See discussions, stats, and author profiles for this publication at: [Link]

net/publication/288967482

A critique of the capability approach

Article · January 2010

DOI: 10.1017/CBO9780511810916.002

CITATIONS READS

49 12,288

1 author:

Thomas Pogge
Yale University
270 PUBLICATIONS 8,520 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Impact Funds - [Link] View project

Amartya Sen Essay Prize View project

All content following this page was uploaded by Thomas Pogge on 09 May 2017.

The user has requested enhancement of the downloaded file.

 A Critique of the Capability Approach*
During the past 25 years, the capability approach, developed by Amartya Sen and Martha
Nussbaum, has come to play a major role in political philosophy and normative economics.
This approach has gained much support, among academics as well as among international
agencies and nongovernmental organizations, at the expense of competing resourcist and
welfarist approaches exemplified, respectively, by John Rawls’s theory and utilitarianism.

Here I examine how the capability approach has been, and might be, justified as superior to,
in particular, its resourcist competition. I reach two main conclusions. First, this question
should not be answered in isolation, but can be plausibly resolved only in conjunction with
other key elements of a conception of social justice. Instead of asking which approach is
superior, we should ask which approach can deliver the most plausible public criterion of
social justice. Second, neither Sen nor Nussbaum has so far shown that the capability
approach can produce a public criterion of social justice that would be a viable competitor to
the more prominent resourcist views.

While I concentrate my critical attention on the capability approach and reject much of the
case made in its favor, my intent is wholly constructive. I am working on the same problem
as the pioneers of the capability approach and I do not have all the answers any more than
they do. If they can learn from my critique even a fraction of what I have learned from their
work, I will be well satisfied.

Section 1 shows that capability theorists have overstated the systematic difference between
the two approaches, which boils down to this: Capability theorists assert, while resourcists
deny, that a public criterion of social justice should take account of the individual rates at
which persons with diverse physical and mental constitutions can convert resources into
valuable functionings. Section 2 examines to what extent a resourcist criterion can be
sensitive to natural human diversity. It also explores the reasons for and against believing
that greater compensatory accommodation of such diversity, as facilitated by the capability
approach, would lead to a more plausible public criterion of justice.

1. Some Supposed Contrasts Between the Capability and Resourcist Approaches

*
This essay is an abridged version of an much longer paper “Can the Capability Approach be Justified?” in
Martha Nussbaum and Chad Flanders eds.: Global Inequalities, special issue 30:2 (Fall 2002) of

pogge 1
Both Sen and Nussbaum hold that, for purposes of assessing alternative feasible institutional
schemes on the basis of how each treats its individual participants, “the appropriate ‘space’ is
neither that of utilities (as claimed by welfarists), nor that of primary goods (as demanded by
Rawls), but that of the substantive freedoms — the capabilities — to choose a life one has
reason to value.”1

As G. A. Cohen has pointed out, this emphasis on freedoms as the hallmark of the capability
approach can be misleading.2 In explicating the meaning of “capabilities,” Sen emphasizes
that he is concerned not with what persons have or are, with their achievements or
functionings, but rather with what they can have or be. Capabilities are options to achieve
valuable functionings.3 This emphasis, however, is one that resourcists can fully share. They
tend to focus not on the goods persons actually have or consume, but on the goods persons
can have or consume. Rawls, for instance, evaluates social positions in terms of the access
they provide to or through certain all-purpose means such as basic liberties, opportunities,
and money. The key question dividing the relevant approaches is then not: Should alternative
feasible institutional schemes be assessed in terms of what their participants have or in terms
of what their participants have access to?

Rather, the key question is: Should alternative feasible institutional schemes be assessed in
terms of their participants’ access to valuable resources or in terms of their participants’
capabilities, that is, access to valuable functionings? In the context of this dispute, the
distinctive feature of the capability approach is its focus on what Cohen calls “midfare”4 or,
in Sen’s words, “on the state of the person, distinguishing it both from the commodities that
help generate that state, and from the utilities generated by the state.” Quoting Cohen, Sen
continues: “We must look, for example, at her nutrition level, and not just, as Rawlsians do,
at her food supply, or, as welfarists do, at the utility she gets out of eating food.”5 In light of
this clarification, the dispute would be better described as being about resources versus

Philosophical Topics, 167-228. I have tried my best to make it self-contained.

1
Sen: Development as Freedom (New York: Anchor Books 2000), 74.
2
G. A. Cohen: “Equality of What? On Welfare, Goods, and Capabilities” in Martha Nussbaum and Amartya
Sen: The Quality of Life (Oxford: Clarendon Press 1993), 9-29.
3
Sen often gives the example of affluent persons fasting by choice. Though they may be undernourished, they
are not to be grouped with other hungry people who cannot afford enough to eat. What matters is their
capability to be well-nourished, in terms of which they are no worse off than other affluent people who are
eating three square meals a day.
4
G. A. Cohen: “Equality of What?,” esp. 18.
5
Amartya Sen: “Capability and Well-Being” in Nussbaum and Sen: The Quality of Life, 43.

pogge 2
functionings rather than about resources versus capabilities — but I will stick to the received
labels.

The resourcist and capability approaches can each be specified in a wide variety of ways. I
leave aside this internal diversity as much as possible to focus instead on the central
disagreement that divides the two approaches. If one side were right about this central
disagreement, then we could infer that we should look on that side for the most plausible
criterion for assessing alternative feasible institutional schemes. We could not infer,
however, that every view on the winning side is more plausible than every view on the losing
side. This second conclusion is mistaken, because a criterion of social justice on the winning
side of the resources-vs.-capabilities dispute may have many other serious flaws that, all
things considered, render it inferior to some criteria on the losing side of this dispute.

It follows that the contrapositive of the second conclusion is equally mistaken: The fact that
one particular capability view is more plausible than one particular resourcist view does not
show that the capability approach is superior to the resourcist approach. Unfortunately, both
Sen and Nussbaum occasionally suggest this invalid argument: They compare an implausible
resourcist view — typically one that assesses alternative feasible institutional schemes by the
average income (or GNP per capita) each would engender — to a more plausible capability
view and then suggest that the capability approach is therefore more plausible than the
resourcist approach.6

Rawls makes the same sort of mistake in the opposite direction. He argues that the reasons
we have for preferring a political over a comprehensive conception of social justice also
show that the way an institutional order treats its participants must be evaluated in terms of
resources rather than capabilities. This is so, according to Rawls, because a capability view
takes a stand on the relative value of the many diverse ends human beings might pursue
whereas a resourcist view like his own can focus more abstractly on certain all-purpose
means that are essential to most any way of life.7 As Sen makes clear, this is a bad
argument.8 A resourcist view might well take a stand, implicitly, on basic moral values and
on the relative merits of various human pursuits; it would be doing so if it focused not on
such abstract and general goods as Rawls emphasizes (e.g., political liberties, income and

6
Cf., for example, Sen: Development as Freedom, 87, and Martha Nussbaum: Women and Human
Development (Cambridge: Cambridge University Press 2000), 60f.
7
Rawls: Collected Papers (Cambridge MA: Harvard University Press 1999), 456.
8
Sen: Inequality Reexamined (Cambridge, MA: Harvard University Press 1995 [1992]), 82-4.

pogge 3
wealth), but on very concrete and partisan goods such as being ruled by pious Catholics or
having access to Verdi performances. More to the point, a capability view can be formulated
quite generally so that it focuses on capabilities that (like Rawls’s social primary goods) are
important to all human pursuits or nearly all. Sen has in fact emphasized capabilities of this
sort, such as being well-nourished and having physical mobility. By ascribing to persons an
interest in such capabilities, Sen is not committing himself to any particular comprehensive
conception of the good life.

To advance the debate between the two approaches, we need a sharper analysis of how they
differ. Sen has listed certain key determinants of quality of life that, he claims, are ignored by
the simpler resourcist criteria of social justice, which focus on income.9 He presents plausible
capability criteria as having the benefit of being sensitive to these determinants. However,
this exercise establishes a reason for preferring the capability approach over the resourcist
approach as such only if even sophisticated resourcist criteria cannot take these determinants
into account. Any time we find such a determinant, which the capability approach can and
the resourcist approach cannot be sensitive to, then we have made progress toward
understanding the contrast between the two approaches. The next question is then whether
this determinant ought to be taken into account. Only when the answer to this question is
affirmative have we found a reason for favoring the capability approach over the resourcist
approach.

Here are the determinants on Sen’s list (with the order changed, but his headings preserved):

(1) Distribution within the family: “intrafamily distribution of incomes is quite a crucial
parametric variable in linking individual achievements and opportunities with the overall
level of family income.”10 The resourcist and capability approaches both share the
commitment to normative individualism. Both agree, therefore, that the comparative
assessment of alternative feasible institutional schemes must focus on individual shares (of
relevant resources or capabilities). Whether a criterion of social justice takes account of
intrafamily distribution depends not on the metric it employs (capabilities, resources,
welfare, or whatever), but on its interpersonal aggregation function: Equalitarian,

9
“It is easy to identify at least five distinct sources of variation between our real incomes and the advantages
— the well-being and freedom — we get out of them” (Sen: Development as Freedom, 70). As this sentence
shows, the following list is meant to provide reasons for preferring the capability approach over a resourcist
approach, in particular.
10
Ibid., 71, cf. also 88f. and Amartya Sen: On Economic Inequality (Oxford: Oxford University Press 1997
[1973]), 212.

pogge 4
prioritarian, and sufficientarian criteria of social justice all do, while sumranking and
averaging criteria do not, take account of intrafamily distribution. Contrary to what Sen
suggests, capability and resourcist criteria of social justice do not differ on the issue of
intrafamily distribution.11

Now it may be true, as Sen suggests in a parallel passage, that systematic, often anti-female
bias in intrafamily distribution of resources “is more readily checked by looking at capability
deprivation (in terms of greater mortality, morbidity, undernourishment, medical neglect, and
so on) than can be found on the basis of income analysis.”12 But this is a point about the best
evidence we can have for unequal and unjust intrafamily distribution. It does not settle the
question what such injustice consists in: The capability theorist will describe the injustice as
family resources being distributed so that males systematically have greater capabilities. The
resourcist will describe the injustice as men and boys systematically receiving larger shares
of family resources than women and girls do. In any case, it is not true, pace Sen, that the
resourcist approach leads its adherents to overlook the very serious problem of intrafamily
distribution.

(2) Differences in relational perspectives: “The commodity requirements of established

patterns of behavior may vary between communities, depending on conventions and
customs. For example, being relatively poor in a rich community can prevent a person from
achieving some elementary “functionings” (such as taking part in the life of the community)
even though her income, in absolute terms, may be much higher than the level of income at
which members of poorer communities can function with great ease and success.”13 The
point Sen makes here — that individual advantage has relative as well as absolute aspects —
is at right angles to the debate between the two approaches: A capability theorist can run
afoul of Sen’s point by disregarding all social and relative functionings and focusing
exclusively on such functionings as being well-nourished and having physical mobility.
Conversely, even a simple income resourcist can accept Sen’s point: by recognizing that the
value of any level of income depends in part on what incomes other participants enjoy and

11
See Susan Okin: Justice, Gender, and the Family (New York: Basic Books 1989) for extensive treatment of
intrafamily distribution within a resourcist conception of social justice. As I explain the longer version of my
essay, I use the word "egalitarian" for criteria of social justice that, in assessing institutional schemes, give
equal consideration to their individual participants. Nearly all criteria of social justice defended today are
egalitarian in this sense. I use the word "equalitarian" for criteria of social justice that assess institutional
schemes by the extent to which they generate substantive equality among their participants.
12
Sen: Development as Freedom, 89, and Sen: On Economic Inequality, 212.
13
Sen: Development as Freedom, 71. Cf. Sen: On Economic Inequality, 212f.

pogge 5
that, partly for this reason, an institutional order may be unjust because the incomes it makes
available to some are too low relative to the incomes it makes available to others. A plausible
resource metric must include all and only the resources human beings need to function
adequately and must weight these resources according to their importance in fulfilling
standard human needs.14 If a reasonably competitive income is important for persons to
function adequately — and there is considerable evidence that this is so15 — then the relative
size of incomes should be incorporated into an appropriate resource metric.16

More sophisticated resourcists will accept Sen’s point with regard to various other goods
they recognize besides income: The value and the moral adequacy of the rights a person
enjoys and of the education, medical care, and employment available to her all depend in part
on how they compare to the rights and educational, medical, and employment opportunities
enjoyed by those around her. In Rawls’s theory, these relative aspects are expressed in the
demand for equal basic liberties and equal opportunities as well as in such intrinsically
relational goods as “powers and prerogatives of offices and positions of responsibility” and
especially “the social bases of self-respect.”17

(3) Variations in social climate: “The conversion of personal incomes and resources into the
quality of life is influenced also by social conditions, including public educational
arrangements, and the prevalence or absence of crime and violence ... epidemiology and
pollution ... the nature of community relationships.”18 This thought can properly be advanced
against any resourcist view that focuses on personal incomes and resources to the exclusion
of social capital and, more generally, public goods and ills. Such a view would presumably
oppose most tax-funded public expenditures: for a criminal justice system, for public health,

14
Does such a reference to standard human needs constitute a capitulation to the capability approach? If it did,
a truly resourcist view would have to select and weight the resources within its metric without regard to our
knowledge about what human beings generally need to function adequately. Such “true” resourcism would
evidently be absurd. The interesting resourcism, here at issue, is distinctive by holding that the justice of
institutional schemes should be judged on the basis of what distribution of valuable resources (rather than
valuable capabilities or utilities) would exist under it. This will become much clearer in the sequel.
15
Consider the economics literature on positional goods as well as more recent evidence that relative poverty
is an important explanatory factor in human health. Cf. Sir Douglas Black, J. N. Morris, Cyril Smith, Peter
Townsend: “The Black Report,” in Peter Townsend and Nick Davidson, eds.: Inequalities in Health (London:
Penguin 1990) and Richard Wilkinson: Unhealthy Societies: The Afflictions of Inequality (London and New
York: Routledge 1996).
16
Within Rawls’s theory, for instance, there is strong reason to include persons’ relative income — expressed
perhaps as the ratio of income over median income — among the social bases of self-respect. Cf. Pogge:
Realizing Rawls (Ithaca: Cornell University Press 1989), 162f.
17
Rawls: Collected Papers, 362f., 454.
18
Sen: Development as Freedom, 70f.

pogge 6
for sports and the arts, and for maintaining national parks and public libraries. To my
knowledge, this is not a position any actual resourcists have taken. They do not focus on
personal (rivalrous or excludable) resources alone, but rather count the prevalence of crime
and violence and the lack of public safeguards against biological and chemical hazards as
diminishing a person’s resources broadly conceived. Our paradigm resourcist Rawls sees
such factors as rendering insecure some of the basic liberties of citizens such as their physical
and psychological integrity and their freedom of movement.19 He also emphasizes the
political liberties and their fair value as a crucial primary good that allows citizens, together,
to shape the social and natural conditions that shape their lives. In these ways, a sophisticated
resourcist view does take account of social conditions and of their potentially differential
impact on persons and groups — albeit in a different way than the capability approach which
is sensitive to these factors only in proportion to the influence they exert on individuals’
capabilities or quality of life.

(4) Environmental diversities: “Variations in environmental conditions ... can influence what
a person gets out of a given level of income. Heating and clothing requirements of the poor
in colder climates cause problems that may not be shared by equally poor people in warmer
lands. The presence of infectious diseases in a region ... alters the quality of life that
inhabitants of that region may enjoy.”20 Again, these are points that resourcists have every
reason to take into account. Thus, if we count a place where one can live and work among
the personal goods or resources, then it makes sense to count such a place as more valuable
when it offers a temperate climate and healthy environment, just as one would count such a
place as more valuable when it offers an abundance of freely available nutritious fruits and
vegetables.

It must be said, however, that actual resourcists have paid insufficient attention to variations
in environmental conditions. They are not discussed by Rawls, for instance, and play no role
in his criterion of social justice. This is so, I conjecture, because he assumes that the citizens
of a just society enjoy freedom of movement. On this assumption, he finds it implausible to
count Susan as worse off than Bill merely because she lives in a less hospitable environment
— she lives there by choice, after all, and could live where Bill lives if she wanted to. (If she

19
Cf. Rawls: A Theory of Justice (Cambridge MA: Harvard University Press 1999 [1971]), 211-13, where he
takes the prevalence of crime and violence to constitute a loss in basic liberties which, as the preeminent
primary goods, are governed by his first principle of justice.
20
Sen: Development as Freedom, 70.

pogge 7
is poorer than Bill and cannot afford to live where he lives for this reason, then Rawls’s
criterion will appropriately count her as worse off than Bill, but not on account of her
environmental conditions but on account of her lesser income and wealth.)

While my conjecture goes some way toward defending Rawls’s criterion, it does not in the
end provide a valid defense. The difference principle compares the socioeconomic positions
of citizens in terms of (simply put) the work-pay packages available to them, without regard
to the geographical area in which such work-pay packages may be available. This can lead to
an intuitively implausible ranking of socioeconomic positions for purposes of identifying the
least advantaged under each feasible institutional order and of comparing such institutional
alternatives by reference to the least advantaged position each generates. Persons whose best
option is to sell fast food in San Diego for a wage of $1500 per month are intuitively better
off than persons whose best option is to perform the same work in Alaska for a wage of
$1510 per month. The reason is that the latter, to do the job, must live in Alaska and must
then also pay more than the former do for clothing and for heating their homes (additional
expenses that exceed $10 per month).

I agree then that a plausible criterion of social justice must take account of environmental
diversities and that actual resourcists have failed to do so. Still, the needed correction is in
the spirit of the resourcist approach: In measuring resources persons have access to, one must
subtract resources standardly needed to enjoy such access — the heavy coat one needs to
withstand the Alaskan winter as much as the uniform one may need in order to hold down a
job as a mail carrier or waitress.

(5) Personal heterogeneities: “People have disparate physical characteristics connected with
disability, illness, age or gender, and these make their needs diverse.”21 How such personal
heterogeneities should be treated is really at the core of the debate between the two
approaches. Resourcists define and consider individual shares without regard to the particular
features of the persons whose shares they are. In selecting the various goods in terms of
which they define their resourcist metric, and in weighting these selected goods relative to
one another, resourcists are guided by some conception of the standard needs and
endowments of human beings. Capability theorists, by contrast, value the goods persons have
access to by reference to the specific needs and endowments of each particular person. In this
sense, capability theorists are, while resourcists are not, sensitive to personal heterogeneities.

pogge 8
Sen tends to overstate the contrast, however, in that some of the features he lists as personal
heterogeneities are ones that resourcists can and do take into account.

(5.1) Sen gives the example of women who, because they are breast-feeding an infant child,
need more nutrition than other persons.22 But surely resourcists are not committed to
counting the needs of infants for naught. To the contrary, they are committed to recognizing
that a share of food or income just adequate for a single adult is less or not adequate if shared
by a woman and her infant child together.

Closely related is Sen’s example of extra food needs during pregnancy.23 Resourcists can
hardly fail to recognize that all human beings have needs even before they are born and must
have these needs met in order to reach their full human potential. Only their mothers can
meet these pre-birth needs and, if they decide to do so, these mothers have additional needs
related to their pregnancy. They must be able to meet these needs in order to meet the pre-
birth needs of their future children. Understanding this, resourcists can easily agree with Sen
that a share of food or income just adequate for a non-pregnant woman is less or not
adequate for a pregnant woman and her unborn child. Any plausible resourcist view will take
account of pregnancy and lactation.

Making this point does not in any way detract from the very great contributions Sen and
Nussbaum have made toward spreading awareness of the economic injustices inflicted
specifically upon women and girls. But it does show that these great contributions do not
presuppose the capability approach. The commitment to feminism is separable from the
commitment to the capability approach: The former does not presuppose the latter and
therefore cannot support it either.24

In fact, the capability approach may even weaken the feminist case by suggesting — falsely
— that women’s terrible and disproportionate suffering in most of this world is due to their
being insufficiently compensated for their inferior natural endowments. Women’s suffering
in the world as it is does not result from social institutions being insufficiently sensitive to

21
Ibid., 70.
22
Sen: Inequality Reexamined, 113.
23
Sen: Inequality Reexamined, 33 and 113.
24
The validity of this point is confined to the debate between capability theorists and resourcists. In the debate
with welfarism, capability theorists can employ feminist arguments. They can point out, for example, that
welfarism does not count as disadvantaged women who have — through social conditioning or by choice
(adaptive preference formation) — accepted their inferior status in their society. This important point against
welfarism can, however, be made just as effectively by resourcists as well.

pogge 9
the special needs arising from their different natural constitution. Rather, it overwhelmingly
results from institutional schemes and cultural practices being far too sensitive to their
biological difference by making sex the basis for all kinds of social (legal and cultural)
exclusions and disadvantages. Women and girls have a powerful justice claim to the removal
of these barriers, to equal treatment (in a resourcist sense). If these barriers were removed, if
our social institutions assured women of equal and equally effective civil and political rights,
of equal opportunities, of equal pay for equal work, women could thrive fully even without
any special breaks and considerations. So why detract from this irresistible demand for equal
treatment in a resourcist sense by staking claims to compensation for greater needs or special
disabilities? Why open the door to all sorts of counter-claims invoking the special needs and
disabilities of males, based perhaps on their naturally lower life expectancy or on their larger
bodies in need of additional food and clothing? Why, as it were, muddy the waters?

Institutional schemes are often facially sex neutral — not openly discriminatory against
women and girls, and yet male centered. For example, access to certain positions can be
subject to restrictions that are not substantially related to the demands of the job and, while
not excluding women as such, nonetheless make it much harder or rarer for women to
qualify. The restrictions do this by targeting features of persons that, for biological or cultural
reasons, are correlated with sex. Such restrictions may, for instance, avoidably exclude
candidates who are pregnant or have small children to take care of or fall below a certain
height or weight or wear “obtrusive” headgear (chadors but not yamulkas).

A criterion of social justice, too, can be facially sex neutral and yet male centered, and may
then be blind to the unjust biases built into some of the institutional schemes it is used to
assess. Such a criterion may, for instance, specify the standard needs and endowments of
human beings in a way that is more appropriate for males than for females. Such a male-
centered criterion of social justice may blind its adherents to the ways in which their
institutional order is covertly discriminatory against women.

This danger is real and important. But it is, once again, orthogonal to the contrast at issue. It
is possible to formulate a resourcist criterion of social justice that, in selecting, formulating,
and weighting valuable resources, takes full account of any divergences in the needs of males
and females and also is appropriately sensitive to covert forms of discrimination by assessing
any institutional order by how it actually works in its social and natural context. Such a
resourcist criterion would employ an unbiased conception of the standard needs and

pogge 10
endowments of human beings and would count an institutional order as just only if it secured
genuinely equal treatment to its female and male participants. On the other hand, it is also
easily possible to formulate a male-centered capability view that bases its selection,
formulation, and weighting of valuable capabilities on the standard needs of males and also is
insensitive to covert forms of discrimination by analyzing any institutional order without
attention to the possibly quite sexist culture within which it is operating.

It is certainly difficult, within either approach, to develop a criterion of social justice that
fully meets this vital challenge — with respect to women as well as to other historically
disadvantaged groups. On this score, the capability theorists are way ahead of most
resourcists, thanks to the great efforts by Sen, Nussbaum, and others. But I see no reason to
doubt that resourcists can do as well. They should certainly make the effort.

(5.2) In response to Sen’s examples of lactation and pregnancy, I have defended resourcism
by invoking the needs human beings have during their infancy and gestation in the womb.
This defense heightens the interest of the next example he adduces to show that a capability
criterion can treat personal heterogeneities in a more plausible way than any resourcist
criterion can. Here Sen argues that when a criterion of social justice makes interpersonal
comparisons in terms of a resource metric, it ignores the fact that human needs vary with
age.25 Yet, in regard to this example as well, resourcism can do rather better than Sen
supposes. Normative individualism, common ground between the resourcist and capability
approaches, holds that any institutional order should be assessed by how it treats its
individual human participants. Sen’s example has force, if we think of such participants as
time slices. But it then poses an equally powerful objection to the capability approach as
well: Whether you think of human beings as needing certain resources or certain
functionings — either way, our needs vary with age.26

It is more common, and plainly more plausible within both approaches, to conceive of
participants as persons over a complete life. They are certainly so conceived by Rawls,27 who
does not regard an institutional order under which all persons enjoy the same upward

25
Sen: On Economic Inequality, 212; Sen: Inequality Reexamined, 113; Sen: Development as Freedom, 70
and 88.
26
Young children do not have much use, for instance, for some of the capabilities Nussbaum lists, for “being
able to participate effectively in political choices that govern one's life” or “being able to search for the
ultimate meaning of life in one's own way” (Nussbaum: Women and Human Development, 78-80, capabilities
4 and 10A).

pogge 11
mobility, with uniformly rising rank and income, as inequalitarian merely because, at any
given time, older persons enjoy higher socioeconomic status than younger ones. His
difference principle does not judge such an order, under which the incomes of the least
advantaged rise with age from $8 to $20 per hour, say, inferior to a feasible alternative under
which the least advantaged earn $9 per hour throughout their lives.28

If participants are conceived as persons over a complete life, age-specific variations in needs
are not a personal heterogeneity. Persons have diverse needs at different times of their lives,
and the adequacy of individual shares depends then not merely on what persons have access
to, but also on when they have such access. Just as having access to ten pounds of food once
a week is less adequate than having access to one pound of food each day, so having access
to dentures as an infant and to mother’s milk in old age is less adequate than the other way
round. Recognizing these points, resourcists can specify the standard needs of human beings
in the various phases of their lives and can then define the adequacy of a person’s lifetime
share as an intertemporal aggregate of how adequate the resources she has access to in each
phase of her life are relative to the age-specific standard human resource needs during this
phase. To be sure, resourcists have not tackled this task,29 but capability theorists have not
tackled it either. Both have clear reason to conceive the participants in institutional schemes
as persons over a complete life (rather than as time-slices of such persons) and to recognize
that the needs of such persons vary with age. Doing so, both approaches can easily avoid the
conclusion Sen rightly finds embarrassing: They can avoid saying that an institutional order
containing special provisions to meet age-specific needs is, for this reason alone, treating its
participants unequally. A society is not inequalitarian merely because it has a public school
system that is open to teenagers but not to septogenarians, if the latter had similar
opportunities when they were young. Nor is it inequalitarian merely because it offers a
special tax break to the aged, if those currently ineligible can receive an equivalent tax break
in their old age.

27
Cf. e.g. Rawls: A Theory of Justice, 56, coining the expression “life prospects”, and Rawls: Collected
Papers, 364.
28
For discussion of such a case, cf. Rawls: Political Liberalism (New York: Columbia University Press 1993),
270f.
29
Rawls has nothing to say about intertemporal aggregation within one life, although this topic is essential for
making his two principles of justice operational in the real world. For a tentative treatment of this topic within
a Rawlsian framework see Thomas Pogge: Kant, Rawls, and Global Justice (Ann Arbor: University
Microfilms International 1983), chapter 3. I also try there to accommodate within Rawls’s resourcist
framework what undeniably is a personal heterogeneity: the fact that persons die at different ages.

pogge 12
Surely, a great deal more could be said about justice between generations, and even what
little I have said about how best to respond to the fact of age-specific needs is not
uncontroversial.30 Still, no more need be said to establish this minimal point: Criteria of
social justice that employ a resource metric are not, for this reason, insensitive to age-specific
needs.

(5.3) Even if participants are conceived as persons over a complete life and age-specific
differences are thus left aside, participants do of course still differ considerably in both body
and mind and therefore also in their capacity to convert valuable resources into valuable
functionings. Even these differences, however, are not ones that resourcist views are
constrained to ignore wholesale. Resourcists can recognize that a person’s mental and
physical constitution is, to a considerable extent, shaped by social factors: by the locality and
family in which one is raised (which greatly influence one’s access to nutrition, medical care,
physical exercise, play, and educational opportunities during childhood and adolescence) and
by the culture and institutional order of one’s society (which determines one’s opportunities
for social and political participation). Resourcists can recognize, therefore, that persons’
developed capacities to derive benefit from resources are co-determined by their prior access
to resources.

In our world, social factors play a huge role in explaining actual mental and physical
differences relevant to persons’ capacity to derive benefit from resources. Vast numbers of
people grow up in conditions where they and their mothers are chronically undernourished
and have inadequate access to basic (including maternal and perinatal) medical care,
immunizations, essential nutrients and safe water. As a consequence, many of them become
mentally and physically stunted and especially vulnerable to diseases. Not having had access
to even a minimal education, hundreds of millions of adults are illiterate. Due to sexist
cultural traditions and practices and sexist institutional arrangements, women are
significantly overrepresented in all these horrifying statistics.

Resourcist views have every reason to take account of interpersonal differences insofar as
these are due — as in fact they overwhelmingly are in the world at large — to past
inequalities in the resource access persons had under their institutional order. Nearly all

30
Some of the complexities and controversies are interestingly discussed in Dennis McKerlie: “Equality and
Time,” Ethics 99 (1989), 475-491, Dennis McKerlie: “Equality Between Age-Groups,” Philosophy and
Public Affairs 21 (1992), 475-495, and Larry S. Temkin: Inequality (Oxford: Oxford University Press 1993),
chapter 8.

pogge 13
persons with special mental or physical needs or disabilities today would be perfectly capable
of leading happy and healthy lives if they were not suffering the effects of severe past (and
present) resource deprivation: lack of effective civil and political rights and inadequate
access to water, food, shelter, health care, and education. Insofar as such resource deprivation
was unjust, a just institutional order will compensate for its effects. The members of a society
or other social system have a duty of justice to mitigate the harms caused by their prior
wrongdoing.

These problems are addressed more compellingly by resourcists than by capability theorists.
Where the latter criticize institutional schemes for their failure to compensate for special
physical and mental frailties, resourcists more powerfully criticize the same institutional
schemes for their failure to compensate for frailties they themselves produce through the
severe mistreatment they impose on so many children and adults.

This criticism is closely linked with another: The reform of such institutional schemes must
not merely ensure that they compensate for frailties they had produced in the past (insofar as
doing so is still reasonably possible), but also that they no longer produce and reinforce such
frailties. This last point is crucial: In order to count as just, an institutional order must not
merely repair, and compensate for, the effects that deprivations it inflicts have on the
physical and mental constitution of its participants — which all too often cannot be done
fully or even at all. It must also and most importantly avoid causing such damage in the first
place insofar as this is reasonably possible.

A plausible capability view will agree. It will hold that an institutional order producing
childhood malnutrition, for instance, deprives its victims of the capability adequately to
develop their physical and mental faculties. And it can then, like a resourcist view, add that
those responsible for maintaining an unjust institutional order that inflicts excessive
deprivations on some of its participants acquire a secondary duty of justice to repair and to
compensate insofar as they reasonably can. Both approaches thus can appropriately respond
to these very real problems.

Sometimes severe deprivations are inflicted by a just institutional order, as when a

motorized-traffic system engenders accidents in which innocent persons get hurt. Here as
well, a plausible resourcist view can recognize a collective remedial duty to repair and to
compensate even for burdens of social cooperation that are not reasonably avoidable. We
authorize and partly finance the traffic system through the democratic political process. And

pogge 14
while we can agree for ourselves to bear the cost of any harm this system may inflict upon
us, we cannot so agree on behalf of others, such as children too young to consent. Justice
requires then that we organize our institutional order so that, when children get hurt in traffic
accidents, funds are available to pay for whatever can reasonably be done to restore their
health and appearance. It may well be true that our modern traffic systems bring vastly
greater gains than losses. But it is not enough that the gainers should be able to compensate
the losers (the Caldor-Hicks criterion). We are entitled at most only to the net gains that
remain after the losers have actually been compensated to the extent this is reasonably
possible.

Two conclusions emerge from this discussion. First, we have learned that yet another
important class of “personal heterogeneities,” namely those that are socially caused, can be
fully taken into account by resourcist views. Second, we have seen that, on a resourcist view,
the causal origins of special needs and disabilities are morally significant. A woman who
became disabled in a childhood accident, paying the statistically predictable price of modern
transportation systems that serve our convenience and affluence, has a stronger claim to
compensation than another whose otherwise equivalent disability is not due to social factors.
Interestingly, Sen indicates that he may agree with this second conclusion:

A person’s capability may be reduced in exactly the same way in two cases:
(1) through a violation of his [sic] liberty (by someone violating her freedom
over a personal domain), and (2) through some internal debilitation that she
suffers. Even though the two cases are not distinguishable in the capability
space, an adequate theory of justice cannot really ignore the differences
between the two cases. In this sense, the capability perspective, central as it
is for a theory of justice, cannot be entirely adequate for it.31

(5.4) While the capability approach may, as Sen suggests in this passage, not be sensitive
enough to the causal origins of special needs and disabilities, the resourcist approach may be
too sensitive to such causal origins by ignoring completely any differences in participants’
mental and physical needs and endowments in whose emergence social institutions are not

31
Sen: Inequality Reexamined, 87; cf. Development as Freedom, 77. Even if this difference is morally
relevant, so that persons in the first category but not those in the second can demand compensation as a right,
we may still want to design our institutional order so that it treats both alike, especially when the distinction is
less than clear-cut or there are only a few cases in the second category. It would not be unjust for a society to
do as much for congenitally blind children as it must do for children whose eyesight was lost due to dangerous
medications or pollutants.

pogge 15
materially involved. Such differences could be due to any combination of three other kinds
of factors: ordinary genetic variations, as when some people have lesser native intelligence or
a faster metabolic rate; self-caused factors, as when people ruin their bodies or minds through
drugs or lack of exercise; and differential luck, as when someone is hurt by lightning or in a
natural disaster.

Resourcists can make three responses to this worry. First, they can point out that special
needs and disabilities in which social causes play no role at all are rare. Most special needs
and disabilities are due to a confluence of diverse causes, social ones included. Social rules
generating excessive poverty may, by forcing many people to live in flimsy huts or at
exposed locations, magnify the harm done by a natural disaster. Social rules governing
advertising may contribute to the damage adolescent smokers are doing to their bodies. And
rules governing social stratification and social mobility influence mating patterns and thereby
possibly the frequency of specific genetic characteristics. The question how a resourcist view
should plausibly be responsive to these complex interdependencies is evidently far too
difficult to be addressed here. But mere awareness of these complexities should suffice to
show that resourcists are not straightforwardly compelled to ignore a wide range of personal
heterogeneities.

Second, resourcists can say that the impact of interpersonal variations in needs and
endowments on individuals is very significantly influenced by the institutional order as well
as by social practices and cultural traditions. To take an obvious example: The mobility of a
walking-impaired person may greatly depend on whether our buildings and public
transportation system are wheelchair accessible. This general point is relevant insofar as
resourcists base their selection, formulation, and weighting of valuable resources on some
account of standard human needs. In doing so, they must avoid the complaint that this
account is modeled mainly on the needs of some and much less appropriate to the needs of
others. We discussed a feminist variant of this complaint above — the charge that some
resourcist views are male centered by employing metrics modeled on the needs of males
considered prototypical of human beings. Resourcist views must avoid analogous complaints
by the disabled: If a resourcist criterion of social justice is to ensure that any institutional
order satisfying it affords all its participants genuinely equal treatment, then its resource
metric must take account of the full range of diverse human needs and endowments. Rawls
tries to solve this problem by working with abstractly defined resources, such as basic

pogge 16
liberties and income, which, as “all-purpose means,” can be usefully employed by persons
with quite different needs and endowments.32

Insofar as a resourcist view satisfies this condition, it has a more attractive way of
accommodating special needs than capability theorists do. The latter say to the disabled
person: “I understand that you have a lesser capacity to convert resources into valuable
functionings. For this reason, we will ensure that you get more resources than others as
compensation for your disability. In doing so, our objective is that, by converting your larger
bundle of resources, you will be able to reach roughly the same level of capability as the rest
of us, that you will be roughly as able as we are to attain the various valuable functionings.”
The resourcist might say instead: “I understand that the present organization of our society is
less appropriate to your mental and physical constitution than to those of most of your fellow
citizens. In this sense, our shared institutional order is not affording you genuinely equal
treatment. To make up for the ways in which we are treating you worse than most others, we
propose to treat you better than them in other respects. For example, to make up for the fact
that traffic instructions are communicated through visible but inaudible signals, we will
provide free guide dogs to the blind. In doing so, our objective is that our institutional order
as a whole should afford you genuinely equal treatment.” This resourcist accommodation
does not, however, apply in all cases — if traffic instructions are communicated through both
visible and audible signals, free guide dogs may not be required by justice. And this
resourcist accommodation rarely results in equal capabilities, as a formerly disadvantaged
group, once it is being treated equally, may still have below-average capabilities — the blind
may still not get around as well as the seeing.33

Third, resourcists can also point out that justice is not the only moral virtue of persons and
institutional schemes. Thus, even if there are some among us who have special needs and
disabilities that the rest of us have no duties of justice to alleviate — because we did not
contribute to their emergence and do not benefit from their existence — we may nonetheless
have other quite stringent moral reasons individually or collectively to engage in such
alleviation: duties of humanity toward all human beings, for example, or duties of solidarity
toward those with whom we share a common political life. Such duties are appealed to when

32
Rawls: A Theory of Justice, 54f., 78-81. Rawls also needs to show that his account of the importance of
these goods relative to one another (as expressed in the lexical priority ranking of his principles of justice) is
equally appropriate for persons with different needs and endowments.

pogge 17
it is said that, though hikers, swimmers, and sailors have no right to be rescued from life-
threatening situations at public expense, it would nonetheless be morally intolerable for a
reasonably affluent society simply to let them die. It would be similarly intolerable for such a
society to allow poor persons with congenital health problems to die untreated.

At the end of this long section, it appears that the resourcist approach is no less able to
address most of the important deprivations and inequalities that so disfigure our world —
though, as we have seen, the two approaches may address them somewhat differently. What
remains as the key theoretical difference is how the two approaches direct institutional
schemes to respond to what one might call pure personal heterogeneities. The word “pure”
here serves to exclude features that, it has turned out, are not genuine personal
heterogeneities at all and should not be ignored by resourcists: lactation, pregnancy, and age.
It is meant to exclude also personal heterogeneities to the genesis of which social factors
substantially contribute — such as the effects of accidents produced by a socially sanctioned
motorized traffic system or the genetic effects of socially sanctioned drugs (thalidomide) or
pollution. To avoid the rather cumbersome language of pure personal heterogeneities, let me
rephrase the topic as how institutional schemes are to respond to natural human diversity,
with the reminder that such natural diversity may arise from any combination of ordinary
genetic variations, self-caused factors, and differential luck.

2. The Real Contrast: Compensation for Natural Differences

Regarding the treatment of natural human diversity, institutional schemes and the criteria of
social justice informing them may be said to be defective in three distinct respects. Only
complaints of the last of these kinds are controversial between the two approaches, and I will
therefore focus on them alone.

Intrinsic discrimination complaints allege that an institutional order is unjustly biased against
persons with certain natural features and demand that such bias be mitigated or removed.
Such a complaint may allege overt bias involving social rules or criteria of social justice that
openly discriminate on the basis of natural features such as skin color or sex. Or it may allege
covert bias involving facially neutral social rules or criteria of social justice that nonetheless
systematically disadvantage persons with certain natural features, such as blacks, women, or

33
It is also possible, of course, that such a formerly disadvantaged group, once treated equally, has above-
average capabilities.

pogge 18
the walking-impaired. Instances of such bias were discussed above (section 1, (5.1) and
(5.4)). Complaints alleging (especially covert) bias can be quite difficult to resolve. But the
problems they pose are equally serious for both approaches.

Extrinsic discrimination complaints allege that an institutional order unjustly fails to be

(sufficiently) biased in favor of persons who are suffering social disadvantages due to
historical or cultural discrimination based on certain natural features. Here the complaint is
not that the institutional order is itself discriminatory, but that it fails (sufficiently) to
dislodge, or to redress the effects of, discrimination extrinsic to itself. To cope well with such
extrinsic discrimination, an institutional order must be shaped in light of three, possibly
competing desiderata: to mitigate the more grievous and pervasive effects of present and past
discrimination, to help dissolve prejudice and discrimination in present practices, attitudes, or
ways of life, and to avoid imposing new discriminatory burdens on participants who are
innocent of the discrimination to be alleviated. This complex task poses equally difficult and
serious problems for resourcists and capability theorists alike, and I can therefore leave it
aside as well.

Compensation complaints allege that an institutional order unjustly fails to be (sufficiently)

biased in favor of persons who are disfavored by what Rawls has called the natural lottery.34
While capability theorists approve of some such complaints and believe that institutional
schemes ought to be designed to even out some natural inequality, resourcists do not.
Capability theorists seek an institutional order under which resources are so distributed that
the resulting distributive pattern of individual capabilities — dependent on individual
endowments and resources — satisfies their preferred criterion. Resourcists, by contrast, seek
an institutional order under which the distributive pattern of resource access satisfies their
preferred criterion. They pay no attention to how this distribution correlates with the
distribution of natural features. How can we make progress toward resolving this dispute?

A compelling ad hominem argument can be made on behalf of the capability approach

specifically against Rawls’s resourcist criterion. The argument shows that the parties in the
original position, as Rawls himself describes them, would favor a criterion of social justice
that is sensitive to natural inequality in the way the capability approach envisions.

Recall Rawls’s central thesis that we should morally endorse that public criterion of social
justice which parties in the original position would endorse prudentially. He imagines these

pogge 19
parties to deliberate on behalf of prospective participants — but behind a veil of ignorance,
so that they know nothing specific about the particular persons they represent. The parties are
made to assume, however, that every prospective participant has three interests, which Rawls
sees as closely connected to their role as citizens in a democratic society (and hence as not
being partisan to any particular religious, philosophical, or ethical worldview or way of life).

Rawls calls these the three higher-order (or fundamental) interests, suggesting both that they
are interests in the content and fulfillment of other interests (like second-order desires are
desires about desires) and also that they are deep, stable, and normally decisive. The first two
higher-order interests are interests in developing and exercising two moral powers,35 namely
“the capacity for a sense of justice: ... to understand, to apply, and to act from (and not
merely in accordance with) the principles of political justice that specify the fair terms of
cooperation [and] a capacity ... to have, to revise, and rationally to pursue a conception of the
good.”36 The third higher-order interest is “to protect and advance some determinate (but
unspecified) conceptions of the good over a complete life,”37 that is, the interest to be
successful in the pursuit of one’s final ends and aims, whatever these may be.

Now it is evident that the extent to which a person’s higher-order interests — especially the
third — are fulfilled depends not merely on her access to resources or social goods but also
on her natural endowments. A person with energy, intelligence, good health, and good looks
is better able to advance her conception of the good than someone who is sickly, gloomy,
unintelligent, and ugly. Even if the parties are made to assume that the institutional order
cannot affect such natural human diversity itself, they would still want this order to promote
the next best thing: a negative correlation between the quality of individuals’ natural
endowments and their access to resources. In particular, the parties will want the institutional
order to ensure that no one scores poorly in regard to both natural endowments and resources
access. By thus choosing a public criterion of social justice pursuant to which the naturally
disfavored are favored in the distribution of social goods and ills, the parties can best protect
the higher-order interests of their clients. Hence they would choose such a compensatory

34
Rawls: Collected Papers, 82; A Theory of Justice, 64.
35
Rawls: Political Liberalism, 74; cf. Justice as Fairness: A Restatement (Cambridge MA: Harvard
University Press 2001), 192.
36
Rawls: Justice as Fairness, 18f.; cf. Political Liberalism, 19.
37
Rawls: Political Liberalism, 74; cf. Justice as Fairness, 192.

pogge 20
criterion — perhaps one involving a capability metric — over the resourcist criterion Rawls
claims they would choose.

If this powerful argument does not move me to abandon resourcism, it is because I see
compelling independent reasons for rejecting the thought experiment of the original position
and the purely recipient-oriented theorizing it so elegantly encapsulates.38 Still, the interests
of recipients — of those who live under the institutional order to be assessed — do matter, of
course, and I should therefore be able to explain why I resist the move from a resourcist to a
capability metric despite the fact that it seems to improve participants’ ex ante prospects. The
following four subsections give my reasons.

2.1 How Resourcism Is Sensitive After All to Unequal Work-Relevant Endowments

My first reason only dilutes the argument for preferring the capability approach. It shows that
while the resourcist approach cannot achieve as nuanced an accommodation of human
natural diversity, it has a good bit more to offer than nothing at all. To bring the points out in
a clear and efficient way, I display them in the context of a stylized economic scenario in
which the approaches contrast starkly.

Imagine then a group of adults of working age stranded together on an island and now
cooperating in food production. In asking how the food acquired through their work should
be distributed among them as income, I assume at first that these adults are such that equal
income affords them equal capabilities. Their mental and physical constitutions are similar
enough or else the pluses and minuses cancel out so that there is no vertical overall
inequality (rendering some evidently better-endowed than others). This highly artificial case
lacks then exactly that empirical feature whose treatment resourcists and capability theorists
really disagree about.

Let us also concentrate on criteria of social justice that are either equalitarian or prioritarian
or sufficientarian or some hybrid of any two of these or of all three. What space or metric or

38
A conception of social justice is purely recipient-oriented if and only if its assessments are based
exclusively on information about the pattern of shares expected to exist under alternative feasible institutional
schemes. In particular, purely recipient-oriented conceptions attach no moral significance to information about
the character of the causal relation between institutional schemes, on the one hand, and individual benefits and
burdens, on the other. Such conceptions attach no moral significance, for instance, to whether an institutional
order gives rise to particular burdens by mandating, authorizing, or engendering them, or through failures to
prevent, to deter, or to compensate. Cf. Pogge: World Poverty and Human Rights, second edition (Cambridge:
Polity Press 2008), 45-50, and John Rawls: His Life and Theory of Justice (New York: Oxford University
Press 2007), 43-53. I criticize there the original position for implicitly endorsing the anonymity condition and
for its purely recipient-oriented theorizing that leads to highly counterintuitive demands on penal institutions.

pogge 21
currency should one want to incorporate into such a criterion for this simple case? There are
three main possibilities.

One straightforward idea is to use the space of income.39 The obvious objection to such an
income metric is that it ignores variations in the amount of work different persons contribute:
The assessment of feasible alternative institutional schemes is based entirely on the profile of
income shares (average annual income over a lifetime, say) each tends to engender, without
regard to how this income profile is correlated with the profile of work contributions. An
equalitarian theory using this space would favor an institutional order under which persons
can decline to work and still be entitled to an equal share of all the food produced by the
others. Under the rules that prioritarian and many sufficientarian theories using this metric
would favor, there would likely be differential rewards as work incentives. But these
incentives would be designed so as to maximize the lowest average annual income (or,
equivalently, so as to minimize the greatest shortfall from sufficiency) and would therefore
lead to a distribution that rewards less the work of those who work more. These people doing
more work may well complain that they are receiving less than equal pay for equal work —
and this for the sake of giving far too good a deal to those who do little work or none at all.

An income metric is certainly defensible nonetheless, especially in conjunction with a

prioritarian interpersonal aggregation function. In fact, Philippe van Parijs has done an
impressive job defending, within the technological and socioeconomic context of
contemporary continental Europe, a conception of social justice that is prioritarian in the
space of income.40

Responding to some remarks by Richard Musgrave,41 Rawls has decided to settle on another
possibility: using the space of some index of income and time worked. He proposes a
specific income/worktime metric that stipulates leisure as another social primary good and
considers those unwilling to put in a full standard workday as having extra leisure.42 This
proposal involves some arbitrariness regarding the length of the standard work day and
regarding the (intrapersonal) aggregation function that sums any quantities of these two

39
Income will be understood as net of taxes and net of expenses that must be incurred by those holding the job
(e.g., warm clothing for that job in Alaska).
40
Philippe van Parijs: Real Freedom for All (New York: Oxford University Press 1995).
41
R. A. Musgrave: “Maximin, Uncertainty, and the Leisure Trade-Off,” Quarterly Journal of Economics 88
(1974), 625-632.
42
Rawls: Collected Papers, 455 n. 7, cf. 252f.

pogge 22
goods (income and leisure) into a single index number. To keep things perspicuous, let us
here consider a more elegant income/worktime metric, which avoids both sources of
arbitrariness by defining the index simply as the ratio of income over hours worked. This
definition is problematic in various ways,43 but it will do well enough for our purposes here
if we add the assumption that all islanders have the same freedom to vary the number of
hours they work at their respective personal rate.44

Such an income/worktime metric faces a milder version of the same objection we saw aimed
at an income metric: Just as it matters how many hours of work persons put in, one might
say, so it also matters how hard or productively they work during those hours. This problem
is clearest when an income/worktime metric is used within an equalitarian theory. Such a
theory favors an institutional order under which workers are entitled to the same hourly
income regardless of how hard they work. To be sure, prioritarian and many sufficientarian
theories are likely to favor incentives that reward hard work. But such incentives would be
designed to maximize the lowest income per hour; so work done slowly will tend to be better
rewarded than the same amount of work done more efficiently. Hard-working people may
well then complain that they are receiving less than equal pay for equal work — and this for
the sake of giving far too good a deal to those who work much below their potential.

There is a related but different objection to an income/worktime metric: that it inconsistently

treats differently two factors that are really on a par: how long people work and how hard
people work. Why should someone who works very hard for only two hours each day be
treated any differently from how she would be treated, if she stretched the same work over
four hours (i.e., if she worked twice as long and half as hard)?

43
Cf. Pogge: Realizing Rawls, 66f., 198f. For example, the assumption that persons working at a higher
hourly wage rate are better off than others working at a lower hourly wage rate may be implausible when the
former cannot choose freely how many hours to put in. In the real world, many jobs are all-or-nothing affairs.
And being free to work either sixty or zero hours weekly at $4 per hour may not be better than being free to
work as many or as few hours as one likes at $3 per hour. And even if persons are free to vary their hours at
will, there may still be important differences, as two persons each earning $4 per hour may face quite different
options at the margins: One may be doing twenty hours a week of unpaid work (caring for children or for a
sick or elderly person) plus twenty hours a week at $8/hour with the option of increasing or decreasing the
latter hours, while another may be working twenty hours a week at $5/hour plus another ten hours at $2/hour
with the option of increasing or decreasing the latter hours.
44
This freedom is, of course, quite rare in the real world. Also, the hourly wage rate available to a person may
be sensitive to how many hours they work (e.g., on account of progressive taxation). And it may also vary
from year to year, which raises difficult issues of intertemporal aggregation (being paid $3/hour during one
decade, then $4/hour during the next, and $5/hour during the one after that makes for a less adequate lifetime
share than being paid $4/hour during all three decades). I will leave such variations aside.

pogge 23
These objections suggest conceiving the cost of cooperation not as time worked, but as
productive contribution, proportional to both how long and how hard persons are working.
This is the third possibility: an income/contribution metric. A simple example of such a
metric is the ratio of income over productive contribution. To be sure, productive
contribution is not an easy concept to operationalize within a complex network of
cooperative interaction: A free-market system in which contributions are valued at what
others are willing to pay for them can at best be a rough approximation, notoriously
undervaluing contributions (by nurses, teachers, homemakers) that are partly motivated
altruistically. Moreover, actual such systems are often quite unfree and noncompetitive in
some respects so that it would be preposterous to infer from current gross income data that
ten US chief executives contribute as much to the global social product as one million
Bangladeshi seamstresses.45 Still we can, in the context of the simple island case, waive these
difficulties for purposes of clarifying theoretically the more prominent metric options.

I have described a very simple world in which there are only two relevant individual
parameters (food and work) and in which all persons have roughly the same mental and
physical constitution, the same endowments, needs, and conversion capacities. And I have
distinguished nine pure criteria of social justice for this world, each constructed of two
components: an interpersonal aggregation function (equalitarian, prioritarian, or
sufficientarian) and a metric for defining and comparing individual shares (income,
income/worktime, or income/contribution). These nine criteria represent the basic options for
adherents of both the capability and resourcist approaches. These approaches have not parted
company thus far because we have not yet introduced into the story any diversities in overall
natural endowments which capability theorists deem fit for compensation. Before doing so,
let me make three orienting remarks.

First, we can expect that, when such vertical natural inequality makes its appearance,
resourcists will still endorse one of the nine criteria of social justice outlined above.
Capability theorists, by contrast, will want to modify the criterion they have advocated for
the simple world so as to incorporate vertical natural inequality. For now let’s assume they
do this not by changing their preferred interpersonal aggregation function (equalitarian,

45
To mention just the most obvious factors: Workers from poor countries are not free to offer their services in
rich countries. What they produce cannot be freely sold in rich countries, but is subject to quotas, tariffs, anti-
dumping duties, and competition from unfairly subsidized rich-country products. And US CEOs play a major
role in appointing members to the boards that determine their salaries, stock options, and other perks.

pogge 24
prioritarian, or sufficientarian), but by modifying only the metric they had advocated for the
simple case. Here is an example of such modification: An equalitarian in the space of
income, learning that persons have different metabolic rates, may change her metric from
income to income/metabolic-rate. She may then require, perhaps, that each islander should
share in the social product in proportion to his or her metabolic food needs. Thus, if the
social product is 115 percent of what is needed for all islanders to be minimally well
nourished, then just rules should assign to every islander 115 percent of what he or she needs
to be minimally well nourished, rather than equal amounts of food to each.

Second, how a capability theorist will be moved to modify her metric will depend on what
metric she had preferred in the absence of vertical natural inequality. It is therefore somewhat
unfortunate that the views set forth by Sen and Nussbaum are not specific enough to address
even the simple case (where vertical natural inequality is assumed absent). If they were
committed to one of the nine possible criteria (or some hybrid), then we could concentrate
our efforts on examining how this criterion might be modified in response to the introduction
of vertical natural inequality. Rawls’s choice from among the nine possible criteria of social
justice for our simple island world seems, by contrast, quite clear: He would want alternative
feasible economic schemes to be assessed by a criterion that is prioritarian in a space of
income/worktime: The islanders should organize their economy so that the lowest hourly
income rate is as high as possible.46

Third, and most important, although resourcists refuse to incorporate reference to vertical
natural inequality into their chosen metric, their nine alternative criteria nonetheless differ in
the relative benefits and burdens they bring to persons with various natural endowments. For
example, a person with poor work-relevant natural endowments will fare much better if the
economy is governed by a prioritarian income/worktime criterion than if it is governed by a
prioritarian income/contribution criterion.

46
This statement simplifies by leaving aside constraints and expenses arising from maintaining the equal basic
liberties and fair equality of opportunity as well as democratically sanctioned expenses for, and opportunity
costs arising from, public goods such as national parks. Also, at issue here are not income rates available to
persons at some particular time, but some aggregate of such available rates over a person’s lifetime (cf. note
44).

pogge 25
With these remarks in mind, let us now set aside our supposition that all islanders of working
age have roughly equivalent natural endowments by introducing (what capability theorists
would regard as) vertical natural inequality.47

Let us begin with the fact that persons differ in how much effort the work they do requires of
them. By “effort” I mean something quite distinct from how intensively or hard persons are
working. Intensity is an objective matter. A person works very hard or intensively when her
average contribution per hour of work is above 80 percent of her capacity, say. Effort,
however, is a subjective matter: Some persons enjoy working and enjoy working hard. They
like working 60 hours a week and also find it quite frustrating to work at less than 70 percent
of their capacity. For them, working very hard at 90 percent of their capacity is rewarding
and requires no special effort or sacrifice. Others are much happier taking it easy. They work
only because they must, and they find that contributing at above 30 percent of their capacity
requires a real effort on their part and that going above 50 percent is positively painful.

Suppose Ann and Sam are equally productive, working equally hard, for the same number of
hours, in the same job. And suppose Ann likes her work while it is painful for Sam. Should
we say that their institutional order is treating Ann and Sam equally only if Sam’s net income
is higher than Ann’s? Arranging such earning differentials may well turn out to be practically
infeasible because of theoretical difficulties about interpersonal comparisons of reluctance
and practical difficulties about dissimulation and perverse incentives. Still, it is interesting to
ask whether one should want to incorporate this natural difference into one’s metric if this
could be done without undermining its workability.

The welfarist answer is a clear yes. Equalitarians, prioritarians, and sufficientarians in the
space of welfare would ideally want the institutional order to be sensitive to the larger
welfare losses Sam incurs from working. An equalitarian welfarist, for example, will want
more income to be channeled to Sam, whose work costs him greater effort, so that his extra
pain from work is compensated by greater pleasures available through extra income. This
reduces the income of Ann, who works hard with little effort or sacrifice and can thus reach
the same welfare level with much less income.

47
We may need the reminder that natural inequalities are defined as ones to which relevant social factors do
not substantially contribute. We might circumvent complications in our simple case by thinking of the
islanders as freshly thrown together by fate and hailing from different parts of the world so that they bear no
responsibility for one another’s physical and mental development.

pogge 26
The resourcist answer is a clear no. Resourcists look at what persons give to and get from
society in objective terms and do not “look behind” the inputs and outputs to discern how
they may affect a person’s subjective states.48 And yet, resourcist views differ in their
implications for persons who differ in their reluctance to work. Except for those with
outstanding work-relevant endowments, persons who like taking it easy will do best under an
income metric and worst under an income/contribution metric. The opposite is true for those
who like hard work. An income/worktime metric, combined with prioritarianism, is a
compromise in that there are rewards for hard work (rewards that those who like hard work
find it easier to capture), but these rewards are designed to maximize the lowest income per
hour (often fetched by those who follow their desire to take it easy). In a Rawlsian society,
the productive contribution of reluctant workers — choosing to work more slowly and/or in
lower-paying jobs — will generally be more highly rewarded than equivalent productive
contributions by those who like hard work.

The answer of the capability approach is unclear. On the one hand, its proponents present it
as an objective approach, as is brought out, for instance, in Sen’s acceptance of Cohen’s
characterization of valuable functionings as midfare.49 On the other hand, they also introduce
subjective capabilities, such as “being happy”50 and “being able to have pleasurable
experiences, and to avoid non-necessary pain.”51 If pleasurable experiences and avoidance of
unnecessary pain are to be included among valuable functionings, then job satisfaction can
hardly be ignored.

Let us turn next to natural differences in persons’ work-relevant natural endowments. Human
beings have a vast number of different talents and aptitudes. Many of these — strength,
stamina, dexterity, intelligence, and so forth — are relevant to what they might contribute to
the social product. Because our endowments are so multidimensional, it is generally true, for
any two persons, that one is more talented than the other in some respects and less talented in
others. Still, it may nonetheless often be true that one has, on the whole, better work-relevant
natural endowments than the other with respect to the existing system of economic
cooperation. This means, roughly, that if each chooses from among what are the most
lucrative jobs for him or her, and if both work equally hard, then one will make a greater

48
Cf. Rawls: A Theory of Justice, 80.
49
Cf. supra text at notes 4 and 5.
50
Sen, Inequality Reexamined, 13.
51
Nussbaum: Women and Human Development, 79.

pogge 27
marginal contribution to the social product than the other. And the question is then how, if at
all, a criterion of social justice should accommodate such endowment-induced interpersonal
productivity differentials in its metric.

We can bring this question into a somewhat more perspicuous form by means of a simple
mathematical model in which any person’s productive contribution is analyzed as the product
of three factors: how long (average annual number of hours) she works (L), how hard she
works on average (H), and how good the work-relevant natural endowments are that she
brings to bear (N): C=LHN. It is uncontroversial that individual shares should be defined to
be sensitive to income: Other things being equal, persons who receive more income than
others under the same institutional order do count as being treated better by this order. It is,
as we have seen, controversial whether individual shares should be defined to be sensitive to
L (and perhaps H as well), so that persons who receive more income than others under some
institutional order may nonetheless not count as being treated better by this order if they also
contribute more by way of L (or LH). The new question is whether individual shares should
be defined to be sensitive also to work-relevant natural endowments, so that persons who
receive more income than others with the same L and H may nonetheless not count as being
treated better if they also contribute more by virtue of a higher N.

Theorists who hold, like van Parijs, that individual shares should be defined without regard
to contribution factors L and H will find this an easy question to answer: Contribution factor
N should also be ignored. If persons with lower incomes count as worse off even if they
choose to work less long and less hard, then surely persons with lower incomes must still
count as worse off even if they bring to bear lesser work-relevant natural endowments. The
reason is that income inequalities arising from work-relevant natural endowments are less
justifiable than income inequalities arising from how long or how hard persons work. Why
so? Consider a scheme of rules under which only half as much income goes to one person
than to another who, because she works longer and harder, gets twice as much work done.
These rules can be justified to the first person by pointing out that he is quite free to work as
long and as hard, and then to earn as much, as the second one does. But this justification is
unavailable for a scheme of rules under which only half as much income goes to one person
than to another who, because she can bring to bear greater work-relevant natural
endowments, gets twice as much work done (without working any harder).

pogge 28
The question is more difficult to answer for theorists at the other end of the spectrum, who
endorsed an income/contribution metric for the simple case. The more libertarian of them
will feel that an institutional order is not treating one person worse than another who earns
twice as much for twice the work contribution, even if the latter’s greater contribution is due
entirely to greater work-relevant natural endowments. They believe that while income
inequalities arising from work-relevant natural endowments may be less justifiable than
income inequalities arising from how long or how hard persons work, they are still justifiable
enough. N may be incorporated in the same way as H and L, and there is then no need to
modify the income/contribution metric preferred for the simple case.

A theorist who favors the income/contribution metric for the simple case but believes that
persons’ lack of control over their work-relevant natural endowments is morally too
significant to be ignored will naturally be drawn to an income/HL metric.52 But this proposal
may not be informationally workable. In the real world, it is all but impossible to tell to what
extent observed differences in persons’ productive contributions are due to how hard they
work or due to their work-relevant natural endowments. The problem here is not merely that
of estimating how much more productive each person could be in the job he or she is in. In a
world with significant differences in work-relevant natural endowments, it also matters how
much more productive each person could be in other jobs. Persons generally have control
over how much they contribute to the social product not only through their choice of how
hard they work in their job, but also through their choice of job: They can choose between
jobs for which they are relatively talented and jobs for which they are relatively untalented
(that is, between jobs in which they can outperform a larger/smaller percentage of their
peers). And they can choose between jobs that tend to contribute more to the social product
and jobs that tend to contribute less. It makes sense to treat such career choices as an element

52
Such a metric may also seem implausible. Consider two persons working equally productively in the same
job for the same number of hours per year. Jean could be twice as productive (in the same or another job),
while Bob is as productive as he can be. In this case, Bob counts as working twice as hard as Jean; and a
system of rules would be treating them equally, by the lights of an income/HL metric, only if Bob had twice
Jean’s income under it. — There are also two further, somewhat less natural possibilities: an (f1(N) +
income)/contribution metric or an f2(N) + income/contribution metric, with f1 and f2 linear increasing
functions. The first of these metrics would achieve equal treatment when persons have gross pay proportional
to their contributions, which is then diminished by a percentage proportional to their work-relevant
endowments. The second of these would achieve equal treatment when persons have gross pay proportional to
their contributions, which is then diminished by a lumpsum tax (or “head tax”) proportional to their work-
relevant endowments. Each of these further possibilities entails that, to treat Jean and Bob equally, an
institutional order must ensure that Bob receives more income than Jean does for the same productive
contribution. The two further possibilities also both run into the informational problem sketched in the
remainder of the paragraph.

pogge 29
of H by defining how hard a person works, during a given work period, as the ratio of what
she contributed to the social product over the most she could have contributed in this period.
Thus, a person who contributes all she can in a job she is not good at does not count as
working hard at all, if she could instead have worked in another job in which, contributing as
much as she can, she would have contributed much more. Estimates of how hard a person is
working are thus all but impossible because of their dependence on a large number of
difficult subjunctives.

Reverting to our simple mathematical model, a person’s productive contribution can be

measured (or at least estimated) and is known to be the product of three factors: the length of
time she works (L), how hard she works on average (H), and the quality of the work-relevant
natural endowments she brings to bear (N). C = LHN. Now, since L is measurable, it is
possible to back out HN as C/L. But it is generally not possible to determine H and N. It is
observable that a person contributes, on average, $20/h to the social product, say, but it is not
observable what her maximum average contribution would be. Could she maximally
contribute $40/h, so that H=0.5, or could she contribute $100/h, in which case H=0.2, or
could she contribute even more? This is quite difficult to know in an economy that is at all
complex. And it would be even more difficult to know in an economy where income rewards
are more sensitive to H than to N and where persons would therefore have an incentive to
dissimulate by playing up their H and downplaying their N.

Given this practical complication, theorists drawn to an income/HL metric thus face a
trilemma: They find it justified to make income sensitive to H and L, and much less justified
(if at all) to make income sensitive to N. But because H and N are empirically inseparable, it
is impossible to do justice to all of these ordinal relations. One can make income as sensitive
to N as income should be to H and L, thereby doing violence to the disparity between N on
the one hand and H and L on the other. This horn leads to the libertarian income/contribution
metric. One can make income as insensitive to H and L as income should be to N, thereby
doing violence to the same disparity between N on the one hand and H and L on the other.
This horn leads to an income metric à la van Parijs. Or one can make income as insensitive to
H as income should be to N, thereby doing violence to the disparity of H and N as well as to
the parity of L and H. This horn leads to the compromise of a Rawlsian income/worktime
metric, which looked inconsistent in the simple case but now gains in stature when seen as

pogge 30
such a practical compromise. The trilemma shows how an informational problem can
plausibly affect our choice from among candidate public criteria of social justice.

Concluding the discussion of work-relevant natural endowments, we find that although

resourcists do not incorporate this natural inequality into their preferred criteria of social
justice, their choice of one such criterion nonetheless has important implications for how
persons with diverse work-relevant natural endowments will fare. Here the difference
principle accommodates persons with lesser work-relevant natural endowments in two
important respects. First, it employs an income/worktime metric, under which those with
lesser work-relevant natural endowments will fare much better, in both relative and absolute
terms, than they would fare under an income/contribution metric. Second, the difference
principle is prioritarian. This means, on the one hand, that there will be income rewards that
may be inaccessible to those with lesser work-relevant natural endowments. But it also
means, on the other hand, that these incentives are constrained so as to optimize the lowest
socioeconomic position in which those with poor work-relevant natural endowments are
overrepresented. Thus Rawls can say with some justification: “the difference principle
represents, in effect, an agreement [that t]hose who have been favored by nature, whoever
they are, may gain from their good fortune only on terms that improve the situation of those
who have lost out.”53

Capability theorists may be moved to concede that a resourcist view like Rawls’s does cope
acceptably well with inequalities in work-relevant natural endowments. They may also
concede that it is acceptable for Rawls to make no allowance for those from whom work
requires greater effort. They see the main weakness of a Rawlsian view in its inability to
accommodate other natural inequalities. And they may even draw ammunition from my
qualified defense of a Rawlsian view by trying to argue that Rawls is inconsistently more
concerned to accommodate those with lesser work-relevant natural endowments than those
with greater natural needs. However, before considering whether such an argument may be
available, one would want to know more about the alternative that capability theorists want
to propose. How do they conceive the burdens of economic cooperation? How do they
propose to accommodate inequalities in work-relevant natural endowments and inequalities
in the amount of effort persons’ work requires of them? Insofar as capability theorists have

53
Rawls: A Theory of Justice, 87.

pogge 31
not offered any criterion of social justice specific enough to answer such questions, their
rationale for preferring the capability approach over resourcism remains obscure.

2.2 The Treatment of Natural Diversity: The Vertical-Inequality Problem

By seeking to shape the institutional order so that the distribution of resources it generates
compensates for natural inequality, capability theorists regard human natural diversity in
vertical terms and human beings as better or worse endowed. Foreswearing any such
compensatory ambitions, resourcists have no use for the very idea of greater or lesser human
endowments and can thus endorse a horizontal conception of human natural diversity. This
difference is of some consequence.

Human beings differ from one another in countless ways in their mental and physical
features. All our lives are immensely enriched by this variety. We can best share the delight
in our natural diversity when we think of it in horizontal terms. We then see persons as
different — in regard to the color of their eyes, for instance — without believing that having
green eyes is either better or worse than having brown ones. History would have gone vastly
better if prominent natural inequalities (sex, height, skin color) had always been seen as
horizontal.

To be sure, many natural inequalities can hardly be seen as horizontal. We speak of bad
posture, bad health, and bad memory and thereby explicitly deny that these are no worse than
their “good” counterparts. And many other predicates carry similarly strong negative
connotations: It is widely considered worse to be unintelligent, obese, balding, frail, tone-
deaf, or short than the opposite. These entrenched valuations are cultural to some extent, and
we may hope to eradicate some of them and to soften others. Still, we cannot realistically —
and perhaps should not — hope for a world in which no one admires some natural features in
another. That such a world is out of reach is no misfortune. For partial verticality —
confined to particular natural features or to particular judges — need not undermine the
shared public sense that human natural diversity overall is horizontal.

As we admire certain features in another — his full dark hair, beautiful eyes, amazing
memory, regal posture, or quickness of mind — we may also realize that we have mental and
physical gifts that the other might admire in us. And while we would be well pleased to trade
our height or memory for another’s, we would be quite reluctant indeed to trade in our full
package of natural features for his. To approve such a trade, to conclude with confidence that
the other is better endowed all things considered, we would not merely need to know in

pogge 32
detail all the myriad features of ourselves and of the other, but also how to value each of
these features relative to all the others. But how do you compare extra hair on your head with
extra musical talent? Good health with good looks? Acne with melancholy? A perfect
memory with perfect teeth? What “exchange rates” would you use?

Insofar as some of us have answers to some or even to all of these questions, their answers
tend to be shaped by their own features, and normally with a bias against the trade: Musical
persons tend to attach great importance to being musical, athletes to being athletic, brainy
people to brains, and — most notoriously — beautiful people to being beautiful.54 This
reinforces our reluctance to trade, our reluctance to envy. And our awareness of the great
diversity in our valuations and of the bias in favor of one’s own endowments militates
against the idea of a socially shared ranking of persons’ overall endowments. Looking at
each person’s full set of endowments from a shared social point of view we can sustain the
conception of natural inequality as horizontal. We can think of humankind as displaying a
wonderful natural diversity rather than, as was done through most of human history, a natural
hierarchy of persons more or less well-born.55 This thought powerfully reinforces the modern
ethos of democratic equality.

While the resourcist approach is supported by this conception of natural inequality as

horizontal, the capability approach requires that natural inequality be conceived as vertical.
When a capability theorist affirms that institutional schemes ought to be biased in favor of
certain persons on account of their natural endowments, she thereby advocates that these
endowments should be characterized as deficient and inferior, and those persons as naturally
disfavored and worse endowed — not just in this or that respect, but overall — not just in the
eyes of this or that observer, but in the eyes of the shared public criterion of social justice.56

Although conceiving human natural diversity as vertical is central to the capability approach,
its intent is opposite to the historically dominant one: Where Aristotle postulated natural

54
Health may be an exception in this regard in that persons in poor health tend to attach more importance to
health than healthy people do.
55
Acceptance of a resourcist criterion does not entail or presuppose these thoughts in celebration of human
natural diversity. My point is that resourcist criteria are, while capability criteria are not, compatible with
these sentiments.
56
As we just saw, Rawls also uses the language of vertical natural inequality (“favored by nature”). But such
notions play no role in the public criterion of justice he advocates. And their use could (and, I think, should)
be confined to contexts where resourcists respond to objections from welfarists and capability theorists. Rawls
might then be saying: “Those whom you call naturally disfavored and whom you want the institutional order
to compensate would actually fare rather well under the difference principle even without being singled out
for special compensatory benefits.”

pogge 33
hierarchy in order to claim greater resources for the better endowed, who alone, he thought,
were able to deploy them toward a truly good life, contemporary capability theorists
postulate natural hierarchy in order to claim greater resources for the worse endowed so as to
make up for their natural deficiencies and to provide them access to the full range of valuable
human functionings. While this concern for the naturally disfavored is noble, it is destructive
of any social conception of human natural diversity as horizontal. In this way, the return to a
natural hierarchy constitutes a social loss — not least for those who, singled out for special
compensatory efforts, are characterized as naturally disfavored or worse endowed.

The capability approach seeks to give to such a person a claim in justice, so she need not ask
for extra resources as a special favor, but can come forward proudly, with her head held high,
insisting on additional resources as her due. But the capability approach must then provide
for such claims a ground that subverts the point of the exercise. In order to justify her claim
in justice, the claimant is made to say that she is overall worse endowed than others. It is not
enough for her to point to one respect in which she has a special limitation, need, or handicap
that renders her worse off than most others are. For there are many other respects in which
the addressees of her claim have special limitations, needs, or handicaps, other respects in
which she may be better endowed than those she is addressing. To have a valid claim that she
is owed compensation as a matter of justice, she must present her special limitation, need, or
handicap as one that outweighs all other particular vertical inequalities and entitles her to
count as worse endowed all things considered.

Would you want to claim that your endowments are inferior, overall, than those of most
others? Would you want to be officially singled out by your society for special compensatory
benefits reserved for the “worse endowed”? Many disability groups resist such labeling,
pointing out that their disability opens realms of human experience and interaction that are
closed to “normal” persons. They see such labeling as undermining their demand that their
condition should not be an admissible ground for selective abortions or for the denial of
fertility treatment and reproductive counseling57 or for assigning them a lower quality of life
for purposes of allocating scarce medical resources.58 They understand that the force of the
label cannot be neatly confined to the grounding of compensation claims.

57
Within so-called Deaf communities, couples have demanded counseling and support for their efforts to
conceive children who will share their condition.
58
According to a widely used method, the achievements of a society’s medical system are measured in terms
of “quality-adjusted life years” (QALYs) or “disability-adjusted life years” (DALYs). This method assigns

pogge 34
Elizabeth Anderson has appealed to something much like the vertical-inequality problem by
way of showing that the capability approach is superior to Dworkin’s view and to
equalitarian welfarism and to variants upon these.59 She succeeds with this claim by tailoring
her own capability view in three significant respects: First, she favors a narrow, modest list
of capabilities: “Negatively, people are entitled to whatever capabilities are necessary to
enable them to avoid or escape entanglement in oppressive social relationships. Positively,
they are entitled to the capabilities necessary for functioning as an equal citizen in a
democratic state.”60 This contrasts dramatically with the wide-ranging and ambitious list
provided by Martha Nussbaum. Second, Anderson is, despite the somewhat misleading
injection of the word “equal,” a sufficientarian: Above some threshold, she counts persons as
functioning as equal citizens even if some are functioning much better than others in the very
respects in which no one is supposed to fall below the threshold.61 Third, Anderson not
merely fails to require compensation for natural inequality above the threshold, but opposes
such compensation. Her view thus is a hybrid which draws on a capability view for
specifying the threshold and for grounding justice claims below it, and on an (unspecified)
resourcist view for grounding justice claims above the threshold.

Having appealed to the vertical-inequality problem, Anderson focuses her diagnosis of

existing injustices on (intrinsic and extrinsic) discrimination complaints — a diagnosis that
resourcists can fully share. However, consistent with the capability component of her view,
she also echoes Sen’s mantra: “Because of differences in their internal capacities and social
situations, people are not equally able to convert resources into capabilities for functioning.
They are therefore entitled to different amounts of resources so they can enjoy freedom as
equals.”62 Remarkably, Anderson seems not to notice that this move exposes her own view to
the vertical-inequality problem she herself had quite polemically pressed against Dworkin
and the welfarists. The problem shows up when she requires the state to give free
wheelchairs to the walking-impaired on the ground that they would be functioning as less
than equal citizens without them. It is apparent also in Anderson’s discussion of the ugly,

greater value to the life years of “normal” than of “disabled” persons and thereby gives the former precedence
over the latter in the allocation of scarce medical resources.
59
Elizabeth S. Anderson: “What is the Point of Equality?” in Ethics 109 (1999), 287-337, at 305f and 316ff.
60
Ibid., 316.
61
Ibid., 318f. Nussbaum works with two capability thresholds. See her “Human Capabilities, Female Human
Beings” in Martha Nussbaum and Jonathan Glover, eds.: Women, Culture and Development (Oxford: Oxford
University Press 1995), 81ff.
62
Anderson: “What is the Point of Equality?,” 320.

pogge 35
who often lack the capability to appear in public without shame.63 Recognizing that practices
of stigmatizing and excluding the ugly cannot be wholly eradicated in a liberal society,
Anderson concludes: “Under these conditions, the better option may well be to supply the
plastic surgery.”64 And so the State Equal Citizenship Bureau of Anderson’s ideal society
will be writing letters to the handicapped and the ugly much like the letters she imagines
might be written in the ideal equalitarian society envisioned by Dworkin and the welfarists:
“Dear Naturally Disfavored, We have determined that, given your inferior natural
endowments, you must receive special compensatory state benefits in order to be capable of
functioning as what we could recognize as an equal citizen....”

Capability theorists may want to respond to this critique that it is less stigmatizing to be
given extra resources toward enhancing one’s otherwise deficient capabilities than toward
enhancing one’s otherwise deficient opportunities for welfare or toward honoring some
hypothetical insurance policy it is presumed one would have purchased. I look forward to
seeing such a response. Meanwhile, I conclude that only the resourcist approach avoids such
stigmatization altogether.

2.3 The Specificity Required of a Workable Public Criterion of Social Justice

Much intuitive appeal of the capability approach derives from the feeling that it is not fair
that some human beings are less healthy, bright, tall, handsome, athletic, or cheerful than
others or just need more food, water, or clothing. Wouldn’t the world be a better place if
human diversity were less vertical, as it were, with people more equal in their health and
needs, and with all people being bright, handsome, and athletic in their own diverse ways?
Failing this, would it not be good if those disfavored by nature at least had additional
resources as compensation?

This intuitive appeal is misleading. The question to which resourcists and capability theorists
give competing answers is not about what makes the world better or worse, but about what
makes institutional schemes more or less just. This latter question is different in many
respects. It is different, for instance, by involving a complementarity of relative gains and
losses. If persons with higher metabolic rates are to be entitled to free supplementary food,
then who will be made to bear to cost of its production? In thinking about the just design of

63
Anderson considers this capability, originating with Adam Smith and often invoked by Sen, to be necessary
for functioning as an equal citizen (ibid., 320).
64
Ibid., 336.

pogge 36
such institutional schemes, we must ask not merely whether we approve of the relative gains
they bring to the “naturally disfavored,” but also whether we can accept the relative losses
they bring to others. And we must ask whether proposed compensation rules achieve equity
among their beneficiaries with their diverse special needs, and equity also among their
contributors. Thus, in addition to the vertical-inequality problem, capability theorists also
face the difficult task of specifying a plausible such criterion of social justice in detail.

Capability theorists usually leave such questions aside. You can read thousands of pages of
their writings without finding any hint about how compensation is to be financed. As we
saw, Anderson seems opposed to taking natural inequality into account above the threshold.
But why? If justice requires special compensatory benefits for the naturally disfavored in
proportion to how disfavored they are, why should justice not also require special
compensatory burdens on the naturally favored in proportion to how favored they are? If a
just institutional order must treat those with an abnormally high metabolic rate better than
those with a normal one, why must it not also treat those with a normal metabolic rate better
than those with an abnormally low one? Conversely, if justice requires or permits
institutional schemes to treat persons above some capability threshold in resourcist terms
(e.g., by taxing them without regard to their particular needs and endowments), then doesn’t
this show that justice is not tied to a capability metric and that our moral concern to meet the
special needs of some is not a concern of justice?65

If each participant’s resource entitlements under a just institutional order are to be

determined in part by the overall quality of his or her natural endowments, then we need
some way of measuring such overall endowments: how naturally favored or disfavored each
participant is. Capability theorists will base their measure of human natural endowments on
their preferred capability metric: Roughly speaking, the more numerous and important are
the valuable capabilities that a particular human endowment helps persons to achieve, and
the greater the contribution that endowment makes to this achievement, the more weight it
merits within the overall measure of human natural endowment. It is obvious that it would be
extremely difficult in the contemporary world to reach agreement on the four key points: on
the list of valuable capabilities, on how to measure achievements with respect to each listed

65
A prioritarian capability theorist may have an answer to this challenge: “Justice requires the highest possible
capability threshold. Because a tax system taking natural inequality into account is more cumbersome and
expensive to operate than a conventional one based on income and/or consumption, the latter should be
preferred on the ground that it yields larger net revenues and thus enables a higher capability threshold.”

pogge 37
capability, on the relative weights of achievements in regard to different listed capabilities,
and on the relative value of diverse overall endowments in respect to overall capability
achievement.

Martha Nussbaum has gone further than anyone toward tackling the first task: toward
facilitating agreement on a widely acceptable list of valuable capabilities.66 Her extensive list
is thoughtful and well constructed. Some of the items she includes — having opportunities
for sexual satisfaction, being able to experience justified anger — may raise eyebrows. But
suppose we could agree on the whole list, or one like it. Could we agree on rules for
evaluating individual natural endowments on the basis of this list so as to compensate for the
greater or lesser suitability of participants’ overall endowments to their achievement of the
full range of valuable capabilities?

In pressing the enormous difficulty of the remaining three tasks, I am not critical of
Nussbaum’s list, but only of a certain use to which it might be put. Her list of capabilities is a
useful heuristic in the development of a resourcist criterion of social justice. It can help us
think of all the personal and public goods and supports that human beings need to flourish
fully, from the school curriculum to the organization of workplaces and organs of democratic
decision making. I have no objections to a list of capabilities being used in this role — not as
the metric within a public criterion of social justice, but as a useful guide in the development
of such a criterion (cf. note 14 supra).

Relatedly, an account of human capabilities can also play an important evidentiary role. The
observed fact that many persons are lacking certain vital functionings may be good reason to
revise our resourcist criterion of social justice. For example, if we find that many persons
living above the international poverty line (and thus counted as having sufficient income) are
in fact malnourished, we have strong reason to conclude that something has gone wrong. If
the observed malnutrition is concentrated in certain countries or years, we may need to
rethink the method used to translate the international poverty line from the base year and
base currency into other years and other national currencies. If the observed malnutrition is
concentrated among girls and women, we may have to jettison the empirical conjecture that a
household’s resources are normally shared equally among its members and may then need to
reform the institutional order so as to achieve a more equitable intra-family distribution of
resources. If malnutrition among the “non-poor” shows no patterns of these kinds, we may

66
Nussbaum: Women and Human Development, 78-80.

pogge 38
conclude that the international poverty line is set too low to express a plausible notion of
sufficiency.67 In these diverse ways, data about important functionings can, as Sen says,
provide a crucial check.68 Resources do not, after all, figure as ultimate ends in human lives,
but as means for meeting human needs and, ultimately, for pursuing all the diverse ends that
persons may set themselves.

While accounts of human capabilities can be quite useful in the two roles just sketched, I find
them distinctly unpromising in the role for which they are chiefly intended by their authors:
as a metric of advantage that governs the compensatory fine-tuning of the distribution of
resources so as to take account of persons’ vertically diverse capacities to convert resources
into valuable functionings. Using a list of capabilities in this way involves grading all
citizens for their natural aptitudes toward each of the capabilities on the list, determining
their specific deficits, and ensuring that these deficits are duly neutralized through suitable
compensatory benefits. Even with a large body of rules and a large administrative
bureaucracy, this task could not be accomplished in a way that is even approximately
equitable.69

Where Nussbaum boldly faces the challenge of constructing an account of valuable

capabilities, Sen skips the first two tasks and then responds to the third with a lesson about
the virtues of partial orderings: Even if we cannot agree on the weight to be attached to each
valuable capability (or natural endowment), we may be able to agree on a certain range
within which these weights should lie.70 Such a vague agreement would not suffice for
deciding, with respect to each pair of participants in some institutional order, how they
should be ranked in terms of overall capabilities or overall endowments, but it may still
suffice for the ranking of some such pairs or even many.

While this response is true, it is also unhelpful. We are seeking a public criterion of social
justice that tells us how an institutional order ought to be designed, and also how existing

67
Cf. Sanjay Reddy and Thomas Pogge: “How Not to Count the Poor” in Sudhir Anand, Paul Segal and
Joseph Stiglitz, eds.: Debates in the Measurement of Global Poverty (Oxford: Oxford University Press 2010),
also at [Link]/sol3/[Link]?abstract_id=893159.
68
Cf. text at note 12.
69
Dworkin seems even more skeptical when he writes: “The idea that people should be equal in their
capacities to achieve these desirable states of affairs, however, is barely coherent and certainly bizarre — why
would that be good? — and the idea that government should take steps to bring about that equality — can you
imagine what steps those would be? — is frightening” (Dworkin: Sovereign Virtue (Cambridge, MA: Harvard
University Press 2000), 302).
70
E.g. Sen: Inequality Reexamined, 46-49, and Sen: Development as Freedom, 78.

pogge 39
institutional schemes fall short and how they should be reformed. For this purpose we need
not merely a partial ordinal ranking, but a complete interval ranking. We need to know what
positive or negative resource compensation each participant should be entitled to on the basis
of his or her specific natural endowments. As an institutional order is fully specific, so is the
public criterion of justice underlying it. Of course, Sen may reasonably believe that there is a
plurality of admissible public criteria of social justice exemplifying the capability approach.
But, for all Sen has published on this topic, he has done little toward ruling out any
candidates within the vast space of conceivable capability views. So far, what he has mainly
proposed is a new language. This language indicates that justice requires compensation for
the naturally disfavored in proportion to how disfavored they are relative to some set of
valuable capabilities. And it also suggests that such compensation is to be levied on the
naturally favored in proportion to how favored they are relative to the same set of valuable
capabilities. But without an account of all these capabilities and their relative weights, we
cannot say of anyone whether she is naturally favored or disfavored, let alone how much she
should owe or be owed.71

You may think that this is an overstatement. Surely it is clear that people with severe
disabilities do not belong among the naturally favored. But do we really know even this
much? Many persons commonly categorized as severely disabled have gifts and talents that
militate against the judgment that they are naturally disfavored. Stephen Hawking, for
example, is seriously handicapped by a motor neuron disease. But is it clear that he is worse
endowed than most others, that justice requires us to tax able-bodied laborers and clerks in
order to help him come as close as possible to their level of physical mobility?

You may think that Stephen Hawking is a rare and special case, that most persons with
physical conditions like his cannot meet their own special needs without receiving
compensatory help. But where this is true, I would respond, it is often true on account of the
social injustices highlighted by intrinsic and extrinsic discrimination complaints: Institutional
schemes treat their diverse human participants quite unequally and, in particular, exclude
some of them from important public facilities — from the spheres of politics, law, health

71
Sen occasionally suggests that these are matters for public discussion (e.g. Sen: Development as Freedom,
110). Indeed they are: A public criterion of social justice should not be dictated by one man, however
eminent, but should be adopted democratically after broad and vigorous debate. But this point in no way
disqualifies Sen from participating in such discussion. He has argued forcefully against resourcist and
welfarist criteria of social justice and in favor of the capability approach. So why should we not be allowed to
learn which capability view is being proposed for discussion and possible adoption by Sen, who has thought
so very long and hard about this question?

pogge 40
care, education, employment, consumption, travel, culture, sports, or entertainment, for
example. Correcting such unjust exclusionary practices will often bring the kind of progress
that capability theorists are demanding.72

Still, the resourcist approach remains vulnerable on two counts. There are, first, a small
number of adults who, even in the most inclusive and otherwise favorable social
environment, cannot possibly meet their own most basic needs.73 On a resourcist view, they
have a justice claim to resources equivalent to what others have (in terms of education and
vaccinations, perhaps), but no justice claim to additional compensatory resources. In a
moderately affluent society, it is certainly morally imperative that such persons be fed,
clothed, bathed, sheltered, and cared for in a dignified way. And it would be no more unjust
to appropriate public funds for such a purpose than it would be unjust to appropriate public
funds for any other morally compelling end. Still, resourcists cannot concede what capability
theorists will insist on: that such persons have a justice claim to compensatory resources.
They can speak of duties of humanity or solidarity instead. Even if these duties are quite
stringent, they are also imperfect in two respects: They do not correlate with rights, and they
do not prescribe that society make every feasible effort toward compensating natural
disabilities. Capability theorists, by contrast, face a serious difficulty in limiting the social
cost of the compensatory justice claims they postulate. Whether they endorse an equalitarian,
prioritarian, or sufficientarian criterion, such theorists must conclude that, so long as
additional expenditures and accommodations can still help raise the lowest level of capability
achievement (below the threshold, if the view is sufficientarian), such expenditures and
accommodations are required by justice. On their account, justice commands indefinite
increases in expenditures on those with the greatest capability shortfalls, provided only that
such additional expenditures can still meaningfully improve the capabilities of at least one
such person.74

There are, second, extremely harsh natural environments where a man with a higher
metabolic rate cannot meet his extra food needs simply by moderating his discretionary

72
Such as the free provision of guide dogs discussed near the end of section 1.
73
Cf. Eva Feder Kittay: Love's Labor: Essays on Women, Equality, and Dependency (New York: Routledge
1999), discussing the condition of her daughter Sesha.
74
But note that this is true only so long as the debate is restricted to criteria of social justice whose
interpersonal aggregation function is equalitarian or prioritarian or sufficientarian or some hybrid of any two
of these or of all three. Other aggregation functions might achieve a more plausible tradeoff between the
interests of persons whose capabilities are very low and very expensive to raise and the interests of the other
participants.

pogge 41
 spending a bit or by working a little overtime. In such a context, decent people will make
every effort to ensure that the man will have enough to eat. They will do so as a matter of
basic human solidarity, realizing that, given his constitution, he simply cannot survive on the
fruits of his own labor. But does he have a justice claim to such support, can he demand it as
a matter of right? Listen to what such a justice claim would sound like: “I have a higher
metabolic rate than you all. As a consequence, I need 50 percent more food each day to be
equally well nourished. Six hours of labor are needed to produce this additional food. You
five therefore owe it to me as a matter of justice to work an extra hour each day along with
me to produce the extra food I need.” If this is not a plausible claim, then we should
recognize, I think, that there are moral requirements that, however stringent and categorical,
are not demands of justice. This recognition would remove both vulnerabilities of the
resourcist approach.

3. Conclusion

Can the capability approach be justified? Can it conceivably deliver at least one candidate
public criterion of social justice that would be as clear and as workable and as plausible as
the leading resourcist criterion, Rawls’s two principles? Would such a capability criterion do
better in addressing and highlighting the horrific injustices of the world in which we live?
The evidence to date suggests that the answers to these questions are no. That the capability
approach has nonetheless done much to advance the discussion of social justice is a great
tribute to its foremost champions: Martha Nussbaum and Amartya Sen.

pogge 42
View publication stats