Envy Test is to be applied diachronically—that is, to measure whether
people were satisfied with their bundles, in comparison to others, over a lifetime. If the islanders
are to enjoy equal resources over a lifetime, then the proceedings of the auction for under- and
un-employment insurance will also be subject to the Envy Test, given its impact on the bundle of
resources that individuals will have. The Envy Test, then, plays a fundamental role in the first
stage of the theory: it is the criterion by which the equality of distributions are judged. Given its
important role in the first stage of the theory, I see no reason why it wouldn’t continue to play a
guiding role in the second stage of the theory.
58
We can, then, reject the post-tax distributions of

58
To clarify, the second stage of the theory is still part of Dworkin’s ideal theory, albeit the part of the theory that
specifies the ideal application of the first stage of the theory. That is to say, the second stage identifies an ideal
system of taxation against which we should judge actual systems of taxation. One way of

distinguishing the two
stages would be to identify the first as the ideal ideal stage and the second as the practical ideal stage of the theory.
59

the second stage if they are inconsistent with the equality of resources specified by the first stage
of the theory. And such distributions will be inconsistent with equality if they fail the Envy
Test.
59

IV AN ALTERNATIVE: THE BURDENS COMPENSATION PRINCIPLE
By pointing out the difficulty of arriving at a distribution that aligns with the luck
egalitarian intuition, I hope to motivate an alternative theory. This section will explore the
advantages of what I call the Burdens Compensation Principle, which requires that the burdens
of socially necessary labor be compensated. Like other general egalitarian principles, we begin
with equality as a default starting position. We then allow deviations from equality only if they
result from compensation for the special burdens of socially necessary labor.
We may think of burdens as, roughly, the features of an occupation that make it
unattractive. Occupations that require long and erratic hours, for example, or place their workers
in hazardous environments are burdensome. Other burdens include health risks, such as the risk
of lung cancer born by coal miners or the long distance truck driver’s risk of developing
diabetes; physical labor and exertion, such as the heavy lifting required of delivery personnel or
construction workers; the risk of injury or death involved in the work of roofers, deep-sea
fishers, and law enforcement agents, among others; exposure to noxious or unpleasant elements,
which sanitation workers endure; and many other such burdens, including the tedium and
boredom endured by factory line workers.
The Burdens Compensation Principle can treat the various cases we’ve thus far explored
in an intuitively appealing way. Consider, first, the case of Sally and Sue. Luck egalitarians
object to a difference in the resources of the two on the grounds that the only relevant difference

59
At least, this is what we should expect, given the rest of Dworkin’s theory. Of course, I argue against the Envy
Test as a sufficient guide for equality, but my argument here is that Dworkin’s theory does not supply us any reason
to reject the Envy Test as applicable to the post-tax distributions of the second stage of his theory.
60

between them is the difference in their genetic luck. But suppose that because Sally catches
more fish throughout the day, she engages in more work than Sue. Say that Sally scales and
filets the fish she catches. Sue also scales and filets the fish she catches, but because she is
unable to catch as many fish as Sally, she scales and filets far fewer fishes. Intuitively, Sally
deserves more remuneration than Sue solely in virtue of the extra work she engages in. The luck
egalitarian may wish to accommodate this intuition, but she will have to trace the extra
remuneration that Sally deserves back to some aspect of her choices. Yet it is difficult to
extricate the choices she makes from the background circumstance that enable her to make those
choices. It is because Sally is so talented at catching fish that she has the opportunity to engage
in more work by processing them. The luck egalitarian, then, will have difficulty in explaining
how we can justify providing greater remuneration to Sally, who is more productive partly in
virtue of her greater talents.
Rather than appealing to the distinction between luck and choice, the Burdens
Compensation Principle considers only whether there is a difference in the burdens born by Sally
and Sue. The two will enjoy incomes that are similar but not the same. They will be
remunerated a similar amount because they both work twelve hours a day spear fishing.
However, given that Sally catches and processes more fishes than Sue, Sally will be remunerated
slightly more to compensate for the extra burdens involved in her greater production. Consider,
for example, that in catching and processing more fish, Sally exerts more physical energy
throughout the day and engages in a greater number of repetitive motions that may increase her
chances of developing joint pain. Although their income is not strictly equal, Sally and Sue still
enjoy equality of resources, all things considered, if the extra income that Sally receives is
merely compensation for the extra burdens she undertakes as part of her greater production.
61

Consider, in addition, how the Burdens Compensation Principle will handle Tim’s
income. While Dworkin’s Envy Test will allow Tim to charge others a steep price for coconuts,
luck egalitarians will denounce the distribution as unjust because it allows Tim to enjoy
advantages that result from unchosen features of his circumstances. But their view does not
provide any more concrete guide to determining an income that is appropriate for Tim. Are there
any morally relevant reasons for why Tim should nonetheless enjoy greater income than the
mango farmers? Perhaps coconut harvesting involves greater risk of injury than mango
farming—in particular, coconut harvesters undertake greater risks of falling and significant
injury. If Tim were to require remuneration that is greater (though not exorbitantly so) than the
mango farmer’s, he would not be unreasonable in doing so. This demand can be justified to the
mango farmer; indeed, the mango farmer would recognize that although Tim earns a slightly
higher income, they enjoy equality of resources, all things considered. The difference between
Tim’s income and the mango farmers reflects only the compensation reasonably demanded by
Tim for the greater risks he undertakes in supplying a product that others want.
V CONCLUSION
My aim in this chapter was to show why luck egalitarians should find Dworkin’s
interpretation of equality of resources problematic. In particular, the theory lacks the proper
apparatus for providing us an account of how we can arrive at distributions that do not reflect the
unchosen features of people’s circumstances. On Dworkin’s view, we are to view the imposition
of an income tax as a proper mechanism for yielding distributions that are sensitive to people’s
preferences and ambitions but insensitive to their native endowments. While he acknowledges
that the income tax is an imperfect mechanism for this task, he nonetheless maintains that it at
least reflects a compromise between two desiderata of equality rather than a compromise
62

between equality and other external values, such as efficiency. I then suggested two plausible
income tax schemes: the Incentivizing tax scheme and the Strictly Equalizing tax Scheme.
Neither of these tax schemes, I argued, yields distributions that are consistent with equality.
First, the Incentivizing tax scheme can be justified only if it can be shown to yield distributions
that pass the Envy test. But this is possible only if we reconstruct Sally and Sue’s preferences in
a problematic way. Even if we allow this problematic reconstruction of Sally and Sue’s
preferences, we are led to the consideration that the Envy test fails to capture the morally
relevant observation that even if an agent may cease to prefer another’s bundle on the
recognition that doing so would be contrary to that agent’s best interests, the agent may
nonetheless wish to switch places with another and so may be said to be dissatisfied with her lot
in the distribution of resources. But if we modify the Envy test to capture this observation—in
other words, if we reject the Envy test in favor of what I called the Positional Test—the
distributions that would result from the Incentivizing Tax scheme will no longer pass our
preferred test for the equality of distributions. Second, I argued that what I called the Strictly
Equalizing tax scheme may be rejected on the grounds that it yields inefficient distributions of
resources. While the charge of inefficiency is not enough, I believe, to justify a rejection of a
distribution, the distributions that would result from the Strictly Equalizing tax scheme are not
only inefficient but fail to secure gains toward equality. Given the difficulties that arise from our
attempt to specify an adequate tax scheme under Dworkin’s theory, I suggested in the last section
of this chapter that the luck egalitarian has good reason to embrace the Burdens Compensation
Principle as an attractive alternative principle for governing the distribution of income, because it
better accommodates the hypothetical cases that proved problematic for the Auction Theory.

63

3 Reciprocity-Based Egalitarianism and Incentives

The luck egalitarian conception of justice—according to which socioeconomic
inequalities resulting from bad brute luck are unjust—was shown in the last chapter to be
inadequately realized by a form of market economy in which taxes on income and wealth are
modeled after a hypothetical auction for unemployment insurance. So while roughly luck
egalitarian in motive, Dworkin’s conception of distributive equality, as modeled by the ideal
auction, seems to allow distributions that fail to adequately correct (dis)advantages resulting
from brute luck. In this chapter, I turn to a discussion of the luck egalitarian critique of
distributions governed by maximin principles. In particular, the focus of my discussion will
center on G.A. Cohen’s famous criticism of Rawls’s Difference Principle, which extends into a
deeper critique of Rawlsian political liberalism.
60
While I do not think that Cohen succeeds in
offering an internal criticism of Rawls’s theory, his objections uncover an important external
problem for the Difference Principle: it does not accommodate the intuitive idea that the special
burdens undertaken by individuals engaged in socially necessary labor are relevant, from a moral
point of view, to the distribution of income and wealth. By casting this problem as an external
criticism, I aim to provide strong reason for turning to an alternative principle of distributive
equality without relying on the strong ethos that Cohen thinks Rawlsian citizens should be
committed to.

60
Cohen had developed and maintained this criticism throughout a series of works over a twenty year period. See
Gerald A. Cohen, "Equality of What? On Welfare, Goods and Capabilities," Recherches Économiques de
Louvain/Louvain Economic Review, (1990); Gerald A. Cohen, "Where the Action Is: On the Site of Distributive
Justice," Philosophy & Public Affairs 26, no. 1 (1997); Cohen, "On the Currency of Egalitarian Justice."; Gerald A.
Cohen, "Incentives, Inequality, and Community," The Tanner lectures on human values 13, (1992); Gerald A.
Cohen, Rescuing Justice and Equality (Harvard University Press, 2009); Gerald A. Cohen, "Facts and Principles,"
Philosophy & Public Affairs 31, no. 3 (2003); Gerald A. Cohen, "Justice, Incentives, and Selfishness," in If You're
an Egalitarian, How Come You're So Rich?(Cambridge: Harvard University Press, 2000).
64

Rawls’s general conception of justice reflects the fundamental assumption that all
individuals are free and equal moral persons and that, accordingly, any deviation from equality
must be justified in a way that is compatible with that fundamental freedom and equality. Hence,
equality of social primary goods is taken to be a benchmark against which improvements are to
be judged by free and equal, rational and reasonable persons who are concerned to engage in
social cooperation under fair terms of reciprocity and mutual advantage.
61
While some social
primary goods—such as rights, liberties, and opportunities—are such that no improvements can
be made over an equal distribution, the Difference Principle moves away from strict equality of
income and wealth in favor of a pareto-superior state of inequality in which everyone is better off
than they would be were income and wealth to be distributed equally.
62
This move from equality
to inequality is defended on the grounds that parties to the original position—a hypothetical
modeling device in which representatives choose principles of justice for trustees in the absence
of knowledge of their vested interests—would prefer and agree to the inequality on the grounds
that all inequalities redound to the greatest benefit of the least advantaged.
Despite its pronounced concern for the least advantaged, however, one may worry that
the Difference Principle makes a problematic concession to efficiency by taking talents—
hitherto regarded as arbitrary from a moral point of view to the distribution of income and
wealth—to be a strong determinant in the outcomes that are permitted by the principle. On the
standard interpretation of the Difference Principle, economic incentives serve to guide the
talented to the occupations for which their talents yield the most social benefit and to encourage
productivity and innovation, thereby benefiting all including the worst-off. Although the worst-

61
Rawls, A Theory of Justice, 62.
62
Rawls’s final formulation of the Difference Principle is: “Social and economic inequalities are to be arranged so
that they are... to the greatest benefit of the least advantaged, consistent with the just savings principle,” ibid., 302.
65

off must benefit from incentives-generated inequalities, the luck egalitarian may nonetheless
regard the resulting distribution in which talented individuals often enjoy greater social and
economic advantages as a result of the good brute luck they enjoy as problematic.
Indeed, such luck egalitarian considerations underlie Cohen’s charge that the Difference
Principle surreptitiously substitutes the competing value of efficiency for equality, to the
detriment of justice.
63
He concedes that if incentives were necessary to maximally raise the
levels of the worst off, then those incentives would not be unjust. However, he contends,
incentives are not truly necessary. On his view, the standard interpretation of the Difference
Principle allows incentives to the talented by adopting an overly permissive understanding of
necessity. Incentives are not regarded as necessary on the grounds that the talented are literally
incapable of working hard in their absence; rather, incentives-generated inequalities are thought
to be necessary in order to shape individuals’ occupational preferences in such a way that would
maximally benefit the worst off. In other words, Cohen argues, incentives are only necessary in
the sense that acquisitive maximizers who “lack a certain sort of commitment to equality and are
set to act accordingly” make incentives necessary through their unwillingness to work
otherwise.
64
If Rawlsian citizens truly affirm the Difference Principle, he argues, they would not
require incentives for productive work and innovation, contrary to the standard interpretation of
the principle. Rather, talented individuals would align their occupational choices and behavior
according to the Difference Principle, forgoing incentives for their productivity and innovation.
As a result, Cohen argues, a truly just society will result in “(virtually) unqualified equality” with
everyone at a level that is better off than the level of the worst off under the standard

63
Cohen, "Where the Action Is: On the Site of Distributive Justice."
64
Cohen, Rescuing Justice and Equality 69.
66

interpretation of the Difference Principle.
65
Societies in which incentives-generated inequalities
are permitted are not fully just, as such incentives are only necessary when acquisitive
individuals make them so through their unwillingness to produce without those incentives.
To support his claim that truly just individuals would not require incentives for the use of
their talents, Cohen launches several arguments. The first argument attempts to demonstrate that
the Incentives Argument is unacceptable in a justificatory community—that is, a society in
which members can justify the policies they enact to one another. The second argument points
to an ambiguity in Rawls’s specification of the basic structure of society. Attempting to resolve
this ambiguity will lead to a dilemma: either the basic structure consists only in the coercive
institutions of society or it includes institutions that generally have a profound and pervasive
effect on people’s lives. Cohen argues that the former specification of the basic structure is
arbitrary while the latter will admit that individual choices and behavior should also be
considered part of the primary subject of justice. And if individual choices and behavior are part
of the basic structure, then they will be regulated by the Difference Principle, resulting in a
society in which truly just, talented individuals would not require incentives for their
productivity.
Although I agree with Cohen that a proper understanding of the Difference Principle (or
of any other plausibly just distributive principle) should not endorse distributions under which
“high-flying marketeers” and “acquisitive self-maximizers” enjoy far higher levels of income
and wealth than others, I do not think that, in order to prevent the Difference Principle from
allowing such distributions, we must be committed to a comprehensive view of justice according
to which individuals are bound by a strong ethos that regulates their lives according to the same

65
Cohen, "Where the Action Is: On the Site of Distributive Justice," 6.
67

principles of justice that regulate the major institutions of their society. My aim in this chapter is
to show that we can find reason to prefer an alternative egalitarian distributive principle—
namely, the Burdens Compensation Principle—even while rejecting the view that both
individuals and institutions as answerable to the very same principles and requirements of
justice.
I. INCENTIVES AND THE JUSTIFICATORY COMMUNITY
Cohen’s objections to the standard interpretation of the Difference Principle reflect a
deeper disagreement over the extent to which facts about the world or the agents in it constitute
constraints on an ideal theory of justice. The content of the ideal of justice, he argues, is
independent of psychological propensities and limitations.
66
The ideal of justice should not be
constructed on a foundation of individual choices and behavior regarded as fixed; rather, Cohen
argues, the ideal of justice issues forth pronouncements on just conduct to which individuals are
expected to conform. On the standard interpretation of the Difference Principle, inequalities are
thought to be necessary relative to individuals’ intentions. In other words, incentives serve to
shape people’s occupational preference structures in such a way that would most benefit the least
advantaged. For example, if Sally prefers sculpting over performing surgeries but could most
benefit the least advantaged by being a surgeon, incentives are thought to be necessary to induce
the desired change in her preference ordering: while she may prefer to sculpt when sculpting and
performing surgery are equally remunerated, she may instead prefer to perform surgery when
doing so would provide her greater remuneration than would sculpting. In Rawls’s system,
freedom of occupation is secured by the lexically prior first principle of justice and hence cannot

66
This is in opposition to Hume’s classic discussion of the circumstances of justice; see Book III, Part II, Section ii
of David Hume, A Treatise of Human Nature, ed. John Cottingham, Oxford Philosophical Texts (New York: Oxford
University Press, 2000).
68

be sacrificed for any gain in material equality.
67
Hence, it is impermissible to coerce Sally into
the medical profession. However, society may instead offer Sally the greater remuneration that
would attract her to the medical profession.
For Cohen, however, greater remuneration would not be needed to attract Sally to the
medical profession if her occupational preferences were guided by an overriding desire to benefit
the least talented.
68
For example, suppose Sally knows that she would contribute more to the
social good as a surgeon rather than a sculptor. If we also suppose that Sally prioritizes her aim
to maximally benefit society by producing needed social goods over her other aims—such as
securing a job that she enjoys—then we can suppose that her preference to take the job in which
she would contribute most to the social good has greater priority or weight to her than her
preference for jobs she enjoys. Knowing that she would contribute more to the social good as a
surgeon than as a sculptor, Sally would then prefer to be a surgeon rather than a sculptor.
Consequently, no material incentives would be needed to shape her preferences and attract her to
the medical profession, since her preferences are already guided by her desire to work toward the
social good.
As a result, the standard interpretation of the Difference Principle—which permits
incentives-generated inequalities—fails to acknowledge that incentives are unnecessary in a
society in which individuals act from an ethos of justice that structures their preferences

67
Throughout his criticism of Rawls’s Difference Principle, it is not clear that Cohen also accepts the lexical priority
of the equal basic liberties. Cohen’s criticism of Rawls, however, is intended to be internal. The point here, for
Cohen, is that given freedom of occupation and what Cohen would consider Sally’s “unjust” preference ordering,
the only permissible means of obtaining an outcome in which Sally chooses to perform surgery would be to
incentivize her, through greater remuneration, to take up the occupation.
68
Because Cohen targets incentives to the talented, he identifies the least advantaged as the least talented. He notes
that this is not to imply that in actual systems, those with greater income and wealth truly are more talented or that
those at the bottom economic stratum really are lacking in talents, but rather that, on the standard interpretation,
inequalities will arise from incentives to the talented for their productivity and innovation, resulting in the most
advantaged group being comprised of those with the greatest marketable talents. One might object, however, that
this is an entirely un-Rawlsian understanding of the most and least advantaged; certainly, Rawls never explicitly
identified the two groups in that way but rather in economic terms, for example with respect to median income.
69

according to the aim of the principles of justice. The standard interpretation ignores the effects
of any such ethos and instead regards individual’s occupational preferences as fixed. Since
individual occupational preferences, in the absence of an ethos of justice, are unlikely to form a
pattern that most benefits the worst off, the standard interpretation deems incentives necessary
for restructuring those occupational preferences in order to produce the right pattern. According
to Cohen, the standard interpretation takes arguments such as the following to be typical forms of
justification for policies that allow the talented to enjoy greater income and wealth:
The Incentives Argument
Economic inequalities are justified when they make the worst off people materially better off.
[Major, normative premise]
When the top tax rate is 40 percent, (a) the talented rich produce more than they do when it is
60 percent, and (b) the worst off are, as a result, materially better off. [Minor, factual
premise]
Therefore, the top tax should not be raised from 40 percent to 60 percent.
69

The incentives argument supports inequalities generated by incentives for the talented to produce
and innovate on the grounds that disallowing such incentives would not only be inefficient, but
more importantly, would result in an inferior absolute level of primary goods for the worst off.
This argument is typically illustrated by hypothesizing what kind of behavior individuals would
have under different systems of progressive taxation. For example, suppose a surgeon who
currently pays 40% of her income in tax might decide to choose a less stressful career if her tax
rate were raised to 60%. If she were to change careers, society would lose out on the
contribution of her valuable skills to the quality of healthcare in addition to whatever difference
in tax payments there is between her former occupation and her new, less well-paying job. The
result is a decrease in the funds available to support government welfare programs, healthcare,
education, public transportation and other publicly provided goods available to the least
advantaged.

69
Cohen, Rescuing Justice and Equality 34.
70

While Cohen does not contest the truth of the premises of the Incentives Argument, he
nonetheless points out that when it comes to arguments regarding public policies, truth is not the
only evaluative norm by which we judge the argument. The justifiability of the argument,
especially to those least advantaged by the proposed policy, is also be an important factor in our
judgments concerning the strength of the argument in favor of the policy. Such policies are
typically constituted of a mixture of factual and normative premises. Even when we agree with
the normative premises of the argument and concede the truth of its factual premises (as well as
recognize the validity of the argument), we might nonetheless come to find the argument as a
whole unacceptable. What we take issue with is not the truth of the premises but rather the
reason-givingness of the factual premises. The strength of the reason-givingness of a factual
premise can be more easily discerned when the argument is offered in interpersonal contexts.
To illustrate, consider the Incentives Argument in the first-person mode, offered by a talented
individual to an untalented one:
The Talented’s Argument
Economic inequalities are justified when they make the worst off people materially better
off. [Major, normative premise]
When the top tax rate is 40 percent, (a) I produce more than I would if the top tax rate
were 60%, and (b) you are, as a result, materially better off. [Minor, factual premise]
Therefore, the top tax should not be raised from 40 percent to 60 percent.

In the third-person mode, the speaker offering the Incentives Argument is merely offering a
valid, and perhaps even sound, argument. Her identity as the speaker of the argument and the
identity of her interlocutor as the addressee have no impact on our intuitive judgment of whether
the argument is reasonably acceptable. In the first person mode, however, the speaker offers an
argument that prompts her interlocutor to demand a justification for the speaker’s role in making
the minor, factual premise true. Intuitively, who is offering the argument and to whom it is
offered impact the justifiability of the policy recommended by the argument.
71

Cohen takes no issue with the normative premise of the Incentives Argument. Rather, his
rejection of the argument as a whole results from his denial that the minor, factual premise
provides sufficient reason for favoring the proposed policy. The factual premise, he argues, is
true only when the talented rich fail to prioritize the aims and underlying ideals of the Difference
Principle. Were the talented rich willing to produce and innovate even in the absence of
incentives, part (a) of the minor premise would be false.
To further illustrate, Cohen likens the relation between the Incentives Argument and the
Talented’s Argument, on the one hand, to the relation between the following Ransom Argument
and Kidnapper’s Argument:
The Ransom Argument
Children should be with their parents. [Major, normative premise]
Unless the kidnappers are paid, the children will not be returned to their parents. [Minor,
factual premise]
So the kidnappers should be paid.

The Kidnapper’s Argument
Children should be with their parents. [Major, normative premise]
Unless you pay me, I shall not return your child. [Minor, factual premise]
So you should pay me.
70

In its third-person mode, the Ransom Argument does not implicate the speaker as playing a role
in making the minor, factual premise true. Subsequently, no demand is made on the speaker to
justify her actions, for she is merely offering a sound argument based on the projected behavior
of some third party—namely, the kidnappers. However, the first-person mode of the
Kidnapper’s Argument does implicate the speaker as playing a role in making the minor, factual
premise true, and the speaker’s interlocutor may reasonably demand that the speaker justify that
role. Although the addressee recognizes that the Kidnapper’s Argument, like the Ransom
Argument, is a sound argument, she nonetheless holds the argument to an additional evaluative

70
Ibid., 39. The titles of these arguments (i.e. “The Ransom Argument” and “The Kidnapper’s Argument”) are my
additions.
72

standard—in particular, a standard of interpersonal justification—that did not apply to the
Ransom Argument.
Analogously, Cohen argues, the first-person mode of the Talented’s Argument prompts
the speaker’s interlocutor to demand a justification for the talented’s involvement in making the
factual premise true. Just as the kidnapper will fail to provide a satisfactory justification for her
role in making the factual premise of the Kidnapper’s Argument true, so too is the talented
unable to justify her the factual premise in the argument she offers to the least advantaged
members of her community.
The Incentives Argument, then, fails to pass what Cohen calls the Interpersonal Test.
This test is to provide a method for determining whether an argument provides justification for a
particular policy:
The Interpersonal Test
If, because of who is presenting it, and/or to whom it is presented, the argument cannot
serve as a justification of the policy, then whether or not it passes as such under other
dialogical conditions, it fails (tout court) to provide a comprehensive justification of the
policy.
71

The test draws a distinction between mere justification and a comprehensive justification. An
argument in favor of a particular policy that relies on facts about the projected behavior of the
members of a society fails to serve as a comprehensive justification for that policy if the
projected behavior of those members itself is not justified.
72
To illustrate, the Ransom
Argument offers a sound argument that recommends a policy of paying the ransom to the
kidnapper. Nonetheless, the Ransom Argument fails to pass the Interpersonal Test because it
fails to offer a comprehensive justification for paying the ransom; the kidnapper who makes the
minor, factual premise of the argument true is unjustified in doing so. The factual premise of the

71
Ibid., 42.
72
Ibid., 41.
73

argument, then, fails to provide a strong enough reason to support the conclusion of the
argument—namely, that the kidnapper should be paid. This is apparent when she offers her
interlocutor the Kidnapper’s argument, in which the first-person mode lays bare her involvement
in making the factual premise of the Ransom Argument true. Likewise, Cohen reasons, the
Incentives Argument fails to pass the Interpersonal Test because, as the Talented’s Argument
reveals, the talented are not justified in their involvement in making the minor, factual premise
true.
The distinction between mere justification and comprehensive justification and its role in
the interpersonal test does not explicitly appear in Rawls’s theory. However, Cohen argues,
insofar as Rawls is concerned with supplying “a public basis in the light of which citizens can
justify to one another their common institutions,” it does not seem problematic to expect that
Rawlsian citizens seek comprehensive justifications for their policies.
73
Members of a
democracy mutually respect each other’s fundamental freedom and dignity by forming policies
jointly, and, as Cohen argues, “we do not make policy together if we make it in the light of what
some of us do that cannot be justified to others.”
74
At the heart of the democratic community lies
the willingness of its members to be open to requests for justification. In other words, fellow
members of a democracy owe each other justification for the policies that they enact, if so
requested. Moreover, such justifications must be reasonably acceptable when offered from one
member of the community to another: it is not enough that an argument for a policy is sound; it
must also be reasonably acceptable to all.
75

73
Ibid., 45. Quotation from John Rawls, "Kantian Constructivism in Moral Theory," The Journal of Philosophy 77,
no. 9 (1980): 561.
74
Cohen, Rescuing Justice and Equality 45.
75
I am speaking loosely here; if the reader so prefers, she may hold that an argument is not justifiable unless it is
such that it cannot be reasonably rejected by anyone.
74

II. REASONABLE UNWILLINGNESS
As we’ve seen in the previous section, Cohen rejects the lax reading of the Difference
Principle on the grounds that we should not consider inequalities truly necessary unless the
talented are literally unable—as opposed to unwilling—to work in the absence of those
incentives. If they are literally unable to work without such incentives, then the Incentives
argument passes the Interpersonal Test, for we cannot hold individuals to a duty that is
impossible for them to fill. However, Cohen maintains that when incentives are necessary only
because the talented are unwilling to work without them, the Incentives Argument does not pass
the Interpersonal Test. As the analogy between the Talented’s and Kidnapper’s Argument
purports to make clear, the talented are not themselves justified in making the factual premise of
the Incentives Argument true.
However, in this section I will argue that the analogy between the Kidnapper’s Argument
and the Talented’s Argument is weak. Unless the reasons for why the talented are unwilling to
work without incentives are specified, we will lack a sufficient basis for condemning their role in
making the factual premise of the Incentives Argument true. While it is immediately clear that
the kidnapper lacks justification in offering the Kidnapper’s Argument, it is not immediately
clear that the talented lacks justification in offering the Talented’s Argument. Hence, while the
Kidnapper’s argument reveals why the Ransom Argument fails to offer a comprehensive
justification for a policy of paying ransom to kidnappers, no such seamless transition holds
between the Talented’s Argument and the Incentives Argument.
Not all reasons for being unwilling to work without incentives are unjustifiable. Granted,
so-called “high-flying marketeers” are unjustified in their demand for incentives if they wish to
accrue exorbitant wealth for its own sake or, for example, to feel superior to others. But, to be
75

fair, not all of the talented’s demands for incentives are of this nature. It may even be helpful to
distinguish between unreasonable and reasonable unwillingness. Consider, for example, the
difference between the marketeer and the surgeon: while both talents fetch a higher-than-average
income on the market, the surgeon employs her talents in an occupation that also requires her to
undertake burdens that provide intuitively compelling reasons for giving her the extra pay. Her
work requires prolonged hours of sustained attention that, in many cases, carries the additional
stress of bearing the weight of responsibility for the health and safety of the patient. It is difficult
to deny that there is a morally significant difference between the high-flying marketeer’s claim
for incentives and that of the surgeon.
To elaborate, consider the following argument, which is not only structurally identical to
the Talented’s Argument but can even be considered an instance of it:
The Nurse’s Argument
Economic inequalities are justified when they make the worst off people materially better
off. [Major, normative premise]
When the top tax rate is 40 percent, (a) I am willing to work the night shift (but not
willing if the tax rate is raised to 60 percent), and (b) you are, as a result, materially better
off. [Minor, factual premise]
Therefore, the tax should not be raised from 40 percent to 60 percent.

Is the nurse justified in making the factual premise of her argument true? This will depend on
her reasons for being unwilling to work the night shift when taxes are raised. It is well known
that working the night shift, compared to engaging in the same work during the day, carries
various health risks. The disruption of the worker’s circadian rhythm can lead to increased
chances of heart disease, diabetes, and depression and can make the worker more prone to
accidents. Moreover, in working the night shift, the nurse may undertake great sacrifices in her
familial life, as her work occupies most of the hours during which her children are at home.
Suppose these considerations constitute the reasons why the nurse would be unwilling to work
76

the night shift if she were to take home the no more pay than she would get working the day shift
instead. It seems that such reasons do support a reasonable demand for greater pay. In this case,
the nurse’s unwillingness to work in the absence of incentives does not incriminate her as
possessing an acquisitive, self-maximizing attitude that is inconsistent with a proper sense of
justice. Rather, her unwillingness may be seen as reasonable and justifiable in light of the
special burdens she undertakes while performing socially necessary labor.
Presumably, Cohen himself would agree that the Nurse’s Argument passes the
Interpersonal Test. However, the aim of my discussion here is to show that it is not at all clear
that the Incentives Argument fails to pass the Interpersonal Test, for the analogy between the
Talented’s Argument and the Kidnapper’s Argument is no stronger than the analogy between the
former argument and the Nurse’s. Indeed, it’s plausible to regard the Nurse’s Argument as one
instance of the Talented’s Argument. Hence, even if we take the “lax” reading of the Difference
Principle—that is to say, even if we accept necessities that are relative to individual’s
intentions—we are not committed, across the board, to accepting inequalities that are
unjustifiable to the least advantaged of society.
Nonetheless, this leaves open the possibility that, unlike the Nurse’s Argument, other
instances of the Incentives Argument may uncover the speaker’s unjustifiable attitudes. The real
target of Cohen’s critique, one might object, is the talented whose demand for extra pay is based
on reasons that her interlocutor may reasonably reject.
However, Cohen does not investigate the kinds of reasons that members of a well-ordered
society of justice as fairness would reject as unjustifiable. Further, the lexical priority of both the
equal political liberties, with a guarantee of their fair value, and fair equality of opportunity
would disallow vast disparities in income and wealth that would undermine these goods. So, the
77

“high-flying marketeers” that Cohen condemns, then, might not be flying so high after all, given
the constraints set by lexically prior principles, whatever their reasons for requiring that pay.
III. INTERPERSONAL JUSTIFICATION IN THE BACKGROUND CULTURE
We have seen that the general strategy that Cohen employs in his demonstration that the
Incentives Argument would not pass the Interpersonal Test is this: if an argument for some
policy can reasonably be rejected by any individual when offered in the second-person mode,
then it fails to offer a comprehensive justification for that policy. One’s failure in offering a
justification to one’s interlocutor when posing an argument in the second-person mode is
sufficient for the argument’s failure to justify a policy in the third-person mode. However, I will
now argue that this conditional is generally false.
A conception of justice gains stability in a pluralistic liberal democracy as individuals
come to form their own justifications for the principles of justice from within their own
comprehensive doctrines. When individuals do so, this forms what Rawls calls a reasonable
overlapping consensus on the principles of justice. Rawls holds that his political liberalism has
three features that enable justice as fairness to be affirmed from within an overlapping
consensus: first, its requirements are limited to the system of interaction among the major social,
political and economic institutions of society (i.e. the basic structure), second, “acceptance of
justice as fairness presupposes no comprehensive view,” and third, the fundamental ideals and
assumptions of justice as fairness are drawn from the shared, public political culture of a liberal
democracy.
76
If a theory of justice were to rely on controversial assumptions resting on any
particular comprehensive doctrine, it is doubtful that it could be affirmed from within all
reasonable comprehensive doctrines, especially when comprehensive doctrines tend to advance
conflicting claims about value and truth. As a result, arguments offered by legislators for

76
Rawls, Justice as Fairness: A Restatement, 33.
78

particular policies must not rely on comprehensive doctrines. Rather, they should rest only on
the shared public values of political society. The constraints on arguments offered in
interpersonal contexts, then, are generally different from the constraints on public policy
arguments offered by the state or any of its representatives.
Cohen’s interpersonal test relies on the idea that any public policy argument can be
rejected if its utterance in interpersonal contexts (i.e. in its second-person mode) is unacceptable
to reasonable interlocutors. Specifically, the rejection of a public policy by a reasonable
interlocutor is sufficient for rejecting the third-person mode of the argument as providing a
justification for the proposed policy. For instance, Cohen rejects the Incentives Argument on the
ground that its second-person mode analogue, i.e. the Talented’s Argument, may be reasonably
rejected by the untalented. But as the Nurse’s Argument shows, there are some instances in
which an individual may be reasonable in her unwillingness to work without incentives (as
Cohen would himself agree). Intuitively, we would want such cases to pass the Interpersonal
Test while others—such as perhaps the high-flying marketeer’s argument—fail. Unlike the
high-flying marketeer, the nurse’s demand for incentives are reasonably acceptable to her
interlocutor and so facts about her projected behavior do provide good reasons to support a
policy of providing her incentives. Reasonable interlocutors would not find her reasons for
offering the Talented’s Argument unacceptable. Hence, in order to properly perform the
Interpersonal Test, we will need to examine the reasons an agent has for making the factual
premise of a policy argument true and, further, judge whether those reasons are reasonable or
not.
However, if we accept that the publicity constraints that apply to public policy arguments
offered by the state or its representatives do not also hold of arguments offered in the background
79

culture, we will see that there will be many instances in which interlocutors may reject each
other’s reasons for making the factual premise of an argument true while nonetheless accepting
the argument as it’s publicly posed.
To illustrate, consider the following scenario. Jack and Jill are both owners of mid-size
retail stores who are in favor of state-mandated and subsidized paid paternity leave. They both
offer the following argument:
Normative premise: Fathers of newborns or newly adopted children should be given paid
paternity leave.
Factual premise: Unless the state mandates and subsidizes paid paternity leave, I will not
provide it for my employees.
Conclusion: The state should mandate and subsidize paid paternity leave.

We can imagine that if Jack and Jill were to engage in a conversation about state subsidized
paternity leaves, they would initially find themselves in agreement with each other. However,
suppose that they begin to discuss their reasons for why they themselves would not offer
paternity leave without state subsidization. Suppose Jack explains that his company is barely
keeping afloat as it is, and that offering paid paternity leave would be financially unfeasible.
Moreover, he explains that were he to offer paid paternity leave, his company would not be able
to compete with other companies that refuse to offer leave. Much as he would like to, Jack
won’t offer paid paternity leave unless it is mandated and subsidized by the state because he
would be driven out of business. Suppose that Jill, however, is a radical in disguise. She
believes that capitalism is evil and that “liberal” measures to correct injustices in the market only
serve to prolong the reign of capitalism. It is better, she thinks, that the injustices of capitalism
be made more and more intolerable so that workers will be induced to revolt and bring about the
communist revolution sooner rather than later. Jack does not share Jill’s reasoning and, not
unreasonably, rejects her reasons for making the factual premise of the Paternity Leave
80

Argument true. However, Jack fully endorses the Paternity Leave Argument and, moreover,
would endorse the Paternity Leave Argument in its third-person mode.
This little scenario illustrates the idea that in a pluralistic liberal democracy, people will
often disagree on matters concerning truth and value but may nonetheless acknowledge their
common agreement over arguments concerning public policy (so long as those arguments rely
only on public reasons). Indeed, Rawls emphasizes that one of the strengths of a political
conception of justice is that it may be affirmed by citizens who hold a wide range of competing
comprehensive doctrines. Despite holding incompatible views about truth and value, citizens
can nonetheless find common grounds for reasoning with each other and agreeing on
fundamental matters of justice:
While in a well-ordered society all citizens affirm the same political conception of
justice, we do not assume they do so for all the same reasons, all the way down. Citizens
have conflicting religious, philosophical, and moral views and so they affirm the political
conception from within different and opposing comprehensive doctrines, and so, in part
at least, for different reasons. But this does not prevent the political conception from
being a shared point of view from which they can resolve questions concerning the
constitutional essentials.
77

While the state and its representatives may only offer arguments that appeal to public reason,
arguments that take place in the background culture may appeal to the various reasonable
comprehensive doctrines that people hold. Indeed, one feature of a healthy democracy is that
political discourse within the background culture actively draws upon a wide diversity of
competing conceptions of the good. At the same time, reasonable individuals will recognize
that, given the burdens of judgment, other reasonable individuals may reasonably reject the
interpersonal arguments they offer but come to support the same public policies on the basis of
their own reasonable comprehensive doctrine. In interpersonal contexts individuals may offer
two types of arguments in their attempt to persuade their interlocutor: those that appeal to

77
Ibid., 32.
81

particular conceptions of the good, on the one hand, and those that appeal to shared political
values on the other. Though arguments of the latter type are more likely to persuade a dissenting
interlocutor, it can often be unclear how those political values should be interpreted relative to a
particular policy. As a result, political discourse in the background culture will often involve
non-public arguments—that is, arguments involving premises that rest on some particular
reasonable comprehensive doctrine.
78

Given that individuals may reject each other’s non-public arguments but ultimately come
to an agreement about a particular policy, Cohen’s Interpersonal Test looks to be too strong. The
test maintains that if, in its second-person mode, an argument can be rejected in interpersonal
contexts, then the argument (in the third-person mode) cannot be comprehensively justified. But
as we’ve seen with the Nurse’s argument and with the exchange between Jack and Jill, whether
we reject an argument as it’s offered in an interpersonal context will depend on our interlocutor’s
reasons for asserting the premises of the argument. In ordinary interpersonal contexts, those
reasons may often invoke comprehensive doctrines that we do not ourselves endorse, as in the
case of Jack and Jill. Or, as with the Nurse’s argument, we may find our interlocutor to be
reasonable in making a factual premise true, while we would reject the same argument offered by
another—such as the high-flying marketeer—on the grounds that his reasons for making the
same factual premise true are unreasonable.
Contrary to Cohen, then, it is far from clear that the Incentives Argument fails the
Interpersonal Test. In some instances, speakers who offer the Incentives Argument in its second-

78
Particularly, when a society is not well-ordered, citizens who are engaged in discourse concerning matters of
constitutional essentials may, not unreasonably, offer reasons grounded in their own reasonable comprehensive
doctrine provided that those reasons, given the historical and social conditions, are required to give strength to the
political conception of justice in which the ideal of public reason could eventually be realized. See Rawls on the
inclusive view of the ideal of public reason: John Rawls, Political Liberalism, expanded ed. (New York: Columbia
University Press, 2005), 250-251.
82

person mode will be regarded, by their interlocutor, as unjustified in making the factual premise
of the argument true; in other instances, however, interlocutors may regard the speaker as
justified in making the factual premise true. Moreover, given that people often agree in their
support of a particular policy while disagreeing in their non-public reasons for supporting that
policy, the Interpersonal Test will reject many policies that would otherwise be intuitively
regarded as justifiable. Interlocutors may be united in their support of a particular policy as well
as their acceptance of a public argument for that policy, but reject each other’s reasons for
supporting the premises of that argument. Intuitively, such arguments may be considered
justifiable, even if the argument may reasonably be rejected in interpersonal contexts, depending
on who offers the argument and to whom it is addressed.
If Cohen wishes maintain a distinction between the Talented’s Argument as it is offered
by the nurse and the high-flying marketeer, then whether an argument can be comprehensively
justified will depend not only on whether interlocutors accept the premises of the argument as
true but also on whether they recognize the speaker as being reasonable or justified in making the
factual premise of the argument true. Intuitively, we would regard the Talented’s Argument as
justified when it is uttered by the nurse, even if we would not regard it as justified when offered
by the high-flying marketeer. So even when an interlocutor does not reject the truth of any of the
premises of an argument, she may nonetheless deny that the argument provides justification for a
policy on the grounds that the speaker is unjustified in making the factual premise of the
argument true. But we may only find a speaker to be unjustified in making some fact true once
we examine her reasons for doing so. When Jack inquires into Jill’s reasons for why she will not
offer paid parental leave unless mandated to, he will likely feel uncomfortable in accepting the
Parental Leave argument, as it is uttered by his radical interlocutor. He may even find Jill
83

unjustified in making the factual premise true. But this does not mean that he rejects the Parental
Leave argument on the whole as unjustified. Indeed, he himself affirms the argument and does
so for his own reasons.
IV FROM EQUALITY OF OUTCOME TO THE DIFFERENCE PRINCIPLE
Let us now consider an alternate criticism that Cohen launches against the Difference
Principle. Cohen reconstructs the reasoning of the parties to the Original Position as proceeding
in two stages, which he calls the Pareto Argument.
79
Cohen explains that the first stage
establishes equality of outcome, and the second stage moves from equality of outcome to a
pareto-superior state of inequality. Note that, strictly speaking, the Pareto Argument is not an
argument. As Cohen has presented it, it is a series of four claims; however, we may take it as
capturing the spirit of an argument whose conclusion is that it is sometimes irrational to favor
equality of opportunity over pareto-efficiency:
The Pareto Argument
The First Stage
(1) True equality of opportunity requires the nullification of all morally arbitrary causes
of inequality.
(2) There exist no causes of inequality that are not morally arbitrary.
The Second Stage
(3) It is irrational to insist on equality when it is a pareto-inferior state of affairs.
(4) Sometimes equality is pareto-inferior.
80

The first stage of the argument moves from equality of opportunity (premise 1) to
equality of outcome (premise 2). If there are causes of inequality that are not morally arbitrary,
then the first premise would allow such causes to affect the distribution of income and wealth.
On its own, the first premise simply requires that morally arbitrary causes be nullified. The
second premise goes on to state that there are no causes of inequality that are not morally

79
Cohen, Rescuing Justice and Equality 87-88.
80
Ibid., 89.
84

arbitrary. As a result, when taken together, the two premises recommend equality of outcome as
the just initial distribution.
First, it is important to note that this is an unfaithful reconstruction of the reasoning of the
parties in the original position. Although Rawls does note that the distribution of natural talents
and endowments should be regarded as a common asset, not only has Rawls never stated (1) or
(2), he is not committed to either of the claims. For Rawls, fair equality of opportunity requires
that “those who are at the same level of talent and ability, and have the same willingness to use
them, should have the same prospects of success regardless of their initial place in the social
system.”
81
Clearly, the principle does not require that all morally arbitrary sources of inequality
be nullified; to see this, one need only reflect on the fact that it is often a matter of one’s luck in
the natural lottery whether one has certain native talents, abilities, or even the willingness to
develop these gifts. Under fair equality of opportunity, those who have been lucky in this sense
will have greater opportunities than those who have not been so lucky. What fair equality of
opportunity prohibits is that opportunities be influenced by one’s initial starting place in the
social system, not that all morally arbitrary sources of inequality be nullified. Such a
requirement is better described as a luck egalitarian view rather than a Rawlsian one. Moreover,
it is mysterious to me why one might think (2) true. Intuitively, if someone works harder than
another, or elects to work while another declines to work, then the inequality that results does not
seem morally arbitrary, for example, but morally justified by the greater labor of that worker.
However, the Pareto Argument is nonetheless worth discussing because criticisms of the
argument would suggest that the luck-egalitarian motivations for equality of outcome seem to
conflict with concessions to pareto-efficiency, an observation that I had previously noted in my
criticism of Ronald Dworkin’s version of luck egalitarianism.

81
See Rawls, A Theory of Justice, 73.
85

The second stage of the argument recommends moving away from equality of outcome
when pareto-superior distributions are available. The third premise is an instantiation of the
more general claim that it is irrational to insist on pareto-inferior states of affairs. That is to say,
if there is a pareto-superior state of affairs available, then one ought not prefer the pareto-inferior
state, on pain of irrationality. By itself, the third premise does not yet recommend moving away
from equality of outcome. It is only with the addition of the fourth premise—that in some cases,
a pareto-superior but unequal state of affairs is available—that the second stage recommends
moving away from equality of outcome in order to achieve more efficient distributions.
Cohen holds that there is an inconsistency between the two stages of the argument;
“Consistent adherence to the rationale of its first move puts the second move into
question…anyone who believes that, because the possible sources of inequality are morally
arbitrary, an initial equality is prima facie just has no reason to believe that the recommended
Pareto improvement preserves justice, even if that improvement should be accepted on other
grounds.”
82
While the first stage of the argument establishes equality as the only just initial
distribution, the second stage of the argument moves away from equality on considerations of
efficiency. For Cohen, the move from the first to the second stage consists in a trade-off
between the two values. On his view, the considerations that establish equality as the only just
initial distribution in the first place should continue to hold sway even as we reflect on the
possibility of moving to pareto-superior states of affairs that do not, as the first stage requires,
nullify all morally arbitrary sources of inequality. The two stages of the argument, then, are at
odds with each other. The first stage recommends equality of outcome on the grounds that all
sources of inequality are morally arbitrary and thus should be nullified; the second stage, on the
other hand, allows inequalities that arise from morally arbitrary sources, so long as those

82
Cohen, Rescuing Justice and Equality 89-90.
86

inequalities are pareto-superior to equality. As a result, it is irrational, according to the second
stage of the argument, to insist on equality of outcome. Though ultimately intended to provide a
critique of the Difference Principle, which regulates inequalities in income and wealth, Cohen
narrows his discussion to the distribution of wages. He then demonstrates that the two stages can
be represented by the following simple chart, which illustrates the move from an equal
distribution of wages to an unequal but pareto-superior distribution of wages:

D1 represents the distribution recommended by the first stage of the Pareto Argument; under
equality of outcome, the wages of the untalented (Wu) are the equal to the wages of the talented
(Wt). D2 represents the distribution recommended by the second stage of the Pareto Argument;
under the standard interpretation, incentivizing the talented will result in increased productivity
that benefits the untalented. The greater wages of the talented, for example, may enable greater
investments in business, which in turn may create more jobs or allow businesses to provide
higher wages to workers.
1 The Dilemma
Cohen responds to the Pareto Argument he has attributed to Rawls by arguing that
commitment to the first stage should prevent acceptance of the second. In particular, he argues
87

that the fourth premise—that equality is sometimes pareto-inferior—relies on the assumption
that the talented can be induced to greater efficiency only when incentivized. On the standard
application of the Difference Principle, the second stage is accepted on the grounds that everyone
would be made better off by allowing an incentivizing structure that would maximally raise the
levels of those who are under the new distribution the least advantaged. In response, Cohen
argues that if it is indeed possible to increase efficiency by offering incentives, then it is not the
case that the talented are literally unable or incapable of being more productive. So, it is possible
for the talented to increase their productivity in the absence of incentives. If the talented were
truly just, moreover, they would indeed increase their productivity, despite the absence of
incentives, by acting from an “ethos of justice.”
83
In a truly just society whose members have
internalized the principles of justice, the talented would guide their occupational choices
according to the aim of the Difference Principle
84
and so would not require incentives in order to
work for the maximal benefit of the least talented. Consequently, a third distribution (D3) would
be recommended:

83
Cohen identifies such an ethos as one of justice; Rawls, however, would most likely regard it as a non-political
(i.e. comprehensive) ethos, such as an ethos of benevolence.
84
The aim, that is, as interpreted by Cohen to be to maximize the well-being of the least talented.
88

On Cohen’s view, the distribution of income and wealth in a truly just society would be
represented by D3. Guided by an “ethos of justice,” the talented would be willing to produce as
much as they would if they were given incentives. As a result, the untalented will be maximally
benefited while preserving equality between the talented and the untalented. Cohen chooses to
interpret the most and least advantaged of society as the most talented and untalented. He seems
to take this to be an unproblematic move. However, such a move opens up the possibility of an
additional distribution, which he should be committed to, given his view that true affirmation of
the Difference Principle requires individuals to modify their choices and behavior so as to
maximally benefit the least talented. Consider, for example, D4, in which Wu > Wt. If we are to
prefer D3 over D2, then by the same reasoning, we ought to prefer D4 over D3 and, moreover,
any subsequent distribution in which the wages of the untalented are brought even further higher
while the wages of the talented lower. Intuitively, this produces the wrong result.
Returning to Cohen’s reconstruction, the first stage of the Pareto Argument, which
establishes equality of outcome as the only just initial distribution, disallows D2 while permitting
D1 or D3. The general claim—that it is irrational to insist on a pareto-inferior state of affairs—
on which the third premise of the argument rests then rejects D1 in favor of D3. In D1, all enjoy
equal wages, but the talented do not make their occupational choices with the aim of benefiting
others in mind. In D3, all again enjoy equal wages, but the talented put their talents to use in
socially beneficial ways. As D3 is pareto-superior to D1, one ought to prefer D3 over D1, on
pain of irrationality. Lastly, the fourth premise of the argument—that sometimes equality is a
pareto-inferior state of affairs—is shown to be false, at least of just societies. There are no other
possible states that are pareto-superior to D3.
89

Moreover, Cohen argues, the move from the first to the second stage of the Pareto
Argument is subject to a dilemma:
Either the talented carry, in D2, a special burden that is compensated for by the
difference between Wt and Wu, in which case it is a mistake to represent movement from
D1 to D2 as issuing in an inequality; or Wt is not required to compensate for any burden,
in which case there is no reason for an egalitarian to regard D2 as acceptable, and every
reason for him to recommend D3. In other words: either the extra money that the
talented get in D2 makes them no better off, all things (including labor burden)
considered, than the untalented, or it does make them better off than the untalented, all
things considered.
85

The imagined proponent of the Pareto Argument aims to provide support for D2 and to deny the
possibility of D3. One way she could do this is to argue that the unequal wages of the talented
and untalented are necessary to compensate the talented for their special labor burdens. In this
case, no tension will arise between the two stages of the argument, for no one can reasonably
deny that special labor burdens are relevant, from a moral point of view, to the distribution of
income and wealth. D2, then, does not consist in the kind of inequality that would be disallowed
by the first stage of the Pareto Argument.
The problem with this strategy, Cohen argues, is that it would not be entirely correct to
say that D2 is a state of inequality. Although the wages of the talented are higher than those of
the untalented, the inequality arises from the extra remuneration needed to compensate the
talented for their special labor burdens. All things considered, then, the talented are not made
better off than the untalented; rather, taking into account special labor burdens, the two groups
are equal.
If, on the other hand, the proponent of the Pareto Argument accepts that the higher wages
of the talented in D2 are not strictly the result of compensating for special labor burdens, then
there will be “no reason for an egalitarian to accept D2 as acceptable.” For any other source of

85
Cohen, Rescuing Justice and Equality 105.
90

inequality between the wages of the talented and untalented, as established by the first stage of
the Pareto Argument, are arbitrary from a moral point of view to the distribution of income and
wealth and so should be nullified. Acceptance of the first stage of the argument would require
the denial of D2 as an acceptable distribution. Moreover, the possibility of D3, a state of
equality with the highest minimum, undermines whatever justification D2 may have gained from
the second stage of the Pareto Argument. Even if one were to accept both the third and fourth
premise (that sometimes equality is pareto-inferior, in which case it would be irrational to insist
on it), D3 will be preferable to both D1 and D2.
2 Equality of Primary Social Goods vs All Things Considered Equality
In the argument just described, Cohen relies on an equivocation of the meaning of the
term equality. As in previous arguments, it is not entirely clear whether Cohen is more
concerned to offer an internal critique of the standard interpretation of the Difference Principle or
to motivate external grounds for rejecting elements of Rawls’s theory. Hence, I’ll evaluate his
argument according to both of these two distinct aims.
First, let us suppose that Cohen wishes to offer an internal criticism. It is important, then,
to measure the equality of people’s levels of advantage in the way that Rawls does—namely, by
measuring their social primary goods. We will say, then, that two people are equal so long as
they have the same or comparable social primary goods. As a result, D2 (in which the wages of
the talented are greater than those of the untalented) is not a state of equality, when we
understand equality to be equality of primary social goods. So, even if D2 is a distribution in
which the greater wages of the talented arise only from compensation for burdensome labor, it
would be incorrect to regard D2 as a state of equality. The first horn of Cohen’s dilemma claims
that if the difference between the wages of the talented and untalented arise from compensation
91

for special labor burdens, then “it is a mistake to represent movement from D1 to D2 as issuing
in an inequality.” This claim, however, will be false if, with Rawls, we measure people’s levels
of advantage in terms of social primary goods, for clearly D2 is a distribution in which the
talented and untalented have different social primary goods. The second horn of the dilemma
will likewise be undermined. If the talented under D2 receive more than what is needed to
compensate them for their labor burdens, then, Cohen claims, the egalitarian will have reason to
reject D2 and favor D3 instead. But regardless of whether the egalitarian rejects D2, it would be
too quick to say that she should prefer D3. Although we have supposed that the greater wages of
the talented are more than what is required to compensate for labor burdens in D2, we have no
assumptions about the distribution of labor burdens in D3. It may be the case, for example, that
D3 represents a distribution in which the talented and untalented receive the same wages even
though the talented undertake more burdensome labor. The egalitarian may then find D3
unattractive, as under this distribution more is asked of some than of others.
So, if Cohen’s primary aim is to offer an internal critique, that criticism would be
unsuccessful. By pointing out in the first horn of his dilemma that it would be a “mistake” to
regard inequalities that arise from compensation for special labor burdens as issuing in an
inequality, Cohen motivates the idea that equality of primary social goods does not capture the
“true” meaning of equality. Indeed, his restatement of the dilemma appeals to the idea of being
better off, all things considered, than another: if the wage difference is the result of burdens
compensation, then though the talented receive higher wages, they are not better off than the
untalented, after their special labor burdens have been taken into account. Given that different
occupations vary in the extent to which they involve special labor burdens, equality of primary
social goods will not be the same as all things considered equality, and vice versa. It is not the
92

case, then, that the following three claims can be simultaneously true of a society: (1) people’s
levels of advantage are measured by their social primary goods (2) people’s levels of advantage
are equal; and (3) burdensome labor is compensated by extra remuneration.
As an external critique, Cohen’s relies on the intuitive force of the idea of all things
considered equality. Although Rawls measures people’s levels of advantages according to their
social primary goods, Cohen’s dilemma suggests that levels of advantage are better understood
according to a broader criterion--how well off people are, all things considered. In other words,
primary social goods alone do not provide an adequate measure of levels of advantage; instead,
Cohen pushes for the view that two people are equal when they are all things considered equal,
not when they have comparable social primary goods. It is on this broader understanding of
equality that two individuals can be equal despite having unequal levels of income and wealth, as
in the special burdens case. Cohen does not explicitly explain what he means by all things
considered equality, but the following two possible interpretations seem the most salient:
(1) equality in social primary goods, were compensation for special labor burdens
subtracted; or
(2) equality in welfare or well being.
The first interpretation roughly proposes that all-things-considered equality is realized when the
distribution of income and wealth reflects only the varying degrees of burdensome labor that
people undertake. In other words, on this interpretation of all-things-considered-equality, each
individual is equally advantaged with respect to others once we have taken the sum of their
income and wealth and their special labor burdens; or, to put it another way, we may measure
level of advantage as income and wealth minus special labor burdens. In a society in which extra
93

pay is given only in compensation for burdensome labor, everyone will turn out to have an equal
level of advantage despite differing in their income and wealth.
The first interpretation, however, is untenable. All things considered equality is achieved
when everyone has a comparable level of advantage, once the labor burdens they undertake has
been subtracted from their income and wealth. However, while primary social goods such as
income and wealth are resources, special labor burdens seem to be better captured in terms of
welfare rather than resources. Intuitively, a strong explanation for why labor burdens should be
compensated is that labor burdens negatively impact people’s well-being in some way or other.
If we understand special labor burdens as a concern of justice because they cause reductions in
well-being, then it is difficult to see how we can measure people’s levels of advantage with
respect to one another by somehow subtracting their special labor burdens from their social
primary goods. While we may be able to perform comparisons between individuals with respect
to their resources or, separately, with respect to their welfare, it does not seem feasible to
perform a comparison that incorporates heterogeneous metrics.
The second interpretation, however, measures people’s advantages entirely according to
their levels of welfare and so does not share the first interpretation’s weakness. One person will
be more advantaged than another if the first has a greater level of welfare than the second. On
this view, two people can be counted as all things considered equal despite having different
social primary goods, so long as they have comparable levels of welfare. On the assumption that
burdensome labor typically causes reductions in well-being, those who undertake burdensome
labor will, all else equal, have lower levels of welfare than those whose work is not burdensome.
If we also suppose that income and wealth can be used to increase one’s level of welfare (by
enabling the recipient to purchase the goods that may contribute to greater welfare), then those
94

who undertake burdensome labor may need to have extra income and wealth to reach a level of
welfare comparable to those whose labor is not burdensome.
The dilemma Cohen had posed considers two possibilities: either the greater wages of the
talented in D2 result from compensation for burdensome labor or they don’t. If they do, then D2
is a state of all things considered equality, despite the inequality in wages. If they don’t, then D2
is a state of all things considered inequality, in addition to the inequality in wages, and the
egalitarian should prefer D3 instead. Again, however, if all things considered equality is,
roughly, equality of welfare, then it is not obvious that the egalitarian should prefer D3. D3 is a
distribution in which all receive comparable wages; it is a state of equality in primary social
goods. It is not, however, necessarily a state in which all have comparable levels of welfare.
Indeed, if some are asked to take on burdensome labor that benefits others but does not provide
additional compensation, then it is plausible that those who undertake greater burdens will have
lower levels of welfare than others. So while D3 will be a state of equality in primary social
goods, it will be a state of inequality in welfare—that is, all things considered inequality. If the
egalitarian cares primarily about equality of welfare, then she will not prefer D3 unless it is
stipulated that D3 is also a state in which all enjoy comparable levels of welfare. However, this
is quite implausible on the plausible assumptions that occupations tend to vary in the extent to
which they involve burdensome labor and that burdensome labor typically results in reductions
in an individual’s level of welfare.
V CONCLUSION
Cohen’s two main arguments against the standard interpretation of the Difference
Principle have been shown to be problematic. His first argument attempts to demonstrate that
standard arguments in favor of allowing greater income and wealth to the talented do not pass the
95

Interpersonal Test, as they cannot be justified in interpersonal settings. However, whether an
argument can be reasonably acceptable to an interlocutor will often depend on the speaker’s
reasons for offering the argument or making the premise of the argument true. While an
interlocutor may reasonably reject an Incentives Argument as it is offered in its second-person
mode by an “acquisitive self-maximizer,” it would not be reasonable for her to reject the
argument as it is uttered by someone who requests greater compensation for the greater burdens
she undertakes in performing socially necessary labor, such as a nurse who demands higher pay
for working the night shift. At best, the interpersonal test may provide a guide for
distinguishing between various utterances of the Incentives Argument, but it cannot rule out the
Incentives Argument altogether, since some instances of the Argument do pass the test.
Moreover, in a pluralistic democracy, interlocutors will often disagree with the other’s reasons
for supporting some of the premises of a public policy argument. This is because people hold a
variety of comprehensive doctrines. Despite their differences, reasonable people will also
recognize the possibility of finding an argument justifiable and acceptable to them, even though
the interlocutor who offers the argument may have vastly different personal reasons for
supporting the premises of the argument. Because we need to look into the speaker’s reasons for
making the premise of an argument true in order to judge the acceptability of an argument
offered in the second-person mode, the Interpersonal Test will prove too strong. Many
justifiable arguments will be rejected if, in their second-person mode, the arguments are uttered
by interlocutors who hold vastly different reasons, stemming from their comprehensive
doctrines, for supporting the premises of the arguments.
The second major argument Cohen offers attempts to demonstrate that the standard
interpretation of the Difference Principle falls prey to a dilemma. I argued that this objection
96

cannot serve as an internal critique of Rawls. Moreover, it fails as an external critique as well, as
more needs to be said in order to defend the claim that a state of equality of primary social goods
can be a state of all things considered equality. On the assumption that occupations involve
varying levels of burden, these two states cannot be the same.
Cohen’s criticisms, however, are nevertheless suggestive. They motivate the important
idea that Rawls’s Difference Principle does not guarantee that burdensome labor will be
appropriately compensated. The idea that, all else equal, greater burdens justify greater
compensation is intuitively compelling but does not play any clear role in Rawls’s theory. While
the two objections discussed in this chapter are not entirely successful, they nonetheless establish
an important concern about the Difference Principle that warrants further attention.

97

4 Burdens Sensitive Egalitarianism

In this chapter I would like to provide a fuller account of what I call the Burdens
Sensitive view, which provides an answer to the question “What are the factors relevant, from a
moral point of view, to the distribution of income and wealth?” The view holds that the
distribution of income and wealth should not reflect the natural distribution of talents and other
native endowments but rather the differences in the various burdens of socially necessary work.
I will begin with a discussion of the most common and widely recognized burdens of various
occupations. By exploring the reasons for why we intuitively regard these as burdensome, we
can begin to see the common features underlying these different types of burdens. Burdens, I
will argue, consist in the work-related harm or risk of harm to an individual’s well-being or
quality of life. While my discussion of burdens draws its examples from various kinds of
occupations, I also give attention to other forms of socially necessary or beneficial work that
aren’t remunerated on the labor market, such as child-rearing and other forms of care. Because
caregiving is such an important part of our social economy and clearly a major concern of
justice, the Burdens Sensitive view has an advantage over other distributive principles that regard
caregiving as a peripheral matter, as a part of non-ideal theory, or as derivative from the
principles of justice that primarily apply to those engaged in market cooperation.
I BURDENS
In an unfettered market, individuals who possess talent, ability, the willingness to work
hard, or prior wealth are far more likely to command a greater income and gain more wealth than
those who lack these resources. In previous chapters, we’ve discussed competing views on how
distributive justice should respond to this fact. Rawls’s Difference Principle allows the more
98

fortunate members of society to enjoy their greater income and wealth only under the condition
that the inequality is necessary to raise the worst off to the greatest extent. Luck egalitarians,
however, question whether the greater life prospects of the talented are truly justifiable. After
all, many talents are the result of good luck in the genetic lottery. And, as Rawls as well points
out, even the willingness to try, make an effort, and work hard is at least partly dependent on
“happy family and social circumstances.”
86
For both Rawls and luck egalitarians, designing the
basic structure of a society in such a way that it unrestrictedly compounds the natural advantages
enjoyed by the lucky with social and economic advantages would foster feelings of envy and
mistrust rather than of civic friendship. Under such systems, natural inequalities are transformed
into social and economic inequalities, and for luck egalitarians especially, the failure to mitigate
the effects of bad brute luck on an individual’s life prospects is a failure of justice.
However, luck egalitarians differ from Rawls in their commitments when it comes to the
advantages of native talents and abilities on the market. Though particular luck egalitarian
theories diverge in their details, they share the core thesis that disadvantages that arise from the
unchosen factors of an individual’s circumstances are to be redressed.
87
On the flipside,
advantages that arise from the unchosen features of an individual’s circumstances are
undeserved. Distributive justice, on this approach, is fundamentally concerned with the
distinction between luck, on the one hand, and choice or responsibility on the other. Luck
egalitarians, then, are committed to the claim that the economic gains that the naturally talented
or able enjoy on the market are undeserved. Typicallly, this will result in a recommendation that
such gains be taxed and redistributed to the genetically unlucky, contrary to Rawls’s theory,

86
Rawls, A Theory of Justice, 74.
87
Conversely, luck egalitarians hold that individuals should be held responsible for disadvantages that arise from
their irresponsible choices.
99

which may allow incentives for the talented to train and use their talents, provided, of course,
that such incentives are necessary to maximally benefit the worst off.
My view shares some common ground with the luck egalitarian’s in agreeing that native
talents and other natural endowments should not be the basis on which an individual should be
remunerated. While not inconsistent with the luck egalitarian position, my view nonetheless
diverges from it by eliminating any mention of the distinction between an individual’s unchosen
circumstances and the choices for which she is to be held responsible.
88
Rather, the Burdens
Compensation Principle holds that inequalities in income and wealth should reflect only the
differences in the burdensomeness of socially necessary labor. Centrally important to this
principle is the notion of burden. In this section, I will argue that we should adopt an objective
set of criteria for identifying the burdens of an occupation. By doing so, members of a society
will be able to interpersonally justify their level of economic advantage by appealing to
publically accessible, objective reasons for their level of economic advantage. In taking
remuneration for work to be compensation for the burdens of that work, my principle will allow
differences in the income and wealth of members of society that are entirely consistent with a
governing ideal of equality. For so long as long as income and wealth is distributed according to
what is necessary to compensate individuals for the burdens that they undertake in contributing
to the larger social good, those who earn a higher income and are able to amass greater wealth do
so only on the condition that their greater economic advantage comes at the cost to other aspects
of their lives, whether it be the result of reduced leisure hours, increased stress or other health
and safety risks, the fatigue that results from intense physical or intellectual exertion, and so
forth.

88
In section II, I will clarify the relationship between the luck egalitarian principle and the Burdens Compensation
Principle with further details.
100

1 A Preliminary List
We have an intuitive idea of what burdensome work is: work that is tedious, physically,
mentally, or emotionally exhausting, stressful, dangerous, or requiring long and/or erratic hours
are all a few examples that quickly come to mind. But while we may, in many cases, be able to
intuitively identify the burdens of a given occupation, we by no means possess a complete list of
such burdens. Moreover, it is easy to see that many borderline cases may arise in which we
could disagree about whether to consider some aspect of an occupation a burden to be
compensated for. Scientific researchers may experience a great amount of frustration or
disappointment when their experiments fail. It is not immediately clear, however, whether these
should be considered burdens for which they should be compensated or simply a forseeable
consequence of the nature of scientific inquiry that all future scientists should be prepared to
accept as par for the course. Disagreement about how to identify burdensome work may also
arise from differences in individual dispositions and tastes. What some intuitively find
burdensome about an occupation may be considered by others to be precisely what makes that
occupation attractive. Individuals who are risk averse may find the high speeds and imminent
sense of danger of professional race car driving to be highly burdensome while others may find
those same features thrilling.
Let us begin untangling these various issues by first laying out a preliminary list of what I
think are uncontroversially regarded as burdens that an occupation can have. The following are
features that I think many would agree describes the burdensome aspects of work:
(i) dangerous to one’s health and safety
(ii) requiring intense physical labor
(iii) involveing long, irregular and/or erratic hours
(iv) highly repetitive, tedious or boring
(v) emotionally or mentally exhausting
(vi) stressful
101

While we will generally agree that these are all aspects of an occupation that are burdensome, it
is worth discussing why we identify these as burdens. By looking at these diverse sources of
burden and identifying the grounds on which they are considered burdens, we can begin to
formulate a systematic account of what a burden is.
After arriving at a definition of a burden, I will explain how my theory will handle the
two questions previously mentioned in the introduction to this section: Should subjective
burdens, such as the frustration frequently felt by scientists, be remunerated? (To be addressed at
the end of Section I.) And, more generally, should remuneration be partly determined by the
individual worker’s judgments of the burdensomeness of their work? (To be addressed in
Section II).
2 Harm and Risk of Harm
A good place to start in fleshing out our notion of a burden is to begin with a discussion
of occupational hazards—that is, the risks and dangers to worker health and safety inherent in a
particular occupation. Construction work, for example, is well known to carry significant risk of
injury or fatality. Many construction workers develop joint and back problems and face a non-
trivial chance of being fatally injured, most commonly as a result of falling, electrocution, being
struck by objects, or being caught in or between objects. While these accidents are sometimes, if
not often, attributable to improper regulation and training, some of these accidents are
unavoidable misfortunes that arise from the inherent nature of the work. It would be
unreasonable to fail to compensate work-related injuries on the grounds that, given the
foreseeable dangers of engaging in such work, construction workers have accepted a risk for
which they should be held responsible. The work they engage in is socially necessary; in fact, it
is vital to society that individuals fill these occupations. When a worker provides a service or
102

product that is not only desired by others in society but needed, it is clearly imperative that
sacrifices undertaken by the worker (such as harm to health or safety) be compensated not just
for practical reasons but for moral reasons as well. Failing to compensate work-related injuries
and harms may result in a problematic scarcity of construction workers. Or, if no scarcity of
workers ensues, its plausible to conjecture that the willingness of individuals to engage in
dangerous work that promises no extra compensation is the result of their lack of adequate
bargaining power. Perhaps the employer does not allow the formation of unions, or there is such
a scarcity of jobs that individuals will accept any work in order to survive. Moreover, a society
that fails to compensate workers for the harms they suffer while doing socially necessary or
desired work would fail to uphold the fundamental ideal of reciprocity that is so central to liberal
theories of justice. These scenarios are unacceptable—one for practical reasons; the other,
moral. So, it seems clear enough that injuries or other harm to worker’s health should be
compensated.
While many construction workers will suffer a work-related injury at some point in their
careers, some may go through their careers without injury. I have briefly explained why injuries
and other harms to well-being should be compensated; one further question we may ask,
however, is whether the risk of a work-related injury should be compensated, and if so, why?
Let’s explore three different answers:
Risk Responsibility
Risks alone should not be compensated. The idea here is that if injuries and other harms
to worker well-being are fully compensated when actually suffered, then there is no need
to compensate the mere risk itself. If a worker suffers an injury, she will be fully
compensated for the injury; if she doesn’t, then the risk has never materialized and no
harm has been done.

Incentivization
Risks, independently of whether they are realized, should be compensated to whatever
extent is necessary to incentivize workers to fill that occupation.
103

Offsetting costs
Risks, independently of whether they are realized, should be compensated because such
compensation is necessary to offset the various costs to the worker’s well-being or
quality of life that the risk alone poses.

Risk Responsiblility fails to take into account the plausible claim that risks alone can impose
some costs to the individual who undertakes them. To illustrate, the significant risk of suffering
a disabling injury that construction workers face can disrupt the worker’s sense of security about
the future. The high chance of being injured poses worries not only about the possibility of
losing important capabilities in the future but also about the effects that such a loss would have
on dependents or spouses. These worries can lead to increased anxiety or stress, which in turn
can have ill effects on the worker’s physical health. Given that the mere risk of injury can
impose actual costs on the individual who undertakes the risk, let’s now turn to the second
answer, which offers an account of the basis on which we should compensate risk.
Incentivization appeals to the idea that rather than attempting to identify a list of objective
burdens, burdens should be regarded as subjective so that workers themselves can determine how
costly they find a risk to be. Whatever their reasons for hesitating to accept the risks to health
and safety involved in a particular occupation, we can measure the compensation needed to
offset the cost of the risk to them by seeing whether, for any given amount of remuneration, they
accept the job or not. The problem, however, with interpreting compensation for risk as
incentives is that the occupational market is not guaranteed to reflect the true value of labor. As
we’ve noted before, if construction workers possess little bargaining power, they might accept
work for remuneration that is lower than what they would consider appropriate to offset the
potential harm to health or loss of capability or life that may result from engaging in that line of
work. So the Incentivizing answer encounters difficulties on moral grounds.
104

The third answer, Offsetting Costs, instead identifies the basis on which to compensate
risks on objective facts about how such risks may affect an individual’s well-being or quality of
life. It recognizes that, even if a risk is never realized, the risk itself imposes some costs to the
individual who undertakes it. As Rawls noted, health is a natural primary good—it is a resource
that anyone, no matter their plan of life or conception of the good, has an interest in possessing.
89

One may add, in agreement with Sen and Nussbaum, that any individual also has a fundamental
interest in possessing capabilities. And, further, individuals also have a fundamental interest in
health and capabilities security—that is, in possessing a reasonable basis for expecting a
continuance of good health and capabilities into the future.
90
Occupations that require workers
to undertake risk of future injury or other harm, or even of loss of life, conflict with their
interests in health and capabilities security. This conflict takes the form of actual harms, as the
prospect of future injury or loss of life results in stress and anxiety in the worker. Moreover, lack
of health and capabilities security may affect myriad other aspects of the worker’s life, most
notably in the individual’s relationships with others. Spouses, for example, may find it difficult
to invest in a future with a worker whose job carries high risk of fatality or disabling injury; and
children of such workers may be more vulnerable to emotional stress. The stress placed on the
worker’s relationships as a result of lack of health and capabilities security is a significant way in
which mere risks of harm can adversely impact the worker’s overall well-being or quality of life.
One thing to note is that the same risk of injury might not warrant the same compensation
across different occupations. Take, for example, a comparison of the risks involved in
gymnastics and construction work. Let’s suppose that both occupations carry the same risk of
disabling injury or fatality. We’ve noted the various ways in which the risk of injury alone can

89
Rawls, A Theory of Justice, 62.
90
Subject, of course, to individual variations in health and to the normal course of decline that accompanies aging.
105

impose costs on the individual who undertakes such risks; let’s also stipulate, for the sake of
example, that these costs are the same for both gymnasts and construction workers. The question
is whether gymnasts should be given the same amount of compensation for risk of injury as
construction workers. Intuitions may not be particularly strong about this case, but I think it is
reasonable to say that while both gymnasts and construction workers should be compensated for
the risk of injury, construction workers should be given greater compensation than gymnasts.
One explanation for this difference appeals to the way in which gymnastics offers its
practitioners various intangible benefits that offset the cost of the risk of injury. Surely,
gymnasts place great value on the preservation of their musculoskeletal integrity. And yet they
accept the significant risk of injury that their profession entails. Perhaps this is because the cost
of the risk does not outweigh the various benefits that come from the development of their skills.
An individual who manages to become a skilled gymnast may enjoy fame, a sense of personal
accomplishment, the increased sense of self-worth and self-confidence that comes from the
perfection of one’s skills, recognition amongst peers, and so on. Construction work, on the
other hand, does not offer as many these intangible benefits. It affords its worker fewer
opportunities for engaging in the intrinsically worthwhile exercise and development of her
various faculties. Particularly talented or hard working construction workers might be able to
gain recognition and perhaps even local fame amongst their peers, and they may gain personal
satisfaction and enjoyment from doing good work. But these are benefits that are present in
almost all occupations and are not particularly forceful or lucrative in construction work in
particular.
Moreover, there is a difference in the degree to which the gymanst’s and construction
worker’s labor is socially necessary. The gymnast contributes to the social good by providing
106

entertainment for those who enjoy watching gymnastics, by providing a source of inspiration or
admiration for the various forms of expression the human body is capable of, or by providing a
sense of communal pride, as when the athlete represents a particular region. However, the
gymnast’s work is less essential to the maintenance of society than is the construction worker’s.
At the very least, a society must have a basic infrastructure to support citizens’ movements,
communication, economic exchanges, subsistence, and so on. All of this would not be possible
without the labor of construction workers. So while we maintain that compensation must
respond to the burdensomeness of labor, it may also respond to the social necessity of the labor.
The recognition that occupations offer benefits as well as burdens helps motivate the idea
that even if the risk of harm is equal across different occupations, different levels of
compensation may be appropriate. To compensate construction workers and gymnasts equally
when there is an inequality in the overall well-being or quality of life that they enjoy would be to
expect one group to carry greater burdens than the other. Moreover, given the greater social
necessity of the labor undertaken by the construction workers, these two considerations together
suggest that there are strong moral reasons for providing greater compensation to the
construction worker.
3 Physical Labor
Occupational hazards are not the only burdens undertaken by workers of a particular line
of work. For example, some occupations involve work that requires intense physical exertion.
While this may be a cause of or contributing factor to the risk of injury or other harm to a
worker’s overall well-being, it is not, considered in isolation, an occupational hazard. (Only the
injuries or risks of injury that result from work that is physically demanding are classified as
occupational hazards). However, I think many would find it intuitive that, all else equal, work
107

that is physically demanding—tiring or exhausting—warrants greater compensation than work
that is not. This is not only because physically demanding work is more likely to carry greater
risk of injury but rather, I think, because we recognize that the extra effort expended by the
worker counts for something—it seems to suggest that, in comparison to the individual whose
work requires less physical exertion, she has engaged in greater labor or, put simply, done more
work. The idea that effort can play a role in our judgments of the morally relevant factors in a
distribution of income and wealth is, of course, familiar. In fact, the sentiment that people
should be rewarded for the effort they take in doing good work seems to be widespread and
commonplace. So I think it is worth exploring whether, in addition to occupational hazards,
extra effort should be counted as a burden that warrants compensation.
Let us first focus on the effort expended by individual workers of the same occupation.
These variations in individual effort do not always correlate with levels of productivity. To
illustrate, consider the various ways two oyster shuckers might vary in the effort they expend in
their work. Suppose Amanda works at a leisurely pace while Beatrice shucks oysters as fast and
she can, continually trying to improve her efficiency. This simple example can lead to several
various outcomes:
(a) Beatrice shucks more oyster than Amanda;
(b) Amanda and Beatrice shuck the same amount of oysters; or
(c) Beatrice, despite her extra effort, shucks fewer oysters than Amanda.

(a) is a pretty straightforward case: intuitively, Beatrice should be remunerated more than
Amanda. And there are several reasons we can offer in an explanation of our intuition: the extra
effort Beatrice expends means that Beatrice works harder than Amanda, Beatrice’s hard work
results in greater productivity, and because Beatrice contributes more to the total social product,
it seems only fitting that she receives greater remuneration in return. The other two cases—(b)
108

and (c)—are not as straightforward or simple. Let’s now suppose that Beatrice is missing a
thumb. In (b), she manages to shuck only as many oysters as Amanda despite her extra effort.
Does Beatrice’s extra individual effort warrant greater remuneration even when the effort does
not result in increased productivity?
One way of answering this question is to see whether the same considerations that
applied to (a) apply to (b). In our response to (a), what seemed to drive our intuition that
Beatrice should receive greater compensation is that her greater contribution to the social product
warrants a greater reward in return. The governing principle here seems to be something like
“the more you contribute, the more you get back in return.” Applied to (b), the principle would
yield the result that Beatrice’s extra effort does not warrant greater compensation. Because
Amanda and Beatrice are equally productive, they receive equal compensation. Applied to (c),
the principle would yield the result that Beatrice receives less compensation than Amanda,
despite being more hard-working.
The problem with the to-each-according-to-her-productivity principle is that it fails to
take into account the morally arbitrary factors that can influence an individual’s productivity.
While there may be some disagreement about whether the principle produces the right result in
(b), most people would feel uncomfortable with its pronouncement in (c). If Beatrice is
compensated less than Amanda, she is at a disadvantage that results not from any responsible
choices or decisions she has made—for example, to take it easy or to focus her attention on
chatting with co-workers instead—but rather from the physical limitations imposed by her
missing thumb. So it seems that effort, independently of its contribution to productivity, is a
relevant factor in deciding how much a worker should be compensated.
109

If individual effort is a relevant factor in the determination of compensation, then it may
turn out that Beatrice should be compensated more than Amanda in (b), despite the fact that they
are equally productive. One thing to note is that compensation need not necessarily take the
form of income but could take the form of other public or private resources. While it may seem
strange to give disabled workers more income on the basis of their disability, it seems clearly
right, on the other hand, to provide them with greater resources in virtue of their handicap.

Likewise, in (c), the extra effort expended by Beatrice need not be compensated by a higher
income than Amanda but may take the form of other kinds of resource.
Capabilities Theorists such as Sen and Nussbaum have criticized Rawls’s theory for
failing to address disability issues. The Burdens Compensation Principle will be an attractive
alternative, for in compensating for the burdens of socially necessary labor, the principle is well
suited to provide the disabled with the greater resources they need in order to participate in social
cooperation. As a result of their disability, disabled workers may expend greater effort than
others engaged in the same occupation, for example. Compensation that responds directly to the
extra effort they undertake as a result of their disability is justified as offsetting the burdens that
they undertake while engaging in social cooperation. The Burdens Compensation Principle also
has the attractive result that in justifying greater compensation for the disabled, it supports the
capabilities needed for the disabled to participate in cooperative production.
4 Hours
Another way in which an occupation can be burdensome is if it requires long, erratic, or
irregular hours. Suppose that a typical nurse’s work week consists of 40 hours, from 8am to
5pm, Mondays through Fridays. Suppose, however, that hospitals often require them to work
overtime or to work erratic shifts in response to varying influxes of emergency patients. Current
110

labor laws, such as that of California, state that workers should be paid 1.5 times the regular hour
rate for each hour worked over the eight-hour workday, and 2 times the regular hourly rate once
the workday has exceeded twelve hours. At first glance, it seems that this policy for overtime
pay conflicts with the way in which the Burdens Sensitive view would remunerate overtime
work. Under the Burdens Compensation Principle, whatever regular hourly rate nurses earn will
reflect the burdens of their occupation: the increased exposure to pathogens and risk of
contracting contagious illnesses, the risk of musculoskeletal injuries that results from frequent
bending or leaning forward, the emotional stress that arises from occasional verbal or even
physical abuse from patients, exposure to toxic chemicals, and so forth. It seems, then, that only
an increase in these aforementioned burdens by 1.5 times would justify a pay increase of 1.5
times for the ninth hour worked in a day. However, the Burdens Sensitive view is by no means
committed to this line of reasoning. In fact, it can easily accommodate overtime pay by pointing
out that long hours generate extra burdens that are distinct from the burdens of the regular work
day.
Consider, for example, some of the reasons for why the ninth hour worked should be
compensated at a greater rate than the eighth hour: after a normal work-day, any additional hour
worked will generate an increasing amount of fatigue for the worker; the risks of injury or harm
inherent during the regular work day increase as a result of the worker’s greater fatigue; the
overtime hour cuts into her expected leisure time, preventing her from spending the hour instead
caring for her children or attending to her other personal affairs; and so forth. Moreover, the
Burdens Compensation Principle would rather recommend that each additional hour of overtime
be compensated at increasingly higher rates. While labor laws currently pay the ninth hour
worked at the same rate as the twelfth hour worked, the Burdens Sensitive view would recognize
111

that the rapidly increasing fatigue and risk of accident for each hour worked overtime constitute
even greater burdens to the worker and hence warrant continually increasing rates of pay.
Work that requires erratic and irregular hours also warrants greater compensation under
the Burdens Sensitive view. Erratic hours are burdensome because they may interfere with our
natural circadian rhythms or other biological cycles. Erratic hours can also make it difficult for
individuals to make plans in advance, and the decreased stability in the worker’s routine can lead
to increased levels of stress as the worker must juggle an unpredictable schedule. Irregular
hours, such as working night shifts, can also interfere with natural sleep/wake cycles. This may
adversely affect the worker’s ability to obtain an adequate number of hours of sleep and the
quality of their sleep. Working night shifts often leads to depression, cardiovascular disease,
diabetes, obesity, and other harms to worker well-being. These are all burdens that are distinct
from the burdens associated with the nature of the work engaged in; consequently, the Burdens
Sensitive view can accommodate the intuition that work that requires erratic or irregular hours
warrants greater compensation.
5 Tedium and Stress
While our discussion of burdens has thus far focused on the ways in which burdens can
primarily affect the physical health or safety of workers, this short section will be devoted to a
discussion in which some occupational burdens can consist in adverse effects on worker’s
overall physical or emotional well-being, quality of life, or capabilities.
Intuitively, work that is highly repetitive and tedious presents its own particular kind of
burden. Take, for example, production line work at a factory. Here, workers are often assigned
a highly specific task that is simple and repetitive. Although these tasks are not necessarily
physically demanding, the mechanical nature of the work can lead to very low feelings of job
112

satisfaction, high levels of frustration, and boredom. This type of work affords individuals little
or no opportunity to exercise and develop their various higher faculties through their work. The
lack of opportunities to make qualified decisions, to learn new skills and abilities and to develop
them, and to meet challenges are all absent in this kind of work. By engaging in work that is
often described as “mind-numbing” or “soul-crushing” drudgery for 40 or more hours per week,
workers suffer from a lower overall quality of life independently of whether their line work
carries occupational hazards or requires intense physical labor.
Work that is emotionally taxing can adversely affect an individual’s psychological well-
being. Social workers who work with victims of trauma can often themselves suffer from
vicarious trauma. In addition, they often encounter verbal harassment and bullying. And the
emotional impact these have on the worker often results in harm to physical health as well.
Social workers may experience fatigue, stress, burnout, increased risk of cardiovascular disease,
and sleep problems, to name just a few of the health issues that can ensue. We have already seen
why harm or risk of harm to health and safety should be compensated; the question now is
whether emotional toll or psychological harm should be compensated as well, independently of
its causal role in the development of harm to physical health and safety or the risk of such harm.
I think it is quite clear that work-related costs to one’s psychological well-being should count as
an occupational burden that warrants compensation, independently of its effects on physical
health. In fact, the concept of an occupational hazard is not limited to harm or the risk of harm to
physical health but rather is concerned with health in general, which includes mental and
psychological health. So while emotional/mental/psychological harm or risk of harm is different
in kind from harm to physical health or safety, we may treat the former in the same way that we
treat the latter. As with risk to physical health and safety, risk of harm to psychological health
113

should also be compensated, independently of whether the psychological harm is realized.
Again this is because the mere possibility of emotional stress and the development of
psychological disorders or neuroses, like the mere possibility of physical injury, can itself affect
the well-being of the individual who undertakes the risk by, for example, leading to increased
worry or anxiety about the future and a lower sense of security.
Lastly, I wish to return to the question concerning frustration mentioned in the
introduction to this section. I’ve listed frustration as one of the burdens of monotonous work on
a production line. If frustration is considered a burden of line work, should the frustration
experienced by scientists when their experiments fail also be considered a burden of the
occupation? My response to this question will be similar to my treatment of the different
remuneration construction workers and gymnasts should receive. Taken in isolation, frustration,
like risk of harm, is a burden that warrants compensation; however, our assessment of the overall
burdensomeness of an occupation will not only take into account its burdens but its benefits as
well. While scientists no doubt often experience great frustrations when their experiments fail,
their line of work also offers them opportunities to experience a deep sense of satisfaction when
their experiments succeed. Moreover, their line of work involves the use of their intellectual
capabilities, which many regard as enjoyable or intrinsically worthwhile. These benefits are
absent in line work in a factory. Hence, while frustration is acknowledged to be a burden of
work in a production line, it will carry greater weight in the overall assessment of the
burdensomeness of line work than in the assessment of the burdensomeness of scientific inquiry.
II THE MARKET
One of the pressing questions facing the Burdens Remuneration Principle is whether and
how burdens can be measured and compared. How, for example, do we determine how much to
114

compensate work that is physically exerting compared to work that is emotionally taxing or work
that is neither physically nor emotionally exhausting but repetitive and tedious? This seems to
suggest to us that the best way of measuring the burdensomeness of an occupation is to find out
how unattractive people find that occupation. This will allow us to avoid having to measure and
compare the different kinds of burdens against each other. However, as we will see, there are
several apparent difficulties with the use of the market as a way of determining the levels of
remuneration needed to compensate various occupations.
Ideally, we can use a market to figure out how unattractive certain occupations are to
people: when the income attached to an occupation is too low, not enough people will choose to
take up that occupation, resulting in a shortage of workers. When the income is raised, open
positions will be filled until the point at which the income is “too high”: in other words, there is a
surplus of willing workers for that particular occupation. But this ideal scenario fails to take into
account many contingencies that interfere with the true measure of how burdensome people will
find a particular occupation. For some occupations, there will be a natural shortage of workers
as a result of the special talents that it requires. To take a familiar example, neurosurgery
requires a high level of skill that few possess. As a result, neurosurgeons will command a higher
income partly in virtue of the fact that few possess the talents required for this line of work. The
higher income attached to neurosurgery will not only reflect the burdens of the occupation but
the natural scarcity of qualified workers as well.
This is a familiar phenomenon in our actual societies that for proponents of meritocracy
are unproblematic. However, luck egalitarians, who maintain that both disadvantages and
advantages that arise from an individual’s unchosen circumstances—such as her talents and other
native endowments—are to be redressed, will find this result unfair. More generally, on the view
115

that talents and other native endowments are morally arbitrary with respect to the distribution of
income and wealth, the inflation of income that results from the natural scarcity of talents
required for neurosurgery will also be found problematic. Moreover, some occupations will
offer levels of income that are lower than are necessary to compensate for the burdens of that
work. Again, this will be partly due to the natural distribution of talents. Some occupations do
not require any special talents and so will naturally be more likely to face a surplus of potential
workers. Even supposing that such an occupation was quite burdensome and unattractive to
many people, individuals may nonetheless be willing to fill the occupation for low pay if their
lack of native talent prevents them from filling other occupations. As a result, the income
attached to occupations requiring little talent may not be enough to compensate individuals for
the burdens they undertake in filling those occupations. Again, for the luck egalitarian, this
result is unfair because individuals suffer disadvantages that result from the unchosen features of
their circumstances—i.e. their lack of talent. According to the Burdens Sensitive view, this
result is problematic because individuals are not adequately compensated for the burdens they
undertake while doing socially necessary or beneficial work; asked to carry greater burdens than
others, such individuals do not enjoy terms of reciprocity and equality.
III CARE
Moreover, the market does not recognize all forms of socially necessary or beneficial
work. Importantly, a vast majority of the socially necessary work of childrearing is neither
considered an occupation nor remunerated as such. But given that it is socially necessary work,
childrearing warrants remuneration in accordance with the burdens it imposes.
While I will not attempt to enumerate all the burdens associated with childrearing here, I
will at least point out that our discussion thus far of the most commonly and widely recognized
116

burdens of various occupations can help us to figure out which aspects of childrearing may be
considered burdens. Childrearing, like other lines of work, can be at times be physically
demanding, emotionally taxing, disrupt natural circadian rhythms, require long and erratic hours,
and can cause a significant amount of stress. Considering that this work is socially necessary
and contributes to the good of society by producing future citizens and workers, individuals who
rear children should be compensated for the burdens they undertake. Moreover, the individual
who rears children must often forego career opportunities and hence realize less of her earnings
potential. This is one of the ways in which being a parent differs from being a child care worker,
such as a teacher, day care worker, or nanny. While there is more to be said about compensation
for childrearing, I would like to at least note that one advantage that the Burdens Compensation
principle has over competing principles such as Rawls’s Difference Principle is that it can be
easily used to explain why work done within the household, such as childrearing, should be a
subject of distributive justice. Feminists who take issue with Rawls’s theory for failing to
include the family as subject to the principles of distributive justice will have motivation to find
the Burdens Sensitive view attractive, a claim that I explore in further depth in the next chapter.
IV A CLOSING NOTE ON EFFICIENCY
While I’ve provided a few familiar arguments against the use of an unrestricted market in
measuring the degrees to which occupations are considered burdensome, the Burdens
Compensation Principle is by no means incompatible with forms of economy that involve a
market. What the principle requires is that if a market is to be utilized in an economy, it must be
appropriately regulated so as to produce outcomes in which individuals are compensated
according to the burdensomeness of their work.
117

The Burdens Compensation Principle states that socially necessary or desirable work
should be compensated. This serves to rule out cases in which we would otherwise be
committed to remunerating burdensome work that is neither socially useful nor desirable. We
wouldn’t, for example, feel the need to remunerate someone who chooses to spend her days in
the utterly useless task of counting the grains of sand on a beach. So social necessity or
desirability provides a necessary requirement for remuneration: the work must be either needed
or wanted by others. Henceforth, I will use the term “social desirability” to stand for both social
necessity and desirability.
One major objection that readers may have is that rather than taking social desirability as
a necessary requirement for remuneration, my theory should incorporate it as a second dimension
along which remuneration can vary: in addition to remunerating for the overall burdensomeness
of work, a principle of distributive equality should also remunerate according to the desirability
of the work. So remuneration varies along two distinct and independent dimensions:
burdensomeness and social desirability.
I would like to first motivate this objection by providing an example that seems to elicit
intuitions in favor of my opponent’s view. Subsequently, I will explain why these intuitions are
at heart motivated primarily by considerations of efficiency. Lastly, I will explain why my
theory rejects such considerations.
First, consider the following example:
The Better Button
Jack and Jill are co-workers. They spend their days pushing a button that triggers a
process that results in diced carrots. Both have the choice to push one of two buttons:
Button A will result in ten carrots being diced; Button B will result in 20 carrots being
diced. Jack pushes button A while Jill pushes button B.

118

While this example is highly artificial, it is nonetheless useful in isolating the factor of efficiency
from other factors such as talents and abilities. Given that Jack and Jill’s work carries exactly
the same burdens, the Burdens Compensation Principle would compensate their work equally.
However, intuitively it seems that Jill should receive greater remuneration.
What are the reasons for our intuition? One of the reasons why we have this intuition
may be that we think that it is better to encourage Jill’s choice so that Jack may follow suit, thus
resulting in greater productivity. Or maybe we think that, relative to Jill’s remuneration, Jack
should be penalized for making a wasteful choice. Or perhaps we straightforwardly think that
greater productivity deserves greater remuneration. Will this be a devastating problem for the
Burdens Compensation Principle?
My answer is to point out that our intuitions are all motivated by considerations of
efficiency. If we wish to encourage Jill’s decision to push button B, it is because doing so may
encourage her to continue to push B and perhaps even coax Jack into pushing B as well, all
ultimately resulting in greater productivity and efficiency. Similarly, if we wish to penalize Jack
for making a wasteful choice, it is because we wish to deter him from continuing to push button
A and guide him toward the more productive choice of pushing button B.
While I do not deny the importance of efficiency in an ultimate judgment of the kind of
society we may wish to actualize, I wish to refrain from incorporating considerations of
efficiency into the Burdens Compensation Principle. Recall that the principle is posed as an
answer to the question “What are the factors relevant, from a moral point of view, to the
distribution of income and wealth?” The value of the Burdens Compensation Principle is that it
can provide an articulation of the morally relevant bases for compensation, thereby elucidating
the conditions under which inequality in economic advantage can be regarded by all as fair.
119

Ultimately, when thinking about the kind of society we wish to actualize, we may decide to
compromise fairness for the sake of efficiency. My discussion of the Burden Sensitive view thus
far is neutral on whether such a compromise will be acceptable and if so, how much of a
compromise is allowable. So, in short, when we consider The Better Button from a strictly moral
point of view and put aside impulses toward efficiency, it becomes less unintuitive that because
there is no difference in the burdensomeness of their work, Jack and Jill should receive equal
remuneration. It is only once we engage in the further stage of theorizing about the kind of
society we wish to have that we should begin to incorporate considerations of efficiency. In
other words, if we are to sacrifice equality for the sake of efficiency, we should at least know
what it is we are sacrificing. To this end, the Burdens Compensation provides an elucidation of
distributive equality from a strictly moral point of view.

120

5 The Burdens of Care and the Reach of Justice

This chapter continues my discussion from Chapter 3 of G. A. Cohen’s charge that an
ambiguity in Rawls’s account of the basic structure makes him vulnerable to a dilemma that
undermines the standard interpretation of the Difference Principle. In response, I develop an
account of the basic structure grounded in Rawls’s identification of its institutions as public
systems of rules. Though informal social conventions and practices can have as significant an
impact on people’s life prospects as formal institutions, some uses of coercive power in
eliminating racist, sexist, or other such unjust attitudes may be unjustifiable in terms of public
reason. This leads to an account of the basic structure as non-arbitrarily limited to formal or
coercive institutions.
I next turn to a feminist criticism of this account of the basic structure. Prevailing gender
norms place greater expectations on women to undertake household labor and the care of
dependent family members. Because these forms of labor and care occur within the private
sphere of the family and not on the market, they are not regulated by the principles of justice.
Hence, it is far from clear that the Difference Principle can accommodate the claims to greater
resources and opportunities of private care-givers, since it regulates only the inequalities of
income and wealth between normally functioning, productive members engaged in social
cooperation. The Burdens Compensation Principle, however, requires that all forms of socially
necessary labor be compensated, thereby acknowledging that private care-giving indeed counts
as contributing to the social good. Moreover, in compensating household labor and care for
dependents, the principle can address the feminist concern that those engaged in private care
work are economically and socially disenfranchised as a result of their absence in the labor
121

market. Lastly, I argue that the principle is compatible with the commitments of political
liberalism. The requirement that household labor and familial care be compensated does not
impose upon citizens any requirement to internally order their families according to substantive
moral claims about the equality between men and women. Thus, both the political liberal and
the feminist have reason to find the Burdens Compensation Principle compelling.
I THE BASIC STRUCTURE AND THE DIVISION OF MORAL LABOR
1 Two Standpoints
Ethics arises in response to the conflict between two standpoints: the personal and
impersonal.
91
From the personal standpoint, individuals take into consideration their various
aims, interests, and values as they deliberate about their actions. In some cases, we need only
consider our own narrow self-interests, but our deliberations about what to do often require us to
take on the impersonal standpoint as well, as many of the actions we undertake have an impact
on others. When we occupy the impersonal standpoint, we identify our interests as just a small
subset of the many competing interests at play. In deliberating about what to do, we will need to
weigh our own personal considerations against those of others and to strike a proper balance
between them when they come into conflict. Such a balancing act will not always come easy, as
it may often be unclear how to weigh the many values and aims that people have. Some values
may be incommensurable, and disagreement may arise about the relative weight of those values
as well as whose values, aims, or interests should be given priority over those of others.
As we arrive at a judgment of what really matters, from an impersonal standpoint, what
matters to us from a personal standpoint can come into conflict with those impersonal values or
aims, and it is not always clear how much personal prerogative we may permissibly exercise in
these cases. To illustrate, Peter Singer has famously argued that it would be morally wrong to

91
Thomas Nagel, for example, takes this to be the central problem of political theory. See Nagel.
122

purchase a luxury item, such as a pair of designer shoes, when one could donate the money to
famine relief instead.
92
From the impersonal standpoint, few would fail to recognize that a
child’s life, for example, matters more than my enjoyment of a fancy pair of shoes. Yet not
everyone is willing to accept Singer’s conclusion that it is morally wrong to refrain from
donating the money to famine relief. They may, for example, maintain that there is a limit to our
individual responsibility for alleviating suffering. Morality, they might argue, allows individuals
a greater scope of personal prerogative than Singer seems to allow, so that it would not be wrong
to prioritize the desires and needs of oneself or one’s close friends and families over those of
distant others. So even if almost everyone agrees with Singer about what matters from an
impersonal standpoint—that is, the alleviation of suffering—disagreement can nonetheless arise
about our individual responsibilities in securing those values.
While the problem of reconciling the personal and impersonal standpoint figures
prominently in ethical deliberation, the problem extends into political theory as well. Liberal
democracies of today are constituted by pluralistic societies whose members possess a wide
range of diverging interests. One of the main tasks of a theory of justice is to provide principles
that strike a fair balance between these competing interests. When applied to the design of the
major social, political, and economic institutions of society, these principles will shape the range
of expectations and aims that individuals acting within those institutions may form and pursue.
2 The Role of the Basic Structure
Theories of justice such as that of John Rawls respond to the conflict between the
personal and impersonal standpoints by proposing a “division of moral labor” between

92
Peter Singer, "Famine, Affluence, and Morality," Philosophy & Public Affairs 1, no. 3 (1972).
123

institutions and individuals in realizing justice.
93
Rawls’s two principles of justice apply to the
basic structure of society: the arrangement of the major social, political, and economic
institutions that distribute the benefits and burdens of social cooperation.
94
In particular, the
basic structure distributes what Rawls calls the social primary goods, “all-purpose means that are
generally necessary to enable citizens adequately to develop and fully exercise their two moral
powers, and to pursue their determinate conceptions of the good;” these all-purpose means
include the basic rights and liberties, opportunities, powers and prerogatives of offices and
positions of authority and responsibility, income and wealth, and the social bases of self-
respect.
95
One of the reasons why the principles should primarily apply to these institutions,
Rawls notes, is that they have a profound and pervasive effect on people’s initial life chances.
96

In distributing the social primary goods, the basic structure provides a framework that both limits
and secures the expectations individuals may have, which in turn shapes the choices they make.
When regulated by just principles, the basic structure goes a long way in securing justice.
Though the principles of justice apply to the basic structure of society, they do not apply
to associations and individuals. It may be objected that a society cannot be fully just unless its
individual members, and not only its institutions, are just as well. How could justice possibly be
achieved when citizens are vicious, for example? The members of a society must also be
subject to rules regarding just conduct, or else even institutions that are initially just will
eventually be destabilized by corruption or lack of compliance.
This objection, however, misses its mark. Rawls’s division of moral labor divides the
task of achieving justice between institutions and citizens in such a way that allows citizens as

93
See, for example, Samuel Scheffler and Véronique Munoz-Dardé, "The Division of Moral Labour," Proceedings
of the Aristotelian Society, Supplementary Volumes, (2005).
94
Rawls, A Theory of Justice, 7.
95
Rawls, Justice as Fairness: A Restatement, 57-59.
96
Rawls, A Theory of Justice, 7.
124

much freedom to pursue their own conceptions of the good as is consistent with the stability of
just institutions. His theory does not exempt individuals from doing their share in supporting
justice. Rather, Rawls holds that while the principles of justice are to be applied to institutions,
individuals and associations are beholden to separate principles.
97
This is because principles that
are appropriate to one domain of society may not be appropriate for another. All citizens may
agree that fair equality of opportunity should be realized in education and the workforce, but it
would be unworkable and undesirable as the governing ideal for, say, friendship. Instead,
individuals are subject to the principle of fairness and various natural duties, such as the duty of
justice, the duty of mutual aid, and the duty not to harm.
98

Moreover, Rawls is careful to note that in order to maintain justice over time, adjustments
within the basic structure must be made. Shifts in the demographic facts about the citizenry,
advancements in technology, and developments in the international arena are just a few examples
of changes that may have destabilizing effects. The appropriate response to such changes is not
to modify our fundamental principles of justice, for we have supposed that they express a
conception of justice that all reasonable citizens could affirm, from one generation to the next.
Rather, the proper response would recommend adjustments in the rules governing the individual
transactions within the institutions that comprise the basic structure.
99
For example,
technological advancements such as the invention of the internet has, in a relatively short period

97
He writes, for instance, “The principles of justice for institutions must not be confused with the principles which
apply to individuals and their actions in particular circumstances. These two kinds of principles apply to different
subjects and must be discussed separately.” Ibid., 54-55.
98
The principle of fairness holds that “a person is required to do his part as defined by the rules of an institution
when two conditions are met: first, the institution is just (or fair), that is, it satisfies the two principles of justice; and
second, one has voluntarily accepted the benefits of the arrangement or taken advantage of the opportunities it offers
to further one’s interests.” Ibid., 111-112. The basic idea is that if we gain from the cooperative efforts of others, we
ought to do our fair share, provided a background of just institutions. The duty of justice “requires us to support and
to comply with just institutions that exist and apply to us.” Ibid., 115. The duty of mutual aid and the duty not to
harm are the familiar duties to help another who is in jeopardy or need, if doing so does not come at excessive cost
to oneself, and to refrain from causing unnecessary suffering, respectively. Ibid., 114.
99
Rawls, Political Liberalism, 284 and 288.
125

of time, allowed for far more efficient ways of disseminating knowledge. To name just one
example, universities may, for relatively little cost, make high-quality courses available for free
to any student who has access to the internet, supporting fair equality of opportunity by removing
economic barriers that might otherwise hinder an economically disadvantaged student from
accessing high-quality education. In other words, this innovation helps to advantage even the
least advantaged.
At the same time, however, some citizens may be burdened by disadvantages that arise
from being “left behind.” Internet service providers may be reluctant to build the infrastructure
necessary for convenient internet access in poor rural areas. Public schools funded by estate
taxes in affluent neighborhoods will have greater resources for providing classes in computer
programming and computer literacy for their students, allowing their students to be better
prepared for the many future occupational opportunities opened up by the internet. And,
moreover, as business, communication, and education increasingly relies on the use of the
internet for everyday transactions, computer literacy and programming ability will increasingly
become necessary basic skills. As a result, a gap in advantages will further widen not only
between developed and underdeveloped countries but between citizens within even developed
countries such as the U.S. as well. The way in which technological advancements may widen
the gap between the advantaged and disadvantaged poses an issue of justice precisely because of
its potential to destabilize the realization of the ideal of fair equality of opportunity. What is
required, then, is not a modification to our ideal of fair equality of opportunity but rather
particular adjustments to our institutions so that the legal enforcement of rules governing the
distribution of internet access will be made consistent with the ideal of fair equality of
opportunity.
126

Limiting the application of the principles of justice to the basic structure allows for the
stability of a conception of justice over time as adjustments to the rules of the institutions that
comprise the basic structure become necessary in response to historical, sociological, and
technological changes. In other words, maintaining the justice of the basic structure from one
generation to the next will require changing the rules that comprise its institutions precisely so
that those institutions will continue to play its role in effectively realizing justice as part of the
basic structure of society. What we have, then, is an institutional division in moral labor
between the basic structure, which is regulated by the principles of justice, and the rules that
constitute the institutions of the basic structure. The basic structure is regulated by the principles
of justice; the institutions that comprise the basic structure, however, are themselves comprised
by their own system of rules, and those rules are governed by a separate set of principles that
must be consistent with, yet distinct from, the principles of justice in order to yield the kind of
flexibility needed to maintain background justice over time.
Moreover, while institutional rules place constraints on the forms of individual
transactions that are permissible within the institution, the principles that apply to institutional
rules will be distinct from the principles that apply to individual conduct. While it is true that
institutional rules impact individual conduct insofar as they constrain the range of permissible
forms of interaction, those institutional rules do not internally order the private lives of
individuals. To illustrate, principles that apply to the rules of an institution like residential
private property may require that landlords be prohibited from rejecting potential tenants based
on religious identity. However, those principles do not also apply to individuals’ private lives.
They do not, for example, also prohibit landlords from rejecting potential friends or other non-
professional associates for religious reasons.
127

Rather, citizens of a well-ordered society are required to comply with and support just
institutions. When effectively regulated by principles of justice, the basic structure provides “the
background social framework within which the activities of associations and individuals take
place,… secur[ing] what we may call background justice.”
100
Although the principles of justice
do not apply to the private lives of individuals, Rawls’s theory of justice is not committed to the
claim that a society whose institutions are just but whose citizens are vicious is still just. Rather,
citizens will be considered unjust in light of a political conception of justice insofar as they fail
to uphold their duty of justice or their duty of civility.
It may be objected that the moral and political requirements that apply to individuals are
insufficiently robust and fail to capture our intuitive notion of what justice requires of citizens.
In the third chapter, we’ve seen one variant of Cohen’s objection to the standard interpretation of
Rawls’s difference principle. Truly just citizens, Cohen argues, would affirm the principles of
justice. Rather than merely complying with the economic institutions selected by the difference
principle, just citizens will apply the difference principle to the economic transactions they
partake in, with the aim of improving the position of the least advantaged. As a result, they will
be willing to forgo incentives for their productivity.
101

In response to G.A. Cohen’s charge that just citizens who affirm the principles of justice
would shape their choices and behavior according to the difference principle, Rawlsians may
point out that in justice as fairness, the difference principle applies only to institutions and not to
individual choices and behavior. This is what Cohen calls the basic structure objection.
Principles of justice apply to the basic structure of society but do not apply to individuals.
Hence, individuals are not required to apply the difference principle to their everyday behavior

100
Rawls, Justice as Fairness: A Restatement, 3.
101
See Cohen, "Where the Action Is: On the Site of Distributive Justice."; Cohen, "Incentives, Inequality, and
Community."; Cohen, Rescuing Justice and Equality
128

and are free to choose whatever courses of action are allowed by the rules of the basic structure.
So long as they support and comply with the rules of just institutions, they’ve done their part in
securing justice.
102

3 The Boundaries of the Basic Structure
In reply, Cohen contends that there is an ambiguity in Rawls’s definition of the basic
structure. Proponents of the basic structure objection take the basic structure to be confined to
the major economic, social, and political institutions whose rules are legally coercive. And while
the principles of justice apply to these legally coercive institutions, they do not apply to the
individual transactions that occur within those institutions. In other words, individuals are free to
choose however they may, so long as they comply with the rules of the basic structure.
103

However, Cohen notes, we might instead interpret the basic structure to include more than just
those institutions whose rules are legally coercive. One of the reasons why Rawls takes the basic
structure to be the primary subject of justice is that the arrangement of the major social, political,
and economic institutions has a profound and pervasive effect on people’s initial life chances.
Yet, Cohen contends, informal social conventions and practices also have profound and
pervasive effects on citizens’ initial life chances and quality of life. For example, pervasive
racial and gender stereotypes can have a significant impact on citizen’s educational,
occupational, and political opportunities, their access to quality health care and their life
expectancies, and the fair value of their various rights and liberties, such as freedom of
movement.

102
Versions of this objection appear in A. J. Julius, "Basic Structure and the Value of Equality," Philosophy &
Public Affairs 31, no. 4 (2003)., Thomas W. Pogge, "On the Site of Distributive Justice: Reflections on Cohen and
Murphy," Philosophy & Public Affairs 29, no. 2 (2000)., Kok-Chor Tan, "Justice and Personal Pursuits," The
Journal of Philosophy 101, no. 7 (2004)., and Andrew Williams, "Incentives, Inequality, and Publicity," Philosophy
& Public Affairs 27, no. 3 (1998).
103
Granted that they do not violate their natural duties.
129

If Rawls applies the principles of justice to the institutions that have a profound and
pervasive effect on people’s initial life chances, then, Cohen reasons, we should expect that the
informal social conventions and practices that also have a profound impact should be subject to
the principles of justice as well. A dilemma then arises: either the basic structure is comprised
solely of legally coercive institutions or it includes whatever institutions—both form and
informal—that have a profound and pervasive effect on people’s initial life chances. To assert
the former disjunct would be to stipulate an arbitrary definition of the basic structure. If the
basic structure is confined to institutions whose rules are legally coercive and excludes informal
social practices and conventions, then the principles of justice will regulate only some of the
institutions that deeply impact people’s initial chances and quality of life while excluding others.
But defining the basic structure as comprised of all institutions that have a profound and
pervasive effect will require the inclusion of the individual behaviors that give rise to many of
our social practices and conventions, thus undermining the standard interpretation of the
difference principle as well as the so-called division of moral labor. As a result, either Rawls
must accept an arbitrarily stipulated definition of the basic structure of society or accept that the
principles of justice apply to citizen’s private lives as well.
In the next section of this chapter, I will respond to Cohen’s dilemma by motivating an
account of the basic structure that elaborates on Rawls’s characterization of the institutions that
comprise the basic structure as public systems of rules. Cohen’s dilemma relies on the claim that
the following two definitions of the basic structure are mutually exclusive and exhaustive: either
(1) the basic structure is solely comprised of legally coercive institutions, or (2) the basic
structure encompasses all forms of social organization that have a profound and pervasive impact
on citizens’ initial life chances. But the dichotomy he presents is false; a plausible argument can
130

be made that the basic structure is comprised of those forms of social organization whose rules
are public. That is to say, the basic structure consists in those forms of organization whose rules,
policies, or laws require justification in terms of public reason. Moreover, although my account
of the basic structure will exclude many of the informal conventions that also have a profound
effect on citizens’ life chances, the definition of the basic structure is not arbitrarily stipulated but
rather motivated from core commitments of Rawls’s political liberalism. In the following
section, I will turn my attention to some of the salient aspects of what Rawls identifies as the
circumstances of justice: a public political culture in which the citizen is identified as free and
equal, the fact of reasonable pluralism as a permanent feature of modern democratic institutions,
and the burdens of judgment. These features, I argue, will help us understand how Rawls may
draw the boundaries of the basic structure in such a way that non-arbitrarily excludes some of
our informal social practices and conventions, despite the fact that these informal institutions
also have a profound and pervasive effect on people’s lives.
II THE INSTITUTION AS A PUBLIC SYSTEM OF RULES
As noted earlier, Rawls describes the basic structure as “a society’s main political, social,
and economic institutions…”.
104
The account I wish to develop stems from Rawls’s description
of the institutions that comprise the basic structure as public systems of rules. Take, for example,
the following passage from A Theory of Justice:
Now by an institution I shall understand a public system of rules which defines offices
and positions with their rights and duties, powers and immunities, and the like. These
rules specify certain forms of action as permissible, and others as forbidden; and they
provide for certain penalties and defenses, and so on, when violations occur.
105

A natural interpretation of this passage may be that the institutions that comprise the basic
structure are simply those that are legally coercive. After all, Rawls understands institutions as

104
Rawls, Political Liberalism., 11.
105
Rawls, A Theory of Justice, 55.
131

public systems of rules that permit or forbid certain forms of action and specify the penalties,
should violations occur. However, the compatibility between the principles of justice and an
institution as a system of rules is not sufficient for the justice of that institution. Rather, Rawls
subsequently notes that it is only the institution as realized that may be properly called just or
unjust. He writes,
An institution may be thought of in two ways: first as an abstract object, that is, as a
possible form of conduct expressed by a system of rules; and second, as the realization in
the thought and conduct of certain persons at a certain time and place of the actions
specified by these rules. There is an ambiguity, then, as to which is just or unjust, the
institution as realized or the institution as an abstract object. It seems best to say that it is
the institution as realized and effectively and impartially administered which is just or
unjust. The institution as an abstract object is just or unjust in the sense that any
realization of it would be just or unjust.
106

Considered abstractly, the rules of an institution are not themselves sufficient for the justice of
the institution. Rather, it is only when the rules effectively bring about certain forms of behavior
that the institution may be said to be just. Rules designed for morally perfect beings—angels,
let’s say—may express the values and ideals that we associate with justice. We may even be
tempted to call them ‘just’ rules. Most likely, though, such rules will fail as an institution when
imposed on ordinary folk who lack the full impartiality of angels. Given its failure to be
effectively realized, we should hesitate to call the system of rules just. It is the system of rules as
realized that is just or unjust.
This leaves us with the following picture. The principles of justice apply, in the first
instance, to the basic structure of society, the way in which the major institutions of society fit
together to distribute the benefits and burdens of social cooperation. These institutions are
public systems of rules that are deemed just when their realization is consistent with the
principles of justice. In other words, once such rules are imposed on reasonable and rational

106
Ibid.
132

persons who possess the two moral powers requisite for citizenship in a well-ordered society, we
may then see whether there arises certain forms of behavior or conduct conducive to or
consistent with the principles of justice.
Indeed, Rawls’s description of the basic structure in his later work Political Liberalism
acknowledges the impact that the design of the basic structure has on the character of its
citizenry:
The initial focus, then, of a political conception of justice is the framework of basic
institutions and the principles, standards, and precepts that apply to it, as well as how
those norms are to be expressed in the character and attitudes of the members of society
who realize its ideals.
107

Later in the same work, Rawls re-emphasizes that the basic structure “in fundamental ways
shapes citizens’ character and aims, the kinds of persons they are and aspire to be.”
108
It is
inaccurate, then, to interpret Rawls as holding the view that individual conduct is entirely
unregulated by principles of justice. Rather, while the principles of justice govern the selection
of the major institutions that make up background framework against which individuals are free
to pursue their own conception of the good, citizens nonetheless play a significant part in
realizing justice insofar as their conduct under the rules of the institutions to which the principles
of justice apply collectively determine the justness or unjustness of the institution.
This is not to say, however, that individual conduct, across the board, is subject to the
principles of justice. In what follows, I will argue that proponents of political liberalism can
argue that the reach of justice is non-arbitrarily limited by the facts that comprise the
circumstances of justice—in particular, the fact of reasonable pluralism, the political conception
of the person (i.e. the concept of the citizen), and the burdens of judgment.
1 The Circumstances of Justice

107
Rawls, Political Liberalism, 11.
108
Ibid., 68.
133

For Rawls, a conception of justice must be realistically utopian—it should “probe the
limits of what’s practically possible.”
109
For modern democratic societies, what’s practically
possible will be framed by the circumstances of justices. Rawls divides the circumstances of
justice into two groups: the fact of moderate scarcity and the necessity for social cooperation
that results,
110
on the one hand, and the fact of reasonable pluralism, on the other. The fact of
reasonable pluralism is the fact that in a modern democratic society, citizens affirm “different,
and indeed incommensurable and irreconcilable, though reasonable, comprehensive doctrines in
the light of which they understand their conceptions of the good”.
111
The fact of reasonable
pluralism arises from a certain normative conception of the citizen together with what Rawls
calls the burdens of judgment. I turn to each of these concepts next.
2 Citizenship
The fact of moderate scarcity leads to the necessity of social cooperation. A theory of
justice that seeks to specify just terms of social cooperation must first settle on a specification of
the relevant features of the social cooperators.
First, in a modern democratic society, people are considered free and equal.
112
For
Rawls, their freedom and equality are grounded in their two moral powers:
The capacity for a sense of justice
Citizens have the capacity to understand, to apply, and to act from (and not merely in accordance
with) the principles of justice.
The capacity for a conception of the good

109
Rawls, Justice as Fairness: A Restatement, 4.
110
In addition to the fact of moderate scarcity are certain assumptions about what makes human cooperation
possible: people live together within a geographical territory and are similar enough in their physical and mental
powers that no single individual can, through those powers alone, dominate the rest. See Rawls, A Theory of
Justice, 126.
111
Rawls, Justice as Fairness: A Restatement, 84.
112
This conception of persons can be found in the public political cultures of democratic societies, as contained in
their constitutions and other basic political texts as well as in historical interpretations of those texts; ibid., 19-20.
134

Citizens have the capacity to have, revise, and rationally pursue a conception of the good.
113

By possessing a capacity for a sense of justice, citizens are reasonable. They are prepared to
propose, acknowledge, and affirm fair terms of social cooperation, even when such terms result
in fewer advantages for them when compared to unfair systems of social coordination.
114
And
their capacity for a conception of the good grounds our conception of citizens as rational.
Citizens are also conceived of as free and equal. They are equal in virtue of having the
moral powers to the extent required for being fully cooperating, normally functioning members
of society.
115
And they are free in virtue of having these two moral powers. Their legal status as
citizens does not depend on their determinate conception of the good; hence, they are free to
exercise their capacity for a conception of the good and to revise those conceptions without
change in their public identity or their status as citizens. They also have a higher-order interest
in the exercise of these powers, and thus citizens in a democratic society think themselves as free
in virtue of being self-authenticating sources of valid claims. In having the capacity to revise
their own conception of the good and an interest in its exercise, they view themselves as entitled
to making certain claims upon society—claims, for example, of non-interference in certain areas
of their lives or for assistance in pursuing their conception of the good. These claims will be
satisfied through their guarantee of primary social goods, which are to be distributed according
to the principles of justice.
Primary social goods are conceived of as the needs of citizens qua citizens; they are not
to be mistaken as the needs of moral persons associated with any particular comprehensive
doctrine, and though some individuals might not need, for example, income and wealth to pursue

113
See ibid., 18-19.
114
For example, a reasonable citizen affirms principles of justice that reject caste systems, even if she would have
enjoyed greater advantages under such a system. Of course, this is assuming that others are also reasonable and
affirm fair terms of social cooperation.
115
Rawls, Political Liberalism, 19.
135

their conception of the good—say, the spiritual devotion that can only be obtained through self-
abnegation and asceticism—the social primary goods are nonetheless necessary for their
citizenship. One aspect of the political freedom of citizens is that their citizenship status is
independent of whatever comprehensive doctrine they may hold. Many individuals continually
revise their conceptions of the good over the course of their lives, and some even undergo radical
transformations in their sense of personal identity as they enter or leave various communities and
acquire or discard comprehensive doctrines. Such changes are made possible by the fact that
their legal status as citizens remain unchanged and that just institutions secure for all citizens the
goods that any rational citizen would want as means for pursuing their own conception of the
good. The ascetic, for example, is, as a citizen, free in virtue of being guaranteed the primary
social goods even if those goods are not identified as needs from within her particular
comprehensive doctrine. Were she to come to hold and pursue a different conception of the
good, her claim as a citizen to the social primary goods would not change.
We have, so far, an account of the citizen as free and equal, and reasonable and rational
in virtue of possessing the capacity for a sense of justice and a capacity for the conception of the
good as well as a higher-order interest in the exercise of those capacities. These aspects of
Rawls’s political conception of the citizen lead us to recognize the burdens of judgment and the
inevitable fact of reasonable pluralism.
3 The Burdens of Judgment and the Fact of Reasonable Pluralism
My claim will be that many areas of our personal lives will be outside of the scope of the
principles of justice, precisely for the reason that facts about the circumstances of justice—for
example, the kinds of psychological facts that parties to the original position are said to have
knowledge of—explain why it would be unreasonable to expect us to accept a conception of
136

justice that requires us to order certain aspects of our lives according to the priorities and aims of
justice. What Rawls calls the burdens of judgment, for example, are a paradigm instance of such
facts. Even reasonable persons may find themselves entrenched in deep and intractable
disagreements. This is because our judgments concerning matters of truth, value, and the
comparative weight of those values are often deeply shaped by our experiences.
116
For example,
someone who has suffered traumatic experiences during war may come to place more weight on
the value of security compared to privacy and be more tolerant of government surveillance. A
civil rights activist, on the other hand, may place more weight on the value of privacy as
necessary to her full and effective exercise of her freedoms of speech, conscience, and
association. At other times, intractable disagreements arise as a result of the complexity or
indeterminacy of the cases at hand. Whatever empirical evidence there is relevant to a case may
be conflicting or difficult to assess. The concepts involved in the case may themselves be vague
and require interpretation.
117
Or it may be difficult to settle the disagreement because normative
considerations of varying force lie on either side of the debate.
118

In a democracy in which reasonable and rational citizens are granted the full exercise of
their rights and liberties, we cannot reasonably expect citizens to come to agree on matters of
truth and value. It may very well be that some fall on the correct side of the debate, and that
there is a matter of fact about whether security is a weightier value than privacy. However,
given the diversity of our life experiences, and the difficulties inherent in our long-standing
disagreements, even reasonable and rational persons who sincerely reflect on the various
empirical and normative considerations involved may find themselves unable to converge on the

116
Rawls, Justice as Fairness: A Restatement, 35.
117
Ibid.
118
Ibid., 36.
137

truth. As free and equal, citizens of democratic societies will inevitably differ in the conceptions
of the good that they form and pursue. This is the fact of reasonable pluralism.
For Rawls, the fact of reasonable pluralism is not to be regretted, for it arises from the
exercise of human reason under conditions of liberty. Given the burdens of judgment, citizens
who exercise their basic rights and liberties—such as those of conscience, association, and
speech—will come to form their own judgments on substantive matters of truth and value.
Because a society cannot be considered free and democratic unless it secures these basic liberties
for its citizens, the fact of reasonable pluralism is permanent and eliminable only through the
oppression of citizens’ two moral powers.
4 The Liberal Principle of Legitimacy
In a modern democracy, a conception of justice must provide an account of how a society
may achieve stability, despite the permanent fact of reasonable pluralism. Conceived of as
having a higher order interest in the exercise of their two moral powers, citizens have an interest
in upholding and supporting just institutions that secure the primary social goods they need in
order to pursue their own conception of the good. Institutions that conflict with that higher-order
interest, however, are unlikely to gain the support and affirmation of citizens who view
themselves as self-authenticating sources of valid claims upon society for assistance or non-
interference in exercising their two moral powers. The imposition of a state religion, for
example, is not likely to achieve stability in a modern democracy in which the citizen is
conceived of as having a higher-order interest in her freedom and equality. Given the burdens of
judgment and the guarantee of the basic rights and liberties, such as freedom of conscience,
speech, and association, citizens will inevitably come to affirm different comprehensive
doctrines from which they form their conception of the good. In other words, some citizens will
138

embrace reasonable comprehensive doctrines that are incompatible with the state religion.
Despite the incompatibility, citizens will nonetheless view themselves as entitled to the pursuit of
the conception of the good that arises from the reasonable comprehensive doctrines they hold.
Hence, as a permanent feature of free democratic institutions, the fact of reasonable
pluralism entails that a conception of justice whose justification rests on any particular
comprehensive doctrine cannot be the subject of an overlapping consensus.
119
This follows from
the assumption that citizens are free and equal in virtue of their two moral powers and that any
democratic society must secure basic rights and liberties such as freedom of speech, conscience,
and association, together with a recognition of the burdens of judgment. The major institutions
of society that distribute the social primary goods, then, must be justifiable in light of the
common human reason that citizens share; their justifications must not rest on any particular
comprehensive doctrine. These reflections lead to a conception of the conditions under which
the use of coercive power is legitimate:
The Liberal Principle of Legitimacy
Our exercise of political power is fully proper only when it is exercised in accordance with a
constitution the essentials of which all citizens as free and equal may reasonably be expected to
endorse in the light of principles and ideals acceptable to their common human reason.
120

It is not immediately clear what constitutes common human reason, but given Rawls’s
conception of the citizen, we may suppose that the reason of the citizen qua citizen is one
articulation of the notion of a common human reason in light of which citizens can reasonably be
accepted to endorse their constitution. All citizens agree upon the political freedom and equality

119
Rawls argues that by appealing only to reasons that are to be found in the public political culture of a modern
democratic society, a political conception of justice is freestanding and as such can gain the adherence of citizens
who hold a wide range of different reasonable comprehensive doctrines. Rawls, Political Liberalism, 144-150.
120
Ibid., 137.
139

of all citizens. That is to say, an individual recognizes that despite whatever reasonable
comprehensive doctrine she holds or rejects, her legal status as a citizen remains unaffected. She
also recognizes that others who hold incompatible doctrines nonetheless enjoy an equal status of
citizenship, no matter how wrong about the truth or value this other person may be. This is
because she recognizes the burdens of judgment and that, as citizens, others are just as entitled as
she is in making claims upon society for assistance or non-interference in the exercise of the two
moral powers. The use of coercive power, then, is only fully proper when it is justifiable in
terms that all citizens would accept. And all citizens will accept justifications that appeal to the
protection of their two moral powers, in accordance with their higher-order interest in their
freedom and equality.
5 Public and Non-public Spheres
The major institutions of society have a profound and pervasive effect on people’s initial
life chances. This is one reason why they should be subject to the principles of justice. To
exempt the economy from the principles of justice, for example, would be to allow distributions
of income and wealth that might unequally enable citizens in their exercise of the two moral
powers. A laissez-faire capitalist economy might yield vast economic inequalities under which
the worst-off do not earn enough for the satisfaction of their basic needs or of the needs of their
dependents. The worst-off under such a system do not have the material conditions that would
enable them to effectively exercise their capacity for a conception of the good, for the exercise of
this capacity already assumes the satisfaction of basic human needs such as food, shelter, and
basic medical care. At best, their capacity for a sense of justice is also undermined; it would be
unreasonable to expect them to be motivated to affirm a system under which their basic needs are
not met. Hence, the application of the principles of justice to the major institutions of society is
140

justified in light of the fact that the arrangement and design of those institutions enable or disable
citizens’ exercise of their moral powers. If left beyond the reach of justice, those institutions
would affect citizens’ life chances in such a way that conflicts with their higher order interest in
the exercise of their two moral powers.
Other institutions, however, may have a profound effect on people’s life chances and yet
remain only indirectly subject to the principles of justice. Consider, for example, a society in
which, although there is no official state religion, a large majority of members share a common
religion. Suppose also that their religious doctrine prohibits women from occupying positions of
religious leadership. This results in an inequality between the range of occupations open to men
and those open to women. Without state interference, women will be unable to effectively
pursue a plan of life that includes religious leadership. This fact may even imply an inequality in
men and women’s freedom in exercising their capacity for a conception of the good. If this is so,
this fact may provide a pro tanto reason in favor of state interference, for the doctrine may be
one that conflicts with some citizen’s (namely, women’s) higher order interest in exercising the
capacity to pursue one’s own conception of the good. However, citizens may acknowledge other
reasons against interference, again appealing to their higher order interest in maintaining their
two moral powers. If the state mandates that positions of religious leadership be open to all
under equality of opportunity, then the state restricts the expression of a religious doctrine that
belongs to a reasonable comprehensive doctrine.
121
We have, then, a conflict between some
individuals’ interest in state interference and others’ interest in non-interference. However, the
reasons against state interference outweigh the reasons in favor of it. Even if state non-
interference results in an inequality between men and women’s exercise of their moral powers,

121
Reasonable, that is, insofar as the religion does not deny the political freedom and equality of citizens; it does
not, for example, prohibit women from voting or hold that members of a particular race should not have the same
political rights as others.
141

state interference would result in an even more severe curtailment of citizens’ religious freedom.
Though some women will have an interest in equality of opportunity for positions of religious
leadership, those women, as well as others, will have a more fundamental interest in the freedom
to follow the prescriptions issued forth by their religious doctrines. State interference threatens
this freedom and so will not be justifiable to citizens who view themselves as entitled to making
claims upon society to protect their capacity to pursue their own conception of the good.
We see, then, that a non-arbitrary distinction can be made between the institutions that
are directly subject to the principles of justice and those that aren’t. The distinction is guided by
what’s required in order to protect citizens’ higher-order interest in the exercise of their two
moral powers: in some cases, state interference is necessary for this protection, and in other
cases, what is required is non-interference. This distinction allows us to demarcate the
boundaries of the public domain: institutions for which state interference is required in order to
protect citizens’ higher order interest are public. These institutions constitute exercises of
political power that are subject to the liberal principle of legitimacy. They must be justifiable in
terms of public reason—in terms, that is, that draw on the concepts to be found in the public
political culture and that appeal to the interests of the citizen qua citizen. On the other hand,
certain domains of life require non-interference in order to allow sufficient room for citizens to
pursue their own conception of the good. Institutions, associations, and conventions that occupy
this domain are non-public. They are not directly subject to the principles of justice, and their
internal workings are not constrained by the Liberal Principle of Legitimacy.
122
Churches, for
example, need not offer justifications for their hierarchies of authority in terms of public reason.

122
This is not to claim, however, that they are entirely “beyond the reach of justice.” The principles of justice may
still indirectly apply to non-public associations insofar as the principles of justice apply to the basic structure, which
sets constraints or limits on the associations and individual transactions that occur within them. Though the
principles of justice do not internally order positions of authority within churches, the basic structure provides
142

Thus far we have the conceptual resources for defending an account of the basic structure
that non-arbitrarily excludes some social practices, despite the fact that they too have a profound
and pervasive effect on individuals’ initial life chances. Citizens have a higher order interest in
the exercise of their two moral powers and view themselves as entitled to lay claims upon society
for assistance or non-interference. In accordance with citizens’ interests, society must establish
institutions that provide citizens with both a fair distribution of the social primary goods and
non-public or private space within which to utilize those goods in order to pursue their own
conception of the good.
Recall that Cohen had noted an ambiguity in Rawls’s account of the basic structure:
either it includes only the legally coercive institutions or it includes whatever has a profound and
pervasive effect on people’s lives. Moreover, Cohen argued, this ambiguity leads to a dilemma.
If the basic structure includes only the legally coercive institutions, then the account is arbitrary:
there is no good reason to suppose that only legally coercive institutions are subject to the
principles of justice when so many other, non-legal institutions have, in some cases, just as
profound and pervasive effect on people’s lives. But if the basic structure should include all
institutions that have a profound and pervasive effect on people’s lives, then the difference
principle should result in virtually unqualified equality as individuals apply it to their daily lives
and engage only in economic transactions that maximally benefit the least advantaged.
I argue that this is a false dilemma. We may understand the basic structure as comprised,
as Rawls puts it, of institutions whose rules are public. They are, in other words, institutions
whose rules are subject to the Liberal Principle of Legitimacy insofar as those rules are exercises
of political (coercive) power. As subject to the Liberal Principle of Legitimacy, they must be

background limits on the permissible forms of action that churches may undertake. They may not, for example, use
force to prevent apostasy. See Rawls, Justice as Fairness: A Restatement, 166.
143

justified by terms that all citizens can reasonably be expected to endorse—justified, that is, in
terms of public reason. This restriction sets limits on the boundaries of the basic structure. State
interference in some areas of our lives will not be justifiable in terms of public reason. In such
cases, interference conflicts with citizens’ higher-order interest in the exercise of their two moral
powers. Thus, although many of our social practices and informal conventions have a profound
and pervasive effect on people’s lives, state interference would conflict with citizens’ interest in
the exercise of their moral powers. Hence, such practices and conventions belong to the non-
public domain and are not directly subject to the principles of justice. To conclude, the basic
structure is confined to, roughly, the legally coercive institutions. Yet, this account is by no
means arbitrary but rather the upshot of Rawls’s conception of the citizen as free and equal in
virtue of their two moral powers, the fact of reasonable pluralism as a permanent feature of
modern democratic societies, and the burdens of judgment. It is the circumstances of justice that
lead to our distinction between the public and non-public realm and the so-called division of
moral labor between them.
III THE FAMILY
The family is included in the basic structure insofar as there are public rules regarding the
composition, creation, and dissolution of families. Laws serve to define the institutions of
marriage and divorce, and they prohibit the neglect and abuse of children and mandate their
education. On Rawls’s theory, the principles of justice regulate the form of family and establish
laws that protect the political freedom and equality of its members. In its role as producing the
next generation of citizens, the family is subject to restrictions intended to ensure that it
continues to play its role in maintaining inter-generational justice. The education of children is
mandated, for example, as a necessary condition for their future status as citizens whose freedom
144

and equality consist in their exercise of their moral powers. The family, however, is also a
private entity. Certain aspects of familial life are not to be intruded upon by the state. For many
people, the freedom to internally order family life according to one’s own reasonable
comprehensive doctrines crucial to their conception of the good. So although the family as a
public institution is subject to the principles of justice, many inter-family dynamics are non-
public and hence only indirectly subject to the principles.
In this section, I briefly survey some of the major objections put forth by feminist
philosophers who criticize the liberal distinction between the family as a public institution and as
a private entity. By exempting certain inter-family dynamics from the principles of justice, they
argue, liberal theories are unable to address the ongoing social injustices that originate in and are
perpetuated by families organized according to sexist comprehensive doctrines. I explain why
there are strong reasons to reject the proposal to subject inter-familial policies to the principles of
justice. The reasons I offer are in line with Rawls’s arguments in favor of drawing a distinction
between the public and non-public domains. I then present Rawls’s response to such feminist
criticisms before pointing out several reasons why the feminist may continue to find his response
unsatisfying particularly when we consider the claims of those who care for the permanently
disabled. Lastly, I argue that feminists will have strong reason to find the Burdens
Compensation Principle to be an attractive alternative to Rawls’s Difference Principle.
1 The Feminist Challenge
We’ve seen how political liberalism justifies a division between the public and non-
public aspects of the family. State interference in the family is, to a certain extent, necessary in
order to ensure that the family plays its role in producing the next generation of citizens and to
protect the status of citizenship of the members of a family. Hence, laws define who and what
145

counts as a family, when a member of the family can leave, and what the minimal standards of
care for dependents are. But the informal social dynamics within families, such as the division
of household labor between the spouses, and whether sons and daughters are equally encouraged
in the development of their faculties also significantly support or thwart the realization of justice
in the institutions that comprise the basic structure. Even if public schools provide the same
instruction for girls as they provide for boys, the educational opportunities of girls and boys will
be unequal if families were to raise girls in such a way that discouraged the development of
qualities, abilities, and ambitions that are encouraged in boys. Similarly, even if laws are
instituted against gender discrimination in the workplace, men and women will have unequal
occupational opportunities if women are expected to take on most of the household labor and are
the primary care-givers for children and elderly parents. They’ll have less time, for example, to
put their efforts into career-advancement projects (such as taking night classes towards a
professional degree) and to engage in after-hours networking opportunities, and they’ll be more
likely to take days off from work in order to care for a sick child. These are some familiar
examples of how inter-family dynamics can have a significant impact on the prospects of those
who have been traditionally assigned gender roles involving household labor and responsibilities
of care.
Most notably, there are two ways in which sexist inter-family dynamics may threaten the
stability of a conception of justice. First, to use Rawlsian terminology, if children who are raised
in sexist households are unable to sufficiently develop the two moral powers necessary for
citizenship, then the possibility of maintaining justice from one generation to the next is
seriously endangered. Second, even if children who are raised in sexist households nonetheless
develop the two moral powers, fair equality of opportunity may be undermined if girls do not
146

form ambitions for higher offices and positions, such as those involving management of
personnel, political influence, or the dispersal of knowledge.
The liberal division between the public and non-public spheres is often targeted by
feminists as fundamentally problematic for reasons such as the two just noted.
123
The above
examples show how some sources of inequality between the sexes (such as intra-family social
dynamics) are beyond the reach of principles of justice. Though principles of justice constrain
the forms of family that people may have and set a minimal standard of care for children, inter-
family social dynamics are regarded as part of the private sphere and to be ordered according to
people’s reasonable comprehensive doctrines. Yet, informal social dynamics are undeniably a
major source of gender-based inequality. One general feminist criticism is that even if all public
institutions are effectively regulated by principles of justice, intuitively, a society will fail to
realize justice if women continue to be disadvantaged in comparison to men as a result of
oppressive gender norms and expectations. In the following, I will focus on the deeper worry
that oppressive gender norms may undermine citizenship by preventing some from individuals
from properly developing their moral powers.
2 Oppressive Gender Norms and Citizenship

123
For a classic overview of several varieties of feminist objections to liberalism, as well as versions of liberal
feminism, see Alison M Jaggar, Feminist Politics and Human Nature (Philosophy and Society) (Rowman &
Littlefield Publishers, 1988). Carole Pateman argues that in sequestering the inner workings of the family to the
private realm, political liberalism places many problems concerning gender inequality outside the reach of justice;
see Carole Pateman, The Sexual Contract (Stanford University Press, 1988). Munoz-Dardé
argues
that
unless
ties

of
sentiment
between
family
members
are
assumed
in
the
original
position,
Rawls’s
theory
will
be unable to
address inequalities between family members; see Véronique Munoz ‐Dardé, "Rawls, Justice in the Family and
Justice of the Family," The Philosophical Quarterly 48, no. 192 (1998). Iris Marion Young maintains that non-
distributive, feminist concerns are left unaddressed by Rawls’s theory as a result of its focus on the basic structure.
Her position, however, is that the proper response is not to conclude that the principles of justice should apply to
more than the basic structure but rather to focus attention on the further work that may be done in demonstrating
how the basic structure may meet these concerns. See Iris Marion Young, "Taking the Basic Structure Seriously,"
Perspectives on Politics 4, no. 1 (2006).
147

We’ve seen that a non-arbitrary account of the basic structure can be defended by
appealing to some of the fundamental concepts of Rawls’s political liberalism, such as the
political conception of the citizen, the burdens of judgment, and the fact of reasonable pluralism.
Given that citizens have a higher-order interest in the exercise of their moral powers and that
citizens are unlikely to converge on the truth of any particular reasonable comprehensive
doctrine, political liberalism identifies only those exercises of political power that can be
justifiable in terms of public reason as legitimate. This sets the boundaries of the basic structure,
since some domains of people’s lives must be protected as private if they are to have effective
freedoms to pursue their own conception of the good.
Sexist doctrines and oppressive gender norms that are enacted in the family, however,
will prove problematic for political liberalism if they prevent individuals from developing the
moral powers necessary for citizenship. In the following, I will consider the conditions under
which sexist doctrines may obstruct the development of a capacity for a conception of the good
or the capacity for a sense of justice. As we will see, Rawls will have a response for each of
these cases. However, in a subsequent section, I will point out one case in particular that will
prove a problematic obstacle to some care givers’ exercise of their capacity for a conception of
the good. The proper response, however, is not to expand the basic structure to include formerly
private areas of family life but rather to ensure that the labor that care givers undertake is
compensated.
Consider, first, the conditions that may disable girls from developing a capacity for a
conception of the good. These will be conditions under which girls fail to develop the basic
reasoning skills or the self-confidence needed to form, pursue, or revise a plan of life. If girls are
denied an education, then it is likely that they will not properly develop the rational capacities
148

needed for capacity for a conception of the good. Sexist doctrines that prohibit girls’ education,
for example, will obstruct the development and exercise of the two moral powers needed for
citizenship.
However, doctrines that prohibit girls’ education will not be considered reasonable since
they deny the political equality of all citizens (or future citizens). Affirming the political equality
of all citizens involves the recognition that all citizens are free and equal in virtue of having and
exercising their two moral powers. The view that girls are not to develop the capacities requisite
for forming and pursuing a conception of the good denies the political equality of all. This case
will not prove problematic for political liberalism because although state-mandated education for
all children encroaches upon some unreasonable individuals’ pursuit of their own conception of
the good, that conception of the good stems from an unreasonable comprehensive doctrine that is
incompatible with the political equality and freedom of all citizens.
Consider, next, whether the enactment of traditional gender roles within the family
hinders boys’ development of a capacity for a sense of justice. A boy who is raised in a
household in which the mother plays a subservient role to her husband may come to internalize
deep-seated beliefs and expectations about the proper roles of women. Such beliefs may even be
largely sub-conscious and so difficult to extricate from oneself. While this case does suggest that
the enactment of gender roles in the family often perpetuates gender inequalities, it is far from
clear that boys who grow up in such households are unable to develop the sense of justice
requisite for citizenship. We may ask, for example, whether such a boy would later, as an adult,
deny that women, too, are citizens and as such are to enjoy the equal basic liberties; it seems
highly unlikely that he will deny the fundamental political equality of women.
149

In reply to the general feminist criticism that in excluding the internal affairs of the
family from the basic structure, political liberalism fails to address gender injustices, Rawls
writes:
Thus, when political liberalism distinguishes between political justice that applies to the
basic structure and other conceptions of justice that apply to the various associations
within that structure, it does not regard the political and the nonpolitical domains as two
separate, disconnected spaces, as it were, each governed solely by its own distinct
principles. Even if the basic structure alone is the primary subject of justice, principles of
justice still put essential restrictions on the family and all other associations. The adult
members of families and other associations are equal citizens first: that is their basic
position. No institution or association in which they are involved can violate their rights
as citizens.
124

Political liberalism is ready to interfere when citizens’ rights are endangered. Moreover, the
basic structure is designed to provide a background framework within which citizens can
exercise their moral powers. Such a framework requires that children be educated so as to
develop the two moral powers and to understand their status as free and equal citizens. A boy
raised in a sexist household may indeed come to hold sexist beliefs later in life, but it is far from
clear that he will have such strongly misogynistic or sexist views that he will be unable to
support or comply with just institutions despite his education.
3 Caring Citizens
Rawls’s theory of distributive justice has been criticized by Capabilities theorists such as
Martha Nussbaum and Amartya Sen for measuring levels of advantage according to resources
(such as Rawls’s primary social goods) while failing to accommodate the differences in people’s
ability to convert resources into well-being.
125
Rawls’s two principles of justice are designed to
regulate the terms of social cooperation between “normal and fully cooperating members of

124
Rawls, Justice as Fairness: A Restatement, 163-164.
125
Amartya Sen, "Equality of What?."
150

society”.
126
Moreover, by taking resources such as his social primary goods, rather than
capabilities, both as the measure of people’s level of advantage and as the currency of
distributive justice, Rawls’s theory fails to adequately address the needs of the disabled and those
who care for them as a core concern of justice.
Issues of justice surrounding disability often intersect with issues of dependency,
although not all who are disabled are dependent and vice versa. All citizens spend a significant
portion of their lives as dependents. Infancy and childhood are states of dependence, and the
elderly require varying degrees of assistance relative to their degree of disability. Moreover,
many citizens, at some point in their adult lives, become temporarily dependent due to illness,
pregnancy, or injury. The permanently disabled comprise yet another class of individuals that
can be further differentiated according to distinct needs and concerns; some of the permanently
disabled are also permanently dependent while others are nonetheless able to participate in the
workforce, their communities, and public life.
While capabilities theorists take issue with Rawls’s theory on the grounds that a theory of
justice should, as far as possible, secure basic capabilities for all, some feminists pose a related
criticism that focuses on dependency.
127
Though not all citizens will experience disability during
the course of their lives, everyone spends a significant portion of their lives as dependents.
Accordingly, most citizens also spend as much time caring for dependents; mothers, in
particular, often shoulder a greater burden in raising children and caring for elderly parents
In this section, I pose a case in which gender roles may indeed threaten a citizens’ freedom in
exercising her capacity for a conception of the good. Traditionally, the task of caring for
children and for elderly parents has fallen upon mothers and wives. While there may be many

126
Rawls, Political Liberalism, 20.
127
See Eva Feder Kittay, Love's Labor: Essays on Women, Equality, and Dependence (New York, NY: Routledge,
1999).
151

other such cases that pose challenges to Rawls’s theory, I will focus in particular on the concerns
of those who care for permanently and severely disabled family members.
Consider a mother who must care for a severely developmentally disabled child. The
amount of care involved in such a case is far beyond the typical demands of child rearing.
Moreover, because the child is permanently disabled, the mother will typically care for the child
over the course of her entire life or the child’s life, unless she has the resources to hire others to
care for the child. The high level of care involved means that the mother must undertake great
sacrifices. Typically, it will be very difficult for her to pursue a career. She will not have as
many opportunities for leisure or as great a range of leisurely activities to enjoy. If the child has
a normal life-span, then the mother may even continue to care for the child well into her late
years, until she is no longer able. Clearly, such a life involves many burdens and great sacrifices.
Specifically, a caregiver in this situation sacrifices the kind of life she may have had if
she did not have the responsibility of caring for the permanently dependent child. This is
significantly unlike the kind of “sacrifice” individuals make when they decide to pursue one
career plan over another, for example. The caretaker’s course of life is, in a sense, not chosen by
her. She could simply refuse to care for the child, but this is clearly not a real option for her to
take. Whatever plan of life the caregiver pursued before acquiring the responsibility of caring for
the permanently dependent child is no longer a viable option for her. It seems plausible to say,
then, that caregivers in these circumstances have less effective freedom in exercising her
capacity for a conception of the good. She lacks the material and social conditions that would
enable her to pursue a plan of life that, for example, involves a full-time career.
It is not entirely clear that Rawls’s theory can adequately address this problem. Fair
equality of opportunity requires that those with similar talents and willingness should have
152

similar chances of success. But the caregiver’s circumstances prevent her from being willing to
apply for positions in the first place. Her duties as a parent, together with prevailing gender
norms, constrain the kinds of ambitions it is feasible for her to form and pursue. The Difference
Principle selects a form of economy in which benefits to the most advantaged redound to the
greatest benefit of the least advantaged. However, it is not clear that the caregiver is guaranteed
enough material resources to both care for the permanently dependent child and pursue her own
plan of life. Because levels of advantage are measured in terms of social primary goods, it is
unclear how the Difference Principle can provide greater economic and social resources that
respond directly to her needs as a care-giver and the labor that she undertakes in providing that
care.
4 Compensating Care
In order to meet the care-giving citizen’s higher-order interest in exercising the two moral
powers, a distributive principle should ensure that all forms of socially necessary labor, including
the care that parents provide, be appropriately compensated. Feminists who argue that Rawls is
unable to adequately address the concerns of the dependent and those who care for them will
have reason to find the Burdens Compensation Principle an attractive alternative. Because
dependent care is socially necessary, those who provide care engage in labor that contributes to
the social good. Democratic societies must meet the basic human needs of all its members and
so has an interest in the care of its permanently disabled and dependent. In providing care, care-
givers are supporting a society in which all members have, at the very least, the basic provisions
they need to lead a decent human life. Reciprocity requires that care-givers, in turn, receive their
share of the benefits of social cooperation. Because the Burdens Compensation Principle calls
for compensation for all forms of socially necessary labor, those who engage in socially
153

necessary labor within the family are not excluded from social cooperation. Rather, even though
they do not offer their labor on the market, they are regarded as nonetheless contributing to the
social good insofar as they provide care for dependents.
The Burdens Compensation Principle will also allow care-givers to enjoy equal
citizenship. Recall that while it is far from clear that the Difference Principle can respond
directly to the claims of care-givers, the Burdens Compensation Principle will compensate care-
givers for the burdens they undertake in performing socially necessary labor. Compensation
need not only come in the form of remuneration. It may be in the caregivers’ best interests to
instead have more flexible work hours, the option to telecommute, employer-provided child-care
or subsidies for personal care attendants. Hence, even though the caregiver does not offer her
labor on the market, the Burdens Compensation Principle will nonetheless ensure that she does
not become economically disenfranchised. Nor need she give up her plans of life as she would if
her labor were left uncompensated.
Moreover, the Burdens Compensation Principle provides, I think, a solution to the
tension between liberal considerations in favor of maintaining inter-family dynamics in the
private realm and feminist concerns over the disenfranchisement of household laborers and
caretakers. In providing compensation for household labor and private care-work, the Burdens
Compensation Principle stays silent on how the family should be internally ordered. Thus, it
would not intrude upon citizens’ interests in internally ordering their family in accordance with
their reasonable comprehensive doctrines. Yet, it ensures that household laborers and private
care-workers are included in scheme of social cooperation and compensated for the socially
necessary labor that they perform. In doing this, the principle supports the citizenship of the
154

caregivers by providing the social and material conditions they need to exercise their capacities
for a sense of justice and a conception of the good.

155

Bibliography

Anderson, Elizabeth. "What Is the Point of Equality?" Ethics 109, no. 2 (1999): 287-337.

Arneson, Richard. "Equality and Equal Opportunity for Welfare." Philosophical Studies 56, no.
1 (1989): 77-93.

Arneson, Richard. "Egalitarianism and Responsibility." The Journal of Ethics 3, no. 3 (1999):
225-247.

Arneson, Richard. "Luck Egalitarianism and Prioritarianism." Ethics 110, no. 2 (2000): 339-349.

Arneson, Richard. "Luck and Equality." Aristotelian Society Supplementary Volume 75, no. 1
(2001): 73-90.

Arneson, Richard. "Luck Egalitarianism: An Interpretation and Defense." Philosophical Topics
32, no. 1 (2006): 1-20.

Barry, Nicholas. "Defending Luck Egalitarianism." Journal of Applied Philosophy 23, no. 1
(2006): 89-107.

Benbaji, Yitzhak. "Sufficiency or Priority?" European Journal of Philosophy 14, no. 3 (2006):
327-348.

Brighouse, Harry and Adam Swift. "Educational Equality Versus Educational Adequacy: A
Critique of Anderson and Satz." Journal of Applied Philosophy 26, no. 2 (2009): 117-128.

Casal, Paula. "Why Sufficiency Is Not Enough." Ethics 117, no. 2 (2007).

Christiano, Thomas and Will Braynen. "Inequality, Injustice and Levelling Down." Ratio 21, no.
4 (2008): 392-420.

Cohen, Gerald A. "On the Currency of Egalitarian Justice." Ethics 99, no. 4 (1989): 906-944.

Cohen, Gerald A. "Equality of What? On Welfare, Goods and Capabilities." Recherches
Économiques de Louvain/Louvain Economic Review, (1990): 357-382.

Cohen, Gerald A. "Incentives, Inequality, and Community." The Tanner lectures on human
values 13, (1992): 263-329.

Cohen, Gerald A. "Where the Action Is: On the Site of Distributive Justice." Philosophy &
Public Affairs 26, no. 1 (1997): 3-30.

Cohen, Gerald A. "Justice, Incentives, and Selfishness." In If You're an Egalitarian, How Come
You're So Rich?, 117-133. Cambridge: Harvard University Press, 2000.
156

Cohen, Gerald A. "Facts and Principles." Philosophy & Public Affairs 31, no. 3 (2003): 211-245.

Cohen, Gerald A. Rescuing Justice and Equality Harvard University Press, 2009.

Crisp, Roger. "Egalitarianism and Compassion." Ethics 114, no. 1 (2003): 119-126.

Crisp, Roger. "Equality, Priority, and Compassion." Ethics 113, no. 4 (2003): 745-763.

Dorsey, Dale. "Equality-Tempered Prioritarianism." Politics, Philosophy & Economics 13, no. 1
(2014): 45-61.

Dworkin, Ronald. "What Is Equality? Part 1: Equality of Welfare." Philosophy & Public Affairs
10, no. 3 (1981): 185-246.

Dworkin, Ronald. "What Is Equality? Part 2: Equality of Resources." Philosophy & Public
Affairs 10, no. 4 (1981): 283-345.

Dworkin, Ronald. "Equality, Luck and Hierarchy." Philosophy & Public Affairs 31, no. 2 (2003):
190-198.

Frankfurt, Harry. "Equality as a Moral Ideal." Ethics 98, no. 1 (1987): 21-43.

Frankfurt, Harry. "Equality and Respect." Social Research 64, no. 1 (1997): 3-15.

Holtug, Nils. "Prioritarianism." In Egalitarianism: New Essays on the Nature and Value of
Equality edited by Nils Holtug and Kaspar Lippert-Rasmussen, 125-156. New York: Oxford
University Press, 2006.

Hume, David. A Treatise of Human Nature Oxford Philosophical Texts, Edited by John
Cottingham. New York: Oxford University Press, 2000.

Huseby, Robert. "Sufficiency: Restated and Defended." Journal of Political Philosophy 18, no. 2
(2010): 178-197.

Jaggar, Alison M. Feminist Politics and Human Nature (Philosophy and Society): Rowman &
Littlefield Publishers, 1988.

Julius, A. J. "Basic Structure and the Value of Equality." Philosophy & Public Affairs 31, no. 4
(2003): 321-355.

Kaufman, Alexander. "Choice, Responsibility and Equality." Political Studies 52, no. 4 (2004):
819-836.

Kittay, Eva Feder. Love's Labor: Essays on Women, Equality, and Dependence. New York, NY:
Routledge, 1999.
157

Knight, Carl. "In Defence of Luck Egalitarianism." Res Publica 11, no. 1 (2005): 1-10.

Knight, Carl. Luck Egalitarianism: Equality, Responsibility, and Justice. Edinburgh: Edinburgh
University Press, 2009.

Munoz‐Dardé, Véronique. "Rawls, Justice in the Family and Justice of the Family." The
Philosophical Quarterly 48, no. 192 (1998): 335-352.

Nagel, Thomas. Equality and Partiality. New York: Oxford University Press, 1991.

Nozick, Robert. Anarchy, State, and Utopia. New York: Basic Books, 1974.

Parfit, Derek. "Equality and Priority." Ratio 10, no. 3 (1997): 202-221.

Pateman, Carole. The Sexual Contract: Stanford University Press, 1988.

Pogge, Thomas W. "On the Site of Distributive Justice: Reflections on Cohen and Murphy."
Philosophy & Public Affairs 29, no. 2 (2000): 137-169.

Porter, Thomas. "In Defence of the Priority View." Utilitas 24, no. 3 (2012).

Rakowski, Eric. Equal Justice. Oxford: Oxford University Press, 1991.

Rawls, John. A Theory of Justice. Cambridge: Belknap Press of Harvard University Press, 1971.

Rawls, John. "Kantian Constructivism in Moral Theory." The Journal of Philosophy 77, no. 9
(1980): 515-572.

Rawls, John. Justice as Fairness: A Restatement. Cambridge: Harvard University Press, 2001.

Rawls, John. Political Liberalism. expanded ed. New York: Columbia University Press, 2005.

Roemer, John. Theories of Distributive Justice. Cambridge: Harvard University Press, 1996.

Scheffler, Samuel and Véronique Munoz-Dardé. "The Division of Moral Labour." Proceedings
of the Aristotelian Society, Supplementary Volumes, (2005): 229-284.

Sen, Amartya. "Equality of What?".

Singer, Peter. "Famine, Affluence, and Morality." Philosophy & Public Affairs 1, no. 3 (1972):
229-243.

Tan, Kok-Chor. "Justice and Personal Pursuits." The Journal of Philosophy 101, no. 7 (2004):
331-362.

158

Tan, Kok-Chor. "A Defense of Luck Egalitarianism." The Journal of Philosophy 105, no. 11
(2009): 665-690.

Temkin, Larry. Inequality. New York: Oxford University Press, 1993.

Temkin, Larry. "Egalitarianism Defended." Ethics 113, no. 4 (2003): 764-782.

Temkin, Larry. "Exploring the Roots of Egalitarian Concern." Theoria 69, no. 1-2 (2003): 125-
151.

Temkin, Larry S. "Equality, Priority or What?" Economics and Philosophy 19, no. 1 (2003): 61-
87.

Williams, Andrew. "Incentives, Inequality, and Publicity." Philosophy & Public Affairs 27, no. 3
(1998): 225-247.

Young, Iris Marion. "Taking the Basic Structure Seriously." Perspectives on Politics 4, no. 1
(2006): 91-97.

Abstract (if available)

Abstract Modern liberal democracies uphold the fundamental freedom and equality of its citizens by guaranteeing them equal rights, liberties, and opportunities regardless of factors such as race, gender, religion, and disability. Citizens also need some amount of income or wealth to cooperate as productive members and to effectively pursue their own plans of life. But does justice require that income and wealth, like rights and liberties, be distributed equally? Or is an unequal distribution of income and wealth consistent with justice, and if so, what are the fair terms under which such a distribution can be reasonably acceptable to all? ❧ In this dissertation, I argue for what I call the ""Burdens Compensation Principle,"" which requires that individuals who undertake socially necessary labor be compensated in proportion to the burdensomeness of that labor. Compensation may take the form of greater income or wealth but may also include provisions, such as child care. Some forms of labor are especially tedious, onerous, dangerous, or physically, intellectually, or emotionally demanding, to name just a few examples of ways an occupation can be burdensome. As different occupations involve different burdens, an unequal distribution of income and wealth may be morally permissible on the grounds that the inequality reflects the differences in the burdens of socially necessary labor. It would be unreasonable to insist that every worker be compensated the same, regardless of the varying degrees to which their labor is burdensome. On the other hand, a citizen may reasonably object to terms under which another gains greater compensation for work that is less burdensome or socially unnecessary

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Intrinsic and extrinsic values of distributional equality

PDF

Defending Rawls's Law of Peoples against cosmopolitan critiques

PDF

Essays on macroeconomics and income distribution

PDF

The EITC, labor supply, and child development

PDF

Responding to harm

PDF

Effects of eliminating the unfunded social security system in an economy with entrepreneurs

PDF

Responsibility and the emotional structure of relationships

PDF

On being bound: law, authority, and the politics of obligation

PDF

Contract, from promise to commodity

PDF

Responsibility and rights: a search for a principled distinction between criminal law and tort law

PDF

Essays on family and labor economics

PDF

Black, Muslim, and gay/queer male allies: an intersectional analysis of men’s gender justice activism

PDF

Negotiating pluralism and tribalism in liberal democratic societies

PDF

The mental states first theory of promising

PDF

Food is the medium: food movements, social justice and the communication ecology approach

PDF

Improvements in women's status, decision-making and child nutrition: evidence from Bangladesh and Indonesia

PDF

Climate work is social work: addressing environmental justice to facilitate achievement of the Grand Challenges for social work

PDF

An aesthetic of comprehension: the distribution of American land art and conceptual art in Germany, 1968-1975

PDF

Coding commons: fun and the Ubuntu labor process

PDF

Reasoning with uncertainty and epistemic modals

Asset Metadata

Creator Chang, Tiffany (author)

Core Title Distributive equality and labor burdens

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Philosophy

Publication Date 07/15/2014

Defense Date 05/28/2014

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag distributive justice,equality,Labor,OAI-PMH Harvest

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Lloyd, Sharon (committee chair), Keating, Gregory (committee member), Watson, Gary L. (committee member)

Creator Email chang.tiffany@gmail.com,changti@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-441744

Unique identifier UC11286677

Identifier etd-ChangTiffa-2696.pdf (filename),usctheses-c3-441744 (legacy record id)

Legacy Identifier etd-ChangTiffa-2696.pdf

Dmrecord 441744

Document Type Dissertation

Format application/pdf (imt)

Rights Chang, Tiffany

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA