Amartya Sen - Inequality Reexamined (Russell Sage Foundation Books) (1992).pdf

This page intentionally left blank

INEQUALITY
REEXAMINED
AMARTYA SEN
RUSSELL SAGE FOUNDATION • NEW YORK
CLARENDON PRESS • OXFORD

This book has been printed digitally a"d produced in a standard specification
in order to ensure its continuing availability
OXPORD
UNIVERSITY PRESS
Great Clarendon Street, Oxford OX2 6DP
Oxford University Press is a department of the University of Oxford.
It furthers the University's objective of excellence in research, scholarship,
and education by publishingworldwide in
Oxford New York
Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan South Korea Poland Portugal
Singapore Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trade mark of Oxford UniversityPress
in the UKand in certain other countries
Published in the United States
by Oxford UniversityPress Inc., New York
© Amartya Sen 1992
The moral rights of the author have been asserted
Database right Oxford University Press(maker)
Reprinted 2006
All rights reserved. Nopart of this publication maybe reproduced,
stored in a retrieval system, or transmitted, in any form or by any means,
without the prior permission in writing of Oxford UniversityPress,
op>r as expressly permitted by law, or under terms agreed with the appropriate
reprographics rights organization. Enquiries concerning reproduction
outside the scope of the above should be sent to the Rights Department,
Oxford University Press, at the address above
You must not circulate this book in any other binding or cover
And you must impose this same condition on any acquirer
fSBN 0-19-828334-2

Contents
Preface ix
Introduction; Questions and Themes 1
1. EQUALITY OF WHAT? 12
1.1 Why Equality? What Equality? 12
L2 Impartiality and Equality 16
1.3 Human Diversity and Basal Equality 19
1.4 Equality versus Liberty? 21
1.5 Plurality and Alleged Emptiness 23
1.6 Means and Freedoms 26
/.7 Income Distribution, Weil-Being and Freedom 28
2. FREEDOM, ACHIEVEMENT AND RESOURCES 31
2.1 Freedom and Choice 31
2.2 Real Income, Opportunities and Selection 34
2.3 Freedom Distinguished from Resources 36
3. FUNCnONINGS AND CAPABILITY 39
3.1 Capability Sets 39
3.2 Value Objects and Evaluative Spaces 42
3J Selection and Weighting 44
3.4 Incompleteness: Fundamental and Pragmatic 46
3.5 Capability or Functionings? 49
3.6 Utility vis-k-v"& Capability 53
4. FREEDOM, AGENCY AND WELL-BEING 56
4.1 Well-Being vis-a-vis Agency 56
4.2 Agency; Instrumentality and Realization 51
4.3 Can Freedom Conflict with Weil-Being? 59
4.4 Freedom and Disadvantageous Choices 62
4.5 Control and Effective Freedom 64
4.6 Freedom from Hunger, Malaria and Other Maladies 66
4.7 The Relevance of Weil-Being 69
5. JUSTICE AND CAPABILITY 73
5,/ The Informational Bases of Justice 73

viil Contents
5.2 Rawlsian Justice and the Political Conception 75
5.3 Primary Goods and Capabilities 79
5.4 Diversities: Ends and Personal Characteristics 85
6, WELFARE ECONOMICS AND INEQUALITY 88
6.1 Space Choice and Evaluative Purpose 88
6.2 Shortfalls, Attainments and Potentials 89
6.3 Inequality, Welfare and Justice 93
6.4 Welfare-Based Inequality Evaluation 95
7, POVERTY AND AFFLUENCE 102
7.1 Inequality and Poverty 102
7.2 The Nature of Poverty 107
7.3 Lowness vis-a-vis Inadequacy of Incomes 109
7.4 Do Concepts Matter? 112
7.5 Poverty in Rich Countries 114
8, CLASS, GENDER AND OTHER GROUPS 117
8.1 Class and Classification 117
8.2 Gender and Inequality 122
8.3 Interregional Contrasts 125
9, THE DEMANDS OF EQUALITY 129
9.1 Questions of Equality 129
9.2 Equality, Space and Diversity 130
9.3 Plurality, Incompleteness and Evaluation 131
9.4 Data, Observations and Effective Freedoms 135
9.5 Aggregation, Egalitarianism and Efficiency 136
9.6 Alternative Defences of Inequality 138
9.7 Incentives, Diversity and Egalitarianism 141
9.8 On Equality as a Social Concern 143
9.9 Responsibility and Fairness 148
9JO Capability, Freedom and Motivations 150
References 153
Index of Names 199
Index of Subjects 205

Preface
THIS monograph, as the title indicates, is about re-examining in-
equality. But it is also about the evaluation and assessment of social
arrangements in general. The former depends on the latter.
Equality of What?
The central question in the analysis and assessment of equality is, I
argue here, 'equality of whatT I also argue that a common charac-
teristic of virtually all the approaches to the ethics of social arrange-
ments that have stood the test of time is to want equality of some-
thing—something that has an important place in the particular
theory. Not only do income-egalitarians (if I may call them that)
demand equal incomes, and welfare-egalitarians ask for equal wel-
fare levels, but also classical utilitarians insist on equal weights on
the utilitiesof all, and pure libertariansdemand equalitywith respect
to an entire class of rights and liberties. They are all 'egalitarians' in
some essential way—arguing resolutely for equality of something
which everyone should have and which is quitecrucial to their own
particular approach. To seethe battle as one between those 'in favour
of* and those 'against* equality (as the problem is often posed in the
literature) is to miss something central to the subject.
I also argue that this common feature of beingegalitarian in some
significant way relates to the need to have equal concern, at some
level, for all the persons involved—the absence of which would
tend to make a proposal lack social plausibility.
Central Equality and Entailed Inequality
The crucial role of the question 'equality of what?* suggests that we
can see the disputes between different schools of thought in termsof
what they respectively take to be the central social exercise inwhich
equality is to be demanded. These demands would then qualify the
nature of the other social decisions. The demand for equality in
terms of one variable entails that the theory concerned may have to
be non-egalitarian with respect to another variable, since the two
perspectives can, quite possibly, conflict.

x Preface
For example, a libertarian demanding equal rights over a class of
entitlements cannot, consistently with that, also insist on equality
of incomes. Or, a utilitarian demanding equal weight on every unit
of utility cannot, consistently with that, also require equality of
freedoms or rights (and, for that matter, cannot insist even on
equating total levels of utilitiesenjoyed by different persons). Want-
ing equality in what is taken to be the 'central' social exercise goes
with accepting inequalityin the remoter 'peripheries'. The disputes
reside ultimately in locating the central social arrangement.
Invariant Demands and Contingent Characteristics
Indeed, the answers that are given to the question 'equality of
what?* can serve as the basis of classifying different ethical
theories of social arrangements. This classiflcatory principle
brings out in each case what the invariant properties are and what
are merely conditional or incidental connections. For example, a
libertarian who sees the central exercise as including the require-
ment that a class of individual liberties be shared equally by all,
need not object-—as a libertarian—to the equality of incomes if,
because of particular circumstances, this were also to come about.
But if the circumstances were different, it is the equality of
liberties that would be preserved, not the circumstantial equality
of incomes.
In this connection, Willard Quine has recently suggested to me
that I should explore the comparison between (1) this classificatory
principle for the ethics of social arrangements based on the
equalities that are preserved (when factual relations are trans-
formed), and (2) the classiflcatory principles used in Felix Klein's
attempted synthesisof geometry (in his Erlanger Programm) in terms
of the properties of a space that are invariant with respect to a given
group of transformations. I think there is an important general con-
nection here, which can prove to be quite illuminating, though I
have not probed this relationship in the present monograph.
Human Diversity and Disparate Equalities
At the practical level, the importance of the question "equality of
whatT derives from the actual diversity of human beings, so that
demanding equality in terms of one variable tends to clash—in fact

Preface xi
and not just in theory—with wanting equality in terms of another.
We are deeply diverse in our internal characteristics (such as age,
gender, general abilities, particular talents, proneness to illnesses,
and so on) as well as in externalcircumstances (such as ownershipof
assets, social backgrounds, environmental predicaments, and so
on), It is precisely because of such diversity that the insistence on
egalitarianism in one field requires the rejection of egalitarianism in
another,
The substantive importance of the question 'equality of what?'
relates, thus, to the empirical fact of pervasive human diversity.
Investigations of equality—theoretical as well as practical—that
proceed with the assumption of antecedent uniformity (including
the presumption that 'all men are created equal*) thus miss out on a
major aspect of the problem. Human diversityis no secondary com-
plication (to be ignored, or to be introduced later on'); it is a fun-
damental aspect of our interest in equality.
Focusing on Freedoms awrf Capabilities
The monograph begins with these arguments, the reasons for
presenting them, and the general implications they have (these issues
are taken up in Chapter 1). The rest of the book follows on this line
of analysis, and I move gradually from discussing the general
nature of equality to exploring one particular way of answering the
question 'equality of what?*
The chosen approach concentrates on our capability to achieve
valuable functioninp that make up our lives, and more generally,
our freedom to promote objectives we have reasons to value. (In-
deed, at one stage the monograph was even called 'Equality and
Freedom'.) I distinguish this approach from other ways of
answering the central question, discussing a whole class of theories
varying from utilitarianism and libertarianism to the Rawbian
theory of 'justice as fairness*. Indeed, my greatest intellectual debt
is undoubtedly to John Rawls. I am led by his reasoning over
quite a bit of the territory, and even when I go in a different direc-
tion (e.g. focusingmore on the extents of freedoms, rather than on
the means—what Rawls calls the 'primary goods"), that decision is,
to a considerable extent, based on an explicit critique of Rawls's
theory.

xii Preface
Methodological and Substantive Claims
So the monograph both develops a general methodological
approach to dealing with issues of inequality and explores a particu-
lar substantive approach as to how social arrangements may be
assessed. In the introductory section of the book ('Introduction;
Questions and Themes'), I have tried to gather together the main
lines of discussion presented in this monograph.
Kuznets Lectures and Other Connections
This monograph draws on the Simon Kuznets Memorial Lectures I
gave at Yale University in April 1988. I am very grateful to the
Economic Growth Center and to its Director, Paul Schultz, for their
invitation, for their hospitality, and for the intellectualstimulation
offered to me. Much of what we know of the natureof the economic
world has been deeply influenced by Simon Kuznets's works, and it
was a great privilegeto be able to pay tributeto hismemory through
these lectures,
There are other connections as well. Various parts of the mono-
graph draw on other lectures, on different but related themes, given
at the Delhi School of Economics (1986), the University of Texas
(1986), Cambridge University (Marshall Lectures 1988), the
University of Pittsburgh (Marion O'Kellie McKay Lecture 1988),
and the Center for Operations Research and Econometrics at
Louvain (1989). I have also lectured on several related themes to the
Royal Economic Society (Annual Lecture 1988), to the International
Economic Association (Presidential Address 1989), and to the In-
dian Economic Association (Presidential Address 1989). I learned a
good deal from the comments and criticisms that were offered in the
discussions that took place on those occasions.
Acknowledgements
I am also most grateful to the Russell Sage Foundation for research
support over a part of the summer of 1988, and to Dr Eric Wanner,
the President of the Foundation, for his interest in, and encourage-
ment of, this research. I have also received much good advice from
Andrew Schuller of the Clarendon Press and from Aida Donald of

Preface xiii
the Harvard University Press, and to them too I am much indebted,
Some of the work was done when I was visiting the London School
of Economics, and I am particularly appreciative of the splendid
working atmosphere at the STICERD research centre there-(under
the Directorship of Nicholas Stern and efficient management of
Luba Mumford).
Earlier versions of this monograph were read by A. B. Atkinson,
Susan Brison, Jean Dreze, James Foster, Siddiq Osmani, Derek
Parfit, Douglas Rae, Gustav Ranis, John Rawls, Emma Roths-
child, Paul Schultz, and Bernard Williams, and I am most deeply
indebted to them for the numerous helpful suggestions they made,
Parts of the monograph were also read by Wilfred Beckerman, Jos
de Beus, Moshe Halbertal, Steven Hawes, Athar Hussain, Robert
Keohane, Peter Lanjouw, Stephen Marglin, James Mirrlees,
Martha Nussbaum, Sanjay Reddy, and Thomas Schelling, and their
comments have been extremely useful. Over the years I have also
had the opportunity of profiting from discussions with a number of
people on these and related subjects, and I would like to acknowl-
edge my debt to Sudhir Anand, Kenneth Arrow, Pranab Bardhan,
Kaushik Basu, Peter Bauer, Andre Beteiite, Charles Blackorby,
Christopher Bliss, John Broome, James Buchanan, G. A, Cohen,
Douglas Dacy, Ralf Dahrendorf, Partha Dasgupta, Claude
d'Aspremont, Angus Deaton, Meghnad Desai, Bhaskar Dutta,
Ronald Dworkin, Wulf Gaertner, Louis Gevers, Jonathan Glover,
James Griffin, Keith Griffin, Peter Hammond, Mahbub ul Haq,
Richard Hare, Albert Hirschman, Eric Hobsbawm, Lai Jayaw-
ardena, Kumari Jayawardena, Ravi Kanbur, Nanak Kakwani,
John Knight, Richard Layard, Isaac Levi, John Mackie, Mukul
Majumdar, John Muellbauer, Amulya Ratna Nanda, Robert
Nozick, Prasanta Pattanaik, Hilary Putnam, Ruth Anna Putnam,
Willard Quine, V, K. Ramachandran, Martin Ravallion, Ashok
Rudra, Thomas Scanlon, A. F. Shorrocks, T, N. Srintvasan, David
Starrett, Hillel Steiner, Frances Stewart, Paul Streeten, Kotaro
Suzuinura, Larry Temkin, Philippe Van Parijs, Patricia Williams,
Bengt Christer Ysander, Stefano Zamagni, and Vera Zanaagni.
I have benefited from valuable research assistance from Chitrita
Banerji, Stephan KJasen, and Sanjay Reddy, I am also grateful for
the organizational support I have received from Jacky Jennings and
Anna Marie Svedrofsky.

xiv Preface
Some Presentational Points
I should, finally, make a few remarks about the presentational
aspects. First, the minor issue of gender and language, 1 refer here
not to the substantive concerns of this monograph, which include
questions of gender inequalityin various forms, but to the appear-
ance of inconsistency in my use of gender-specific pronouns—
sometimes usingthe 'he*, sometimes 'she', sometimes 'he or she*. The
exclusive use of 'he* to refer to both women and men is open to the
charge of some implicit sexism; the exclusive use of'she* can sound
somewhat self-conscious and precious (and also open to a similar
charge of discrimination from the opposite direction); and using'he
or she' everywhere every time is verbose, cumbrous, and ugly. If no
great symbolicsignificance isto beattached to the gender form of the
pronoun, then the natural route to take is to use the different forms
interchangeably, and to refuse to seeany inconsistency in this. This is
exactly what I have tried to do.
Second, since I have been anxious to reach a wider audience than
that of formally trained economists, I have tried to use as few tech-
nical concepts and mathematical expressions as possible. This has
sometimes left room for further specification, but I have also refer-
red to other writings (includingmy own) in which those issues have
been further discussed.
Third, I have been told that the long list of references included in
the monograph may 'raise some eyebrows*. But the literature is
varied and vast, and having had the benefit of profiting from it, it
would be wrongnot to acknowledge the presence of this vast bodyof
writings. The monograph is not a synthetic contribution, and I do
want to take the discussion of inequality in a somewhat different
direction—away from the prevailing traditions. But to do that 1have
to establish the co-ordinates of past and ongoing work, if only to
ascertain where a departure might be due. Many of the works cited
are also disputed, but even whenthey are, I have often been left with
residual gain from the dialectic encounter. The bibliography may
also, I hope, be of some useto readers-—different parts, I imagine, to
different readers, since they cover distinct aspects of inequality
analysis. Be that as it may, I am working on steeling myself against
elevated eyebrows.
A.K.S.

Introduction
Questions and Themes
The idea of equality is confronted by two different types of
diversities." (1) the basic heterogeneity of human beings, and (2) the
multiplicity of variables in terms of which equality can be judged.
This book is concerned with both these diversities. It is also specifi-
cally concerned with the relation between the two. The hetero-
geneity of people leads to divergences in the assessment ofequality
in terms of different variables. This adds significance to the central
question: equality of what?
Diverse Humanity
Human beings are thoroughly diverse. We differ from each other
not only in external characteristics (e.g. in inherited fortunes, in the
natural and social environment in which we live), but also in our
personal characteristics (e.g. age, sex, proneness to illness, physi-
cal and mental abilities). The assessment of the claims of equality
has to come to terms with the existence of pervasive human di-
versity,
The powerful rhetoric of 'equality of man* often tends to deflect
attention from these differences. Even though such rhetoric (e.g.
"all men are born equal*) is typically taken to be part and parcel of
egalitarianism, the effect of ignoring the interpersonal variations
can, in fact, be deeply inegalitarian, in hiding the fact that equal
consideration for all may demand very unequal treatment in
favour of the disadvantaged. The demands of substantiveequality
can be particularly exacting and complex when there is a good
deal of antecedent inequality to counter.
Sometimes, human diversities are left out of account not on the
misconceived 'high* ground of 'equality of human beings", but on
the pragmatic low* ground of the need for simplification. But the
net result of this can also be to ignore centrally important features
of demands of equality.

2 Introduction: Questions and Themes
Diversity of Foots
Equality isjudged by comparing some particular aspect of a person
(such as income, or wealth, or happiness, or liberty, or oppor-
tunities, or rights, or need-fulfilments) with the same aspect of
another person. Thus, thejudgement and measurement of inequality
is thoroughly dependent on the choice of the variable (income,
wealth, happiness, etc.) in terms of which comparisons are made, I
shall call it the 'focal variable'—the variable on which the analysis
focuses, in comparing different people,
The chosen focal variable can, of course, have an internal plur-
ality. For example, freedoms of different types may be put together
as the preferred focus of attention, or the variable selected may
involve a combination of freedoms and achievements. The multiple
features withina chosen focal variable have to be distinguished from
the diversity betweenthe chosen focal variables. Some variables that
are often taken to be elementary and uniform do, in fact, have much
internal plurality (e.g. real income or happiness).'
To use the kind of language for which we economists are often—
not unreasonably—teased, this is the question of the choice of
'space' in which different persons are to be compared. That spatial
analogy, despite its demonstratively Cartesian pretensions, is a
useful classiflcatory device, and I shall invoke it to separate out the
problem of the choice of focal variables ('the choice of space') from
other issues in the assessment of inequality,
Links and Disharmonies
The characteristics of inequality in different spaces (such as income,
wealth, happiness, etc,) tend to diverge from each other, because of
the heterogeneity of people. Equality in terms of one variable may
not coincide with equality in the scale of another. For example,
equal opportunities can lead to very unequal incomes. Equal in-
comes can go with significant differences in wealth. Equal wealth
can coexist with very unequal happiness. Equal happiness can go
with widely divergent fulfilment of needs. Equal fulfilment of needs
can be associated with very different freedoms of choice. And so on.
If every person were much the same as every other, a major cause
1
I have discussed these issues elsewhere, addressing also the problem of overall
ranking and aggregate valuation of inherently plural variables (in Sen 1980-1,
!982a).

Introduction: Questions and Themes 3
of these disharmonies would disappear. If the rankinp of equality in
different spaces coincide, it would then be less important to have a
clear answer to the question: equality of what? The pervasive diver-
sity of human beings intensifies the need to address the diversity of
focus in the assessment of equality,
Diverse EgaUtarianism
It is convenient to begin with the observation that the major ethical
theories of social arrangement all share an endorsement of equality
in terms of some focal variable, even though the variables that are
selected are frequently very different between one theory and
another. It can be shown that even those theories that are widely
taken to be 'against equality* (and are often described as such by the
authors themselves) turn out to be egalitarian in terms of some other
focus. The rejection of equality in such a theory in terms of some
focal variable goes hand in hand with the endorsement of equality in
terms of another focus.
For example, a libertarian approach (such as the entitlement
theory forcefully developed in Robert Noziek's Anarchy, State and
Utopia2
) may give priority to extensive libertiesto be equally guaran-
teed to each, and this demands rejecting equality—or any 'pat-
terning*—of end states (e.g. the distribution of incomes or happi-
ness). What is taken—usually by implication—to be a more central
focus rules the roost, and inequalities in the variables that are, in
effect, treated as peripheral must, then, be accepted in order not to
violate the right arrangements (including equality) at the more cen-
tral level.
Plausibility and Equality
There is a reason for this apparently ubiquitous 'egalitarianism'.
Ethical plausibilityis hard to achieve unless everyone is given equal
consideration in some space that is important in the particular
theory (Chapter 1). While it may be too ambitious to claim (as some
have done) that this is a logical necessity, or simply a part of the
discipline of the language of morals,3
it is difficult to see how an
2
See Noaek (1973, 1974). For a reassessment and some revision, see Nozick
(1989).
3
See particularly Hare (1952, 1963).

ethical theory can have general social plausibility without
extending equal consideration to all at some level.
While the question 'why equality?' is by no meansdismissible, it is
not the central issue that differentiates the standard theories, since
they are all egalitarian in terms of some focal variable. The
engaging question turns out to be 'equality of what?'
To that question—'equality of what?*—-different theories give
different answers. The different answers are distinguishable in prin-
ciple and involve different conceptual approaches. But the practical
force of these distinctions depends on the empirical importance of
the relevant human heterogeneities which make equality in one
space diverge from equality in another.
Achievement and Freedom
Sources of divergence between different approaches can, of course,
go well beyond the identification of the space itself, and may be
concerned instead with the way the space is utilized. In the standard
theory of inequality measurement, these problems of "appropriate
indices* have tended to receive much attention. The analysis can
fruitfully proceed on the basis of postulating—explicitly or im-
plicitly—acceptable axioms for inequality assessment in that space.
While the focus of this book is on the choice of space and its impli-
cations, it is not my intention to deny the practical importance of
these indexing problems in a given space (this was in fact the main
subject of analysis in my previous book on inequality4
).
One of the aspects of inequality assessment that has received less
attention than it deserves relates to the distinction between achieve-
ment and the freedom to achieve. The nature, reach, and relevance of
that distinction between achievement and freedom are briefly dis-
cussed in Chapter 2, makinguse of conceptual grounds for discrimi-
nation as well as analytical procedures used in modern economics.
Functionings and Capability
The monograph then proceeds to identify, develop, and defend a
particular choice of space and its use in terms of the freedom to
achieve (Chapter 3). A person's capability to achieve functionings
4
On Economic Inequality; Sen 1973« in bibliography. Since I shall have to cite it
frequently (mainly to avoid having to repeat myself), I shall refer to it in a more
easily recognizable form, viz, OBI.

Introduction; Questions and Themes 5
that heor she has reason to value providesa general approach to the
evaluation of social arrangements, and this yields a particular way
of viewing the assessment of equality aad inequality,
The functionings included can vary from most elementary ones,
such as being well-nourished, avoiding escapable morbidity and
premature mortality, etc,, to quite complex and sophisticated
achievements, such as having self-respect, being able to take part in
the life of the community,and so on. The selection and weighting of
different functionings influence the assessment of the capability to
achieve various alternative functioning bundles.
The roots of this approach can be traced to Aristotelian distinc-
tions, but its ramifications can take various different forms. The
particular class of possibilities developed here is less assertive and
less insistently complete than some possible alternatives. But it is
also less demanding on interpersonal agreement and more tolerant
of unresolved disputes.
Evaluation of Effective Freedom
The concentration on the freedom to achieve and not just on thelevel
of achievement raises some deep questions about the connection
between the appraisal of the alternative achievements and the value
of thefreedom to achieve them (Chapter 4). Even the freedom-based
perspective must pay particular attention to the nature and valueof
the actual achievements, and inequalities in achievement can throw
light on inequalities in the respective freedoms enjoyed. This recog-
nition requires us to reject such proposed rules of freedom assess-
ment as the counting of the number of alternatives in the 'range of
choice'. More constructively, it suggests practical ways of using
observable data regarding achievements to get a partial but signifi-
cant view of the freedoms enjoyed by different persons.
In this context I also discuss the difference between well-being
objectives and the other objectives a person may have. This
difference not only leads to some plurality within the idea of
freedom itself, it also has important implications on the divergence
between the perspective of achievements and that of freedoms.
One of the related issues taken up here is the possibility that more
freedom can be disadvantageous to" a person, which—if generally
true—can undermine the rationale of judging inequality in terms of
freedoms. I argue that the real conflict is between different types of

freedoms, and not between freedom tout court and advantages in
general.
Distinctions: Capability and Utility
The focus on the space of functionings—and on the capability to
achieve functionings—differs quite substantially from the more tra-
ditional approaches to equality, involving concentration on such
variables as income, wealth, or happiness (Chapters 3 and 4). The
fact of human diversity is closely related to substantive conflicts
between focusing on different informational bases for assessing
equality, efficiency, and justice.
In particular, judging equality and efficiency in terms of the
capability to achieve differs from the standard utilitarian
approaches as well as from other welfarist formulations. Welfarism
in general and utilitarianism in particular see value, ultimately,
only in individual utility, which is defined in terms of some mental
characteristic, such as pleasure, happiness, or desire,5
This is a re-
strictive approach to taking note of individual advantage in two
distinct ways; (I) it ignores freedom and concentrates only on
achievements, and (2) it ignores achievements other than those re-
flected in one of these mental metrics. In so far as utility is meant to
stand for individual well-being, it provides a rather limited
accounting of that, and it also pays no direct attention to the
freedom to pursue well-being—or any other objective (Chapter 3).
This way of seeing individual advantage is particularly limitingin
the presence of entrenched inequalities. In situations of persistent
adversity and deprivation, the victims do not go on grieving and
grumbling all the time, and may even lack the motivation to desire a
radical change of circumstances. Indeed, in terms of a strategy for
living, it may make a lot of sense to come to terms with an ineradic-
5
There issome ambiguity in the characterization of the 'preference* viewof utility,
as itcan be—and has been—defined inquitedistinct and divergentways. If it is defined
entirely in terms of individual choice (as in Samuelson 1938), then the preference view
does not give any immediate content to interpersonal comparisons and thus does not
yield any straightforward judgements of inequality. The devised meaningsthat can be
somewhat artificially constructed (e.g. Harsanyi's 1955 engaging proposal that we
consider our preference regarding who we would choose to be) involve conceptual
problems as well as empirical difficulties (I have discussed this issue in Sen 1982a),
On the other hand, if preference is defined in terms of a person's sense of desire or
satisfaction (as in Edgeworth 1881 or Hicks 1939), then the preference approach is in
line with the mental-metric views considered in the text.

able adversity, to try to appreciate small breaks, and to resist pining
for the impossible or the improbable. Such a person, even though
thoroughly deprived and confined to a very reduced life, may not
appear to be quite so badly off in terms of the mental metricof desire
and its fulfilment, and in terms of the pleasure-pain calculus, The
extent of a person's deprivation may be substantially muffled in the
utility metric, despite the fact that he or she may lack the oppor-
tunity even to be adequately nourished, decently clothed,
irinimally educated, or properly sheltered.
The misleading nature of utility metrics may be particularly
important in the context of stable differentiation of class, gender,
caste, or community, It contrasts with the focus on capabilities,
which provides a straightforward account of the lack of freedom of
the deprived people to achieve those elementary functionings
(Chapter 3).
Capability amd Opportunities: Equality and Efficiency
The capability perspective also differs from various concepts of
'equality of opportunities' which have been championed for a long
time. In a very basic sense, a person's capability to achieve does
indeed stand for the opportunity to pursue his or her objectives. But
the concept of 'equality of opportunities* is standardly used in the
policy literature in more restrictive ways, defined in terms of the
equal availability of some particular means, or with reference to
equal applicability (or equal rtOB-applicability) of some specific
barriers or constraints,
Thus characterized, 'equality of opportunities' does not amount
to anything likeequalityof overall freedoms.This is so because of (1)
the fundamental diversity of human beings, and (2)the existence and
importance of various means (such as income or wealth) that do not
fall within the purview of standardly defined 'equality of oppor-
tunities'. In terms of the position outlined and defended in this
monograph, a more adequate way of considering 'real' equality of
opportunities must be through equality of capabilities (or through
the elimination of unambiguous inequalities in capabilities, since
capability comparisons are typically incomplete).
But equalityis not the only social charge with whichwehave to be
concerned, and there are demands of efficiency as well. An attempt
to achieve equality of capabilities—without taking note of aggrega-

live considerations—can lead to severe curtailment of the capabili-
ties that people can altogether have. The demands of equality of
capabilities has to be $een in the context of the contending claims of
efficiency and, in general, of aggregative concerns. Indeed, it will be
argued that the import of the concept of equality cannot even be
adequately understood without paying simultaneous attention also
to aggregative consideration—to the 'efficiency aspect', broadly
speaking (Chapter 9),6
Differences with the Rawlsian Focus
A particularly important contrast is that between capability-based
evaluation and RawSs's (1971) procedure of focusing on the holding
of 'primary goods' (including resources such as incomes, wealth,
opportunities, the social bases of self-respect, etc.). This is a part of
his 'Difference Principle*, which is an integral component of the
Rawlsian theory of 'justice as fairness'. While my own approach is
deeply influenced by Rawls's analysis,1
1 argue that the particular
informational focus on which Rawls himself concentrates neglects
some considerations that can be of great importance to the substan-
tive assessment of equality—and also of efficiency.8
The importance of the contrast once again turns on the funda-
mental diversity of human beings. Two persons holding the same
bundle of primary goods can have very different freedoms to pursue
their respective conceptions of the good (whether or not these con-
ceptions coincide). To judge equality—or for that matter effi-
ciency—in the space of primary goods amounts to giving priority to
the meansof freedom over any assessment of the extents of freedom,
and this can be a drawback in many contexts. The practical impor-
tance of the divergence can be very great indeed in dealing with
* In several ethical frameworks, the insufficient attention paid explicitly to effi-
ciency considerations is combined with choice of somewhat insensitive indicators for
the assessment of inequality, and thus the neglect of efficiency does not yield immedi-
ately unappealing results. But this kind of "double limitation" does little justice
either to equality or to efficiency. I argue that something of this problem is present
even in the Rawlsian formulation of the Difference Principle (Ch. 9).
1
In fact, one reason for my concentration on the difference between Rawls's
analysis and what I have proposed is precisely my indebtedness to Rawls. By
specifying in tome detail {Chs. 5 and 9) the departures from Rawls's position, it is
possible to clarify what exactly is being claimed and why.
* A similar remark can be made, though for somewhat different reasons, about
Ronald Dworkin's (1981) arguments for "the equality of resources* (see Ch, 5, and
also Sen 1984: ch. 13).

inequalities related to gender, location, and class, and also to
general variations in inherited characteristics.
Economic Inequality and Poverty
The perspective of functionings and capabilities suggests particular
approaches to the evaluation of economic inequality. They differ
from the standardly used informationalfocus in welfare economics,
which tends to concentrate on incomes, wealth, and utilities
(Chapter 6), They also raise some questions about the analytical
procedures of inequality evaluation commonly used in economic
theory.
The theory of inequality evaluation has close links with that of
assessment of poverty, and the choice of space becomes a central
concern in identifying the poor and in aggregating the information
about the states of those identified. If poverty is seen as the depriva-
tion of some minimum fulfilment of elementary capabilities, it
becomes easier to understand why poverty has both an absolute
and a relative aspect. These considerations are important in
dealing with poverty in any country (rich or poor), but are particu-
larly relevant in understanding the nature of poverty in the richer
countries, such as the USA and those in Western Europe (Chapter
7), The persistence of poverty in otherwise affluent countries is an
apparently puzzling phenomenon that is beginning to get serious
attention in contemporary debates. The understanding and the
remedying of this problem can both be helped by explicit con-
sideration of the relation between deprivations in different spaces,
especially between incomes and the capability to lead secure and
worthwhile lives.
Clas$,. Gender and Other Categories
In Chapter 8 the relevance of the capability perspective is discussed
in the context of differences in class, gender, and other social
features. Once again, the fact of humandiversityplays a crucial part
in strengthening the significance of the informational departure in
moving from the spaces of incomes, opportunities, happiness, pri-
mary goods, etc., to takingdirect note of the achievement of func-
tionings and the capability to achieve them (Chapter 8).
The capability perspective is more sensitive than utility-based

10 Introduction: Questionsand Themes
approaches to problems of entrenched deprivation, which can lead
to defensiveadjustmentof desires and expectations (thereby distort-
ing the metric of utilities). It can also be fairer in dealing directly
with freedoms rather than concentrating on the means of freedoms.
These differences are significant in assessing inequality and injustice
across the barriers of class, gender, and other social divisions.
Equality, Efficiency and Incentives
While the last chapter does not 'summarize' the monograph (or list
the main conclusions), it contains a fairly wide-ranging discussion of
many of the issues covered in the book. It also tries to link the
methodological arguments on equality with the substantive
analysis of the capability perspective as the basis of judging
equality.
In this context, it is argued that the demands of equalitycannot be
properly assessed without seeing them in the context of other de-
mands, especially those of aggregative objectives and of overall
efficiency. When equality is viewed isolated from other concerns, the
evaluation of equality tends to get distorted because of the unneces-
sary load it has to carry (proxying for efficiency objectives that can
be better accommodated elsewhere). This consideration has some
bearing on the formulation of theories of justice, including
Rawlsian theory.
In considering the conflict between aggregative and distributive
objectives, the incentive problem proves to be less of a force against
egalitarianism when the inequalities are generated by entrenched
antecedent diversities (as they typically are in the case of class,
gender, and other non-adjustable and identifiable barriers). Since
the problem of inequality can be particularly serious in the presence
of extensive human diversities, this question is of some relevance for
economic and social policy.
Methods and Substance
This monograph is concerned with methodological as well as sub-
stantive issues. The attention is mostly concentrated on conceptual
clarification in the early chapters, emphasizing the importance of
the question 'equality of what?' and relating it to the fact of extensive
human diversity. The substantive claims take the form of proposing

a particular way of answering that question and suggestingvarious
implications of that answer. The implications, I have argued, are not
only of theoretical interest, they also have some practical impor-
tance.

1
EQUALITY OF WHAT?
1.1. WHY EQUALITY? WHAT EQUALITY?
Two central issues for ethical analysis of equality are; (1) Why
equality? (2) Equality of what? The two questions are distinct but
thoroughly interdependent. We cannot begin to defend or criticize
equality without knowing what on earth we are talking about, i.e.
equality of what features (e.g. incomes, wealths, opportunities,
achievements, freedoms, rights)? We cannot possibly answer the
first question without addressing the second. That seems obvious
enough.
But if we do answer question (2), do we still need to address
question (1)? If we have successfully argued in favour of equality
of* (whatever that x is—some outcome, some right, some freedom,
some respect, or some something else), then we have already
argued for equality in (hat form, with x as the standard of com-
parison. Similarly, if we have rebutted the claim to equality of x,
then we have already argued against equality in that form, with x
as the standard of comparison. There is, in this view, no 'further',
no 'deeper', question to be answered about why—or why
not—'equality'. Question (1), in this analysis, looks very much like
the poor man's question (2).
There is some sense in seeing the matter in this way, but there is
also a more interesting substantive issue here. It relates to the fact
that every normative theory of social arrangement that has at all
stood the test of time seems to demand equality of something—
something that is regarded as particularly important in that theory.
The theories involved are diverse and frequently at war with each
other, but they still seem to have that common feature. In the
contemporary disputes in political philosophy, equality does, of
course, figure prominently in the contributions of John Rawls
(equal liberty and equality in the distribution of 'primary goods'),
Ronald Dworkin ('treatment as equals', 'equality of resources'),
Thomas Nagel ('economic equality'), Thomas Scanlon ('equal-
ity'), and others generally associated with a 'pro equality'

Equality of What? 13
view.' But equality in some space seems to be demanded even by
those who are typically seen as having disputed the 'case for
equality' or for 'distributive justice*. For example, Robert Noziek
may act demand equality of utility or equality of holdings of pri-
mary goods, but he does demand equality of libertarian rights—no
one has any more right to libertythan anyoneelse, JamesBuchanan
builds equal legal and political treatment—4ndeed a great deal
more—into his view of a good society.2
In each theory, equality is
sought in some space—a space that is seen as havinga central role in
that theory.3
But what about utilitarianism? Surely, utilitarians do not, in
general, want the equality of the total utilities enjoyed by different
people. The utilitarian formula requires the maximization of the
sum-total of the utilities of all people taken together, and that is, in
an obvious sense, not particularly egalitarian.4
In fact, the equality
that utilitarianismseeks takes the form of equal treatment of human
beings in the space of gains and losses of utilities. There is an insist-
ence on equal weights on everyone's utility gains in the utilitarian
objective function.
This diagnosis of 'hidden' egalitarianism in utilitarian philos-
ophy might well be resisted on the ground that utilitarianism really
involves a sum-total maximizingapproach, and it might be thought
that, as a result, any egalitarian feature of utilitarianism cannot be
more than accidental. But thisreasoning isdeceptive. Theutilitarian
approach is undoubtedlya maximizing one, but the real question is
' See Rawls (1971,I988«), R. Dworkin (1978, 1981), Ntgel (1979,1986), Scanlou
(1982, 1988ft). The positions taken by the modem utilitarians raise a more complex
question (on which more presently), bwt the starting-point is something like 'giving
equal weightto the equal interests of all the parties"(Hare 1982:26), or a procedure to
'always assign the same weight to all individuals' interests*(Haraanyi 1982: 47),
2
SeeNozick (1973,1974), J. M. Buchanan (1975,1986). Seealso J. M, Buchanan
and Tultecfc (1962).
J
This does not, obviously, apply to those critiques of equality (in some space) that
do not include a proposal for somethingconstructive instead. It is the presentation or
defence of such a constructive proposal that can be expected to entail—often im-
plicitly—the demand for equality in some other space. Nor is the expectation of a
demand for equality in some other space likely to apply to theories that do not refer
to human beings »t all, e,g, proposals that advocate 'maximization of the total
market value of wealth*. It is in a constructive proposal making use of some human
condition that an implicit demand for some type of equality is likely to occur,
4
In my earlier book on inequality (On Economic Inequality, Sen 1973« in the
bibliography), I had discussed in some detail (see ch. 1) why utilitarianism is in-
egalitarian in some important respects. As indicated in the Introduction, that book is
referred to in this monograph as OKI.

14 Equality of What?
what Isthe nature of the objective function it maximizes. That objec-
tive function could have been quite inegalitarian, e.g. giving much
more weight to the utilities of some than to those of others. Instead,
utilitarianism attaches exactly the same importance to the utilities
of all people in the objective function, and that feature—coupled
with the maximizing format—guarantees that everyone's utility
gains getthe same weight in the maximizingexercise.The egalitarian
foundation is, thus, quite central to the entire utilitarian exercise,
Indeed, it is precisely this egalitarian feature that relates to the
foundational principle of utilitarianism of 'giving equal weight to
the equal interests of all the parties* (Hare 1981: 26), or to 'always
assign the same weight to all individuals' interests' (Harsanyi 1982:
47).5
What do we conclude from this fact? One obvious conclusion is
that being egalitarian (i.e. egalitarian in somespace orotherto which
great importance is attached) is not really a 'uniting' feature.6
In-
deed, it is precisely because there are such substantive differences
between the endorsement of different spaces in which equality is
recommended by different authors that the basic similaritybetween
them (in the form of wanting equality in some space that is seen as
important) can be far from transparent. This is especially so when
the term 'equality1
is defined—typically implicitly—as equality in a
particular space,
For example, in his interesting essay, The Case against Equality',
with which William Letwin (1983) introduces an important collec-
tion of papers by different authors on that theme (the volume is
called Against Equality), he argues against equal distribution of
incomes (or commodities) thus: 'Inasmuch as people are unequal, it
5
John Rawis (197!) has argued that 'there is a sense in which classical utilitar-
ianism fails to take seriously the distinction between persons* (p. 187). In so far as a
utilitarian theorist argues simply for the maximization of the amount of happiness,
pleasure, etc., with no attention beingpaid to the fact that these thingsare featuresof
particular persons, Rawls's claim has much force. But a utilitarian can also see
utility as an irredueibly personal feature demanding attention precisely because the
well-beings of the persons involved command respect and regard. On this see
Bentham(1789), Mill (1861), Bdgeworth (1881), Pigou (1952), Hare (1981), Harsanyi
(1982), and Mirrlees (1982). This limited 'defence' of utilitarianism should not be
seen as supporting it as an adequate ethical or political theory. Utilitarianism does
have serious deficiencies (I have tried to discuss them elsewhere: Sen 1970a, 1979*,
1982&), but not takingthe distinction between diflferent persons seriously may not be a
fair charge against utilitarianism in general,
6
On this and related issues, see B. Williams(I973a), Suppes (1977), Sen (I980a),
R. Dworkin (1981), Rae (1981), Beteille (19836).

is rational to presume that they ought to be treated unequally—
which might mean larger shares for the needy or larger shares for the
worthy' ('A Theoretical Weakness of Egalitarianism*, 8). But even
the demand for equal satisfaction of 'needs' is a requirement of
equality (in a particular space), and it has indeed been championed
as such for a long time. Even though the idea of individual 'worth'is
harder to characterize, the usual formulations of the demand for
larger shares for the worthy' tend to include equal treatment for
equal worth, givingto each the same reward for worth as is given to
another. Thus, these critiques of egalitarianism tend to take the form
of being—instead—egalitarian in some other space.7
The problem
again reduces to arguing, implicitly, for a different answer to the
question 'equality of what?'.
Sometimes the question 'equality of what?' gets indirectly
addressed in apparently discussing 'why equality?', with equality
defined in a specific space. For example, Harry Frankfurt's (1987)
well-reasoned paper attacking 'equality as a moral ideal* is con-
cerned mainly with disputing the claims of economic egalitarianism
in the form of 'the doctrine that it is desirable for everyone to have
the same amounts of income and wealth (for short, "money")'
(p. 21).8
Though the language of the presentation puts 'egalitarian-
ism* as such in the dock, this is primarilybecause Frankfurt uses that
general term to refer specifically to a particular version of 'econ-
omic egalitarianism': 'This version of economic egalitarianism (for
short, simply "egalitarianism") might also be formulated as the
doctrine that there should be no inequalities in the distributionof
money' (p.21).
The choice of space for equality is, thus, central to Frankfurt's
main thesis,9
His arguments can be seen as disputing the specific
demand for a common interpretation of economic egalitarianism by
1
Similarly,Peter Bauer's (1981) forceful argument in favour of the same right tor-
all to enjoy what they have 'produced* is also an egalitarian demand to that chosen
space—that of receiving a reward commensurate with one's productive contribution,
8
See also J.R.Lucas (1965,1980). For a pointedcritique of Frankfort's thesis, see
Ooodin (1987).
* Indeed, the nature of the space is crucial to all axioms that take the form of
demanding or rejecting equality. For example, the 'weak equity axiom' stated in my
OE1indicated a preference for equality in the space or overall well-being. While that
condition was possibly over-strong, since it incorporated a lexicographic priority of
equality over aggregative considerations, some of the critiques of the condition have
been misplaced in interpreting the formal requirement in other spaces, e.g. in the
allocation of specialist medical care among persons (see J. Griffin 1981, 1986; see
also Brandt 1979, and my response to his critique in Sen 1980-1).

arguing (1) that such an equalityis of no great intrinsic interest, and
(2) that it leads to the violation of intrinsically important values-
values that link closely to the need for payingequal attention to all
in some other—more relevant—way.
Wanting equalityat something—something seen as important—is
undoubtedly a similarityof some kind, but that similaritydoes not
put the warringcamps on the same side. It only shows that the battle
is not, in an important sense, about 'why equality?", but about
'equality of what?*.
Since some spaces are traditionally associated with claims of
'equality' in political or social or economic philosophy, it is
equality in one of those spaces (e.g. incomes, wealths, utilities) that
tend to go under the heading 'egalitarianism*, I am not arguing
against the continued use of the term 'egalitarianism' in one of
those senses; there isno harm in that practice ifit isunderstood to be
a claim about equality in a specific space (and by implication,
against equality in other spaces), But it is important to recognize
the limited reach of that usage, and also the fact that demanding
equality in one space—no matter how hallowed by tradition—can
lead one to be anti-egalitarian in some other space, the comparative
importance of which in the overall assessment has to be critically
assessed.
1,2, IMPARTIALITY AND EQUALITY
The analysis in the last section pointed to the partisan character of
the usual interpretations of the question 'why equality?'. That ques-
tion, I haveargued, has to be faced, just as much, even by those who
are seen—by themselves and by others—-as 'anti-egalitarian', for
they too are egalitarian in some space that is important in their
theory. But it was not, of course, argued that the question t*why
equality?' was, in any sense, pointless. Wemay bepersuaded that the
basic disputations are likely to be about 'equality of what?', but it
might still be asked whether there need be a demand for equality in
some important space or other. Even if it turns out that every sub-
stantive theory of social arrangements in vogue is, in fact,
egalitarian in some space—a space seen as central in that theory—
there is still the need to explain and defend that general characteris-
tic in each case. The shared practice—even if it were universally
shared—would still need some defence.

The issue to address is not so much whether there must be for
strictly formal reasons (such as the discipline of 'the language of
morals'), equal consideration for all, at some level, in all ethical
theories of social arrangement.l
° That is an interesting and hard
question, but one I need not address in the present context; the
answer to it is, in my judgement, by no means clear, I am more
concerned with the question whether ethical theories must have this
basic feature of equality to have substantive plausibility in the
world in which we live,
It may be useful to ask why it is that so many altogether different
substantive theories of the ethics of social arrangements have the
common feature of demanding equality of something—something
important. It is, I believe, arguable that to have any kind of plau-
sibility, ethical reasoning on social matters must involve ele-
mentary equal consideration for all at some level that is seen as
critical. The absence of such equality would make a theory arbi-
trarily discriminating and hard to defend. A theory may accept-
indeed demand—inequality in terms of many variables, but in de-
fending those inequalities it would be hard to duck the need to
relate them, ultimately, to equal consideration for all in some
adequately substantial way.
Perhaps this feature relates to the requirement that ethical
reasoning, especially about social arrangements, has to be, in
some sense, credible from the viewpoint of others—potentially all
others. The question 'why this system?* has to be answered, as
it were, for all the participants in that system. There are
some Kantian elements in this line of reasoning, even though
the equality demanded need not have a strictly Kantian struc-
ture."
Recently Thomas Scanlon (1982) has analysed the relevance
and power of the requirement that one should *be able to justify
one's actions to others on grounds that they couid not reasonably
10
For a classic exposition and defence of such an analytically ambitious claim, see
Hare (1952, 1963).
ij por reasons for taking note of differences (e.g. of personal commitments or obli-
gations) that tend to be ignored, at least in some versions of the Kantian uniformist
format, see Williams (1981), Hampshire (1982), Taylor (1982). On some related
issues, see Williams (1973a), where it is also discussed why 'the various elements of.
the idea of equality' pull us in 'different directions* (p.24S). But the acknowledge-
ment of the importance of different obligations and commitments does not, of course,
do away with the general need to make our ethics credible to others.

reject',n
The requirement of fairness' on which Rawls (1971)builds
his theory ofjustice can be seen as providing a specific structure for
determining what one can or cannot reasonably reject," Similarly,
the demands of 'impartiality*—and some substantively exacting
forms of "universalizability'—invoked as general requirementshave
that feature of equal concern in some major way.M
Reasoning ofthis
general type certainlyhas much to do with the foundationsof ethics,
and has cropped up in different forms in the methodological under-
pinning of substantive ethical proposals.K
The need to defend one's theories, judgements, and claims to
others who may be—directly or indirectly—involved, makes
equality of consideration at some level a hard requirement to avoid.
There are interesting methodological questions regarding the status
of thiscondition, in particular: whetherit is a logical requirement or
a substantive demand,"and whetherit isconnected with the need for
"objectivity' in ethics.17
I shall not pursue these questions further
here, since the main concerns of this monograph do not turn on our
answers to these questions.l8
What is of direct interest is the plausibilityof claiming that equal
consideration at some level—a level that is seen as important—is a
demand that cannot be easily escaped in presenting a political or
ethical theory of social arrangements. It is also of considerable
pragmatic interest to note that impartiality and equal concern, in
12
See also Scanlon (19880). On related matters, see Rawls (1971, 1988c), B.
Williams (S972, 1985), Mackie (I978o), Ackerman (1980, 1988), Parflt (1984),
O'Neill (1989).
11
See also Rawls's later—more explicit—analysis of this connection, in Rawls
(1985, 19880, 1990).
14
See Mackie (1978a). Impartiality-based reasoning is used by Harsanyi (1955)
and Hare (1963) to defend the choice of utilitarian ethics. The idea of equal concern,
in the form of the requirement of impartiality, is invoked even in setting up theories
that explicitly take an 'anti-egalitarian' form. For example, in presenting his ease for
'morals by agreement', Gauthier (1,986) asserts—correctly in terms of his particular
definition of equality—that 'equality is not a fundamental concern in our theory*,
but goes on immediatelyto explain: "we have appealed to the equal rationality of the
bargainers to show that their agreement satisfies the moral standard of impartiality'
(p.270, emphasis added).
15
On this see Sen (197Qa: eh. 9).
'* This issue can be compared with John Mackie's (1978a) examination of whether
the need for universalization is 'a logical thesis* or 'a substantive practical thesis'
(P-96).
17
On the scope of objectivity, see Nagel (1980, 1986), McDowell (1981, 1985),
Wiggins (1985,1987), H. Putnam (1987,1991), and Hurley (1989). Onthe other hand,
see also Harman (1977), Mackie (1978a, 1978*) and B. Williams (1981, 1985).
l
* Some particular aspects of this question are discussed in Sen (19836, 1985a).

some form or other, provide a shared background to all the major
ethical and political proposals in this field that continue to receive
argued support and reasoned defence.19
One consequence of all
this is the acceptance—often implicit—of the need to justify dispar-
ate advantages of different individuals in things that matter. That
justification frequently takes the form of showing the integral con-
nection of that inequality with equality in some other important—-
allegedly more important—space,20
Indeed, it is equality in that more important space that may then
be seen as contributing to the contingent demands for inequality in
the other spaces. The justification of inequality in some features is
made to rest on the equality of some other feature, taken to be
more basic in that ethical system. Equality in what is seen as the
'base* is invoked for a reasoned defence of the resulting inequalities
in the far-flung 'peripheries*.
1.3, HUMAN DIVERSITY AND BASAL EQUALITY
Human beings differ from each other in many different ways. We
have different external characteristics and circumstances. We begin
life with different endowments of inherited wealth and liabilities.
" The remark here applies specifically to social arrangements—and thus to
theories in political philosophy rather than personal ethics. In the ethics at personal
behaviour, powerful arguments have been presented in favour of permitting or re-
quiringexplicit asymmetries in the treatmentof different people. Such arguments may
relate, for example, to the permissibility—perhaps even the necessity—of paying
special attention to one's own interests, objectives and principles, vis-a-vis those of
others. Or they may relate to the requirement of assuming greater responsibility
towards one's own family members and others to whomone is 'tied*. Different types
of asymmetries involved in persona! ethics are discussed in B. Williams (I973a,
19736, 1981), Mackie (19780), Nagel (1980, 1986), Scheflter (1982), Sen (19826,
19836), Regan (1983), and Parfit (1984). While these requirements can also be seen in
terms of demandsfor equalityof rather special types, they would tend to go against
the usual political conceptions of 'anonymous* equality (on this see Sen 1970a),
20
This greater importance need not be seen as intrinsic to the space itself. For
example, equality of primary goods in Rawls's (1971,1982,1985,1988a) analysis, or
of resources in Ronald Dworkin's (1981, 1987) theory is not justified on grounds of
the intrinsic importance of primary goods or of resources. Equality in these spaces is
seen as important because they are instrumental in giving people equitable opportu-
nity, in somesense, to pursue theirrespective goals and objectives. This distance does,
in fact, introduce—I would claim— some internal tension in these theories, since the
derivative importanceofprimary goods or resources depends on the respective oppor-
tunities to convert primary goods or resources into the fulfilment of the respective
goals, or into freedoms to pursue them. The conversion possibilities can, in fact, be

We live in different natural environments-—some more hostile than
others. The societies and the communities to which we belong offer
very different opportunities as to what we can or cannot do. The
epidemiologieal factors in the region in which we live can pro-
foundly affect our health and well-being,
But in addition to these differences in natural and social environ-
ments and external characteristics, we also differ in our personal
characteristics (e.g. age, sex, physical and mental abilities). And
these are important for assessing inequality. For example, equal
incomes can still leave much inequality in our ability to do what
we would value doing. A disabled person cannot function in the
way an able-bodied person can, even if both have exactly the same
income. Thus, inequality in terms of one variable (e.g. income)
may take us in a very different direction from inequality in the
space of another variable (e.g. functioning ability or well-being).
The relative advantages and disadvantages that people have,
compared with each other, can bejudged in terms of many different
variables, e.g. their respective incomes, wealths, utilities, re-
sources, liberties, rights, quality of life, and so on. The plurality of
variables on which we can possibly focus (the focal variables) to
evaluate interpersonal inequality makes it necessary to face, at a
very elementary level, a hard decision regarding the perspective to
be adopted. This problem of the choice of the 'evaluative space*
(that is, the selection of the relevant focal variables) is crucial to
analysing inequality.
The differences in focus are particularly important because of
extensive human diversity. Had all people been exactly similar,
equality in one space (e.g. incomes) would tend to be congruent
with equalities in others (e.g. health, well-being, happiness). One
of the consequences of 'human diversity' is that equality in one
space tends to go, in fact, with inequality in another.
For example, we may not be able to demand equality of welfare
levels and other such 'patterning*—-to use Nozick's helpful
description—once we demand the equality of libertarian rights as
specified by Nozick (1974). If equal rights, in this form, are
accepted, then so must be all their consequences, and this would
very diverse for different people, and this does, Iwould argue, weaken the rationale of
the derivative importance of equality of holdings of primal^ goods or resources.On
this, see Chs, 3 and 5 (also Sea 1980«, 19906).

include all the generated inequalities of incomes, utilities, well-
being, and positive freedoms to do this or be that.
I am not examining, here, how convincing this defence is.2
' The
important issue in the present discussion is the nature of the strategy
of justifying inequality through equality. Nozick's approach is a
lucid and elegant example of this general strategy. If a claim that
inequality in some significant space is right (or good, or acceptable,
or tolerable) is to be defended by reason (not by, say, shooting the
dissenters), the argument takes the form of showing this inequality
to be a consequence of equality in some other—more centrally
important—space. Given the broad agreement on the need to have
equality in the 'base*, and also the connection of that broad agree-
ment with this deep need for impartiality between individuals (dis-
cussed earlier), the crucial arguments have to be about the reason-
ableness of the 'bases* chosen. Thus, the question: 'equality of
what?* is, in this context, not materially different from the enquiry:
'what is the right space for basal equality?" The answer we give to
'equality of what?* will not only endorse equality in that chosen
space (the focal variable being related to the demands of basal
equality), but will have far-reaching consequences on the distribu-
tional patterns (includingnecessary inequalities) in the other spaces.
'Equality of what?' isindeed a momentous—and central—question.
1.4. EQUALITY VERSUS LIBERTY?
The importance of equality is often contrasted with that of liberty.
Indeed, someone's position in the alleged conflict between equality
and liberty has often been seen as a good indicator of his or her
general outlook on political philosophy and political economy.
For example, not only are libertarian thinkers (such as Nozick
1974) seen as anti-egalitarian, but they are diagnosed as anti-
egalitarian precisely because of their overriding concern with
liberty,22
Similarly, those diagnosed as egalitarian thinkers (e.g.
Dalton 1920, Tawney 1931, or Meade 1976) may appear to be less
concerned withliberty precisely because they are seen as beingwed-
ded to the demands of equality.
In the light of the discussion in the previous sections, we must
21
Some criticismsof that approach can be found in Sen (1982A, 1984).
22
I refer here specifically to Nozick (1973, 1974). For a reassessment and refine-
ment of his position, see Nozick (1989).

argue that this way of seeing the relationship between equality and
liberty is altogether faulty. Libertarians must think it important
that people should have liberty. Given this, questions would
immediately arise regarding: who, how much, how distributed, how
equaF! Thus the issue of equality immediatelyarises as a supplement
to the assertion of the importance of liberty.23
The libertarian pro-
posal has to be completed by going on to characterize the distribu-
tion of rights among the people involved.24
In fact, the libertarian
demands for liberty typically include important features of 'equal
liberty*, e.g. the insistence on equal immunityfrom interferenceby
others. The belief that liberty is important cannot, thus, be in
conflict with the view that it is important that the social arrange-
ments be devised to promote equality of liberties that people have.
There can, of course, be a conflict between a person who argues
for the equality of some variable other than liberty (such as income
or wealth or well-being) and someone who wants only equal
liberty. But that is a dispute over the question 'equality of whatT
Similarly, a distribution-independent general promotion of liberty
(i.e. promoting it wherever possible withoutpaying attention to the
distributive pattern) could, of course, conflict with equality of
some other variable, say, income, but that would be (1) partly a
conflict between concentrating respectively on liberty and on in-
comes, and (2) partly one between a concern for distributive pat-
terns (of incomes in this case) and non-distributive aggregative con-
siderations (applied to liberty). It is neither accurate nor helpful to
think of the difference in either case in terms of 'liberty versus
equality'.
Indeed, strictly speaking, posing the problem in terms of this
latter contrast reflects a 'category mistake*. They are not alterna-
tives. Liberty is among the possible fields of application of equality,
13
There can be quite different ways of defending the importance of liberty. One
distinction relates to the different concepts of goodness and Tightness. First, liberty
can be seen as a good thing that people should have, and the violation of liberty may
be seen as making the state of affaire less good. Second, liberty may be taken to be not
a part of the idea of goodness, but a feature of right social arrangements. There are
distinctions—not unrelated to the above contrast—also between what duties others
have if someone's liberties are violated. I have tried to discuss these questions else-
where {see Sen 197ft», 19826, 1983«, 1992a), and will not pursue them further here.
24
See, in this context, Rawis's (1971) discussion of the priority of 'equal liberty'
(ch. 4). Seealso Berlin (1955-6,1969), Wollheim (1955-6), Hayek (1960,1967), Buch-
anan (1975, 1986), Haksar (1979), Gutmann (1980), Goodin (1988), Suppes (1988),
and Lukes (1990).

and equality is among the possible patterns of distribution of
liberty,"
As was discussed earlier, the need to face explicitly the choice of
space is an inescapable part of the specification and reasoned evalu-
ation of the demands of equality. There are, at one end, demands of
equal libertarian rights only, and at the other end, various exacting
demands of equality regarding an extensivelist of achievements and
also a corresponding list of freedoms to achieve. This study is much
concerned with this plurality and its manifold consequences.
1.5, PLURALITY AND ALLEGED EMPTINESS
The recognition of plurality of spaces in which equality may be
assessed can raise some doubts about the content of the idea of
equality. Does it not make equality less powerful and imperative as
a political idea? If equalitycan possibly speak with so many voices,
can we take any of its demands seriously?2
*
Indeed, the apparent pliability of the contents of equality has
appeared to some analysts as a source of serious embarrassment for
the idea of equality. As Douglas Rae (1981) has put it (in his
meticulous and helpful exploration of the various contemporary
notions of equality), 'one idea that is more powerful than order or
efficiency or freedom in resisting equality' is 'equality itself
(P-151).
While Rae argues that the idea of equality is, as it were, 'over-
full', others have argued, on similar grounds, that equality is 'an
empty idea*—it is 'an empty form having no substantive content of
its own'.27
Since equality can be interpreted in so many different
K
There can, of course, be some ambiguity regarding what is called a 'pattern'.
Sometimes the term 'pattern' may be used to impose particular specifications of
constituent characteristics, e.g. the Union Jack demands some blue and some red.
The appropriate analogy for equality tnd liberty is with the distinction between, say,
the pattern of intensities of colours (e.g. the same intensity for each unit, or maximal
intensity altogether), and the use of particular colours (e.g. blue) the intensities of
which arc examined,
24
There is also a related but distinct issue as to whether equality can provide a
deep enough justification for any social structure. Robert Goodin (1988) asks an
interesting question as to whether the 'apparent egalitarianism" underlying 'welfare
state practices* are ultimately just 'epiphenomenaP (pp. 51-49). The argument
depends, as Goodin notes, on how equality is defined, and his affirmative answer to
the question draws on the conflict between different views of equality (including that
implicit in what he calls 'impartiality*).
27
Wcsten (1982: 596).

ways, the requirement of equality cannot, in this view, be taken to
be a truly substantive demand.
It is certainly true that merely demanding equality without say-
ing equality of what,cannot be seen as demanding anything specific,
This gives some plausibility to the thesis of emptiness. But the thesis
is, I believe, erroneous nevertheless. First, even before a specific
space is chosen, the general requirement of the need to value
equality in somespace that is seen to beparticularly important is not
an empty demand. This relates to the discipline imposed by the
need for some impartiality, some form of equal concern, At the very
least, it is a requirement of scrutiny of the basis of the proposed
evaluative system, It can also have considerable cutting power, in
questioning theories without a basal structure and in rejecting those
that end up without a basal equality altogether. Even at this
general level, equality is a substantive and substantial require-
ment.
Second, once the context is fixed, equality can be a particularly
powerful and exacting demand. For example, when the space is
fixed, demands for equality impose some ranking of patterns, even
before any specific index of equality is endorsed. For example, in
dealing with the inequality of incomes, the so-called 'Dalton prin-
ciple of transfer* demands that a small transfer of income from a
richer person to a poorer one—keeping the total unchanged—must
be seen to be a distributive improvement.n
In its context, this is a
fairly persuasive rule in ranking distributions of the same total
income by the general requirement of equality without invoking
any specific index or measure.
In addition to such ordering of patterns in a givenspace, even the
broader exercise of the choice of space itself may have clear links
with the motivation underlying the demand for equality. For
example, in evaluating justice, or social welfare, or living stan-
dards, or quality of life, the exercise of choice of space is no longer
just formal, but one of substantive discrimination. As I shall try to
show in the chapters that follow, the claims of many of these spaces
can be forcefully disputed once the context is fixed. Though this
need not lead us to one precise characterization of the demands of
equality that is important in every context, this is far from a real
» On this see Dalton (1920), Koto (1969), Atkinson (19706, 1983). On some
further normative implications of this property, see Dasgupta, Sen, and Starrett
(1973) and Rothschild and Stiglitz (1973), and also OBJ.ch. 3.

embarrassment. In each context, the demands of equality may be
both distinct and strong.
Third, the diversity of spaces in which equality may be de-
manded really reflectsa deeper diversity, to wit, different diagnoses
of objects of value—different views of the appropriate notions of
individual advantage in the contexts in question. The problem of
diversity is, thus, not unique to equality evaluation. The different
demands of equality reflect divergent views as to what things are to
be directly valued in that context. They indicate different ideas as
to how the advantages of different people are to be assessed vis-a-
vis each other in the exercise in question. Liberties, rights, utilities,
incomes, resources, primary goods, need-fulfilments, etc., provide
different ways of seeing the respective lives of different people, and
each of the perspectives leads to a corresponding view of equality.
This plurality—that of assessing the advantages of different per-
sons—reflects itself in different views not merely of equality, but
also of any other social notion for which individual advantage
substantially enters the informational base. For example, the no-
tion of 'efficiency* would have exactly the same plurality related to
the choice of space.2
* Efficiency is unambiguously increased if there
is an enhancement of the advantage of each person (or, an advance-
ment for at least one person, with no decline for any), but the
content of that characterization depends on the way advantage is
defined. When thefocal variable is fixed, we get a specific definition
of efficiency in this general structure.
Efficiency comparisons can be made in terms of different vari-
ables. If, for example, advantage is seen in terms of individual
utility, then the notion of efficiency immediately becomes the con-
cept of 'Pareto optimality', much used in welfare economics. This
demands that the situation is such that no one's utility can be
increased without cutting down the utility of someone else. But
efficiency can also be similarly defined in the spaces of liberties,
rights, incomes, and so on. For example, corresponding to Pareto
optimality in the space of utilities, efficiency in terms of liberty
would demand that the situation is such that no one's liberty can
be increased without cutting down the liberty of someone else.
There is, formally, an exactly similar multiplicity of efficiency
9
While the plurality is exactly similar in principle, it is possible that empirically
there may be more space-related divergence between inequality comparisons than
between efficiency comparisons; on this see Sen (19926).

notions as we have already seen for equality, related to the plurality
of spaces.
This fact is not surprising, since the plurality of spares in which
equality may be considered reflects a deeper issue, viz, plurality
regarding the appropriate notion of individual advantage in social
comparisons. The choice between these spaces is undoubtedly an
integral part of the literature of inequality evaluation. But the plur-
ality of spaces really reflects diversities in substantive approaches to
individual advantage, and in the informational base of interper-
sonal comparisons. Space plurality is not a unique problem—nor of
course a source of special embarrassment—for the idea of equality
as such.
1.6. MEANS AND FREEDOMS
It was suggested earlier that the class of normative theories of social
arrangements with which we are concerned demand—for reasons
that we discussed—equality in some space or other. This equality
serves as the 'basal equality*of the system and has implications on
the distributive patterns in the other spaces. Indeed, basal equality
may be directly responsible for inequalities in the other spaces.
It may be useful to discuss an example or two of the choice of
space and its importance. In modern political philosophy and eth-
ics, the most powerful voice in recent years has been that of John
Rawls (1971), His theory of'justice as fairness* provides an interest-
ing and important example of the choice of space and its conse-
quences. In his "Difference Principle', the analysis of efficiency and
equality are both related to the individual holdings of primary
goods.30
With that system, the diversity of inherited wealth and of talents
30
It is the Difference Principle which isconcerned with the distribution of primary
goods in the Rawlsian two principles of justice as fairness. It is worth noting—so as
not to oversimplify matters—that (!) Rawls's first principle, which has priority, deals
only with personal liberties (and demands equal liberty); (2) the Difference Principle
is concerned not only with distributiveconsiderations but also with efficiency (in the
form that any change that improves the position of all—including the worst off
group—is regarded as an improvement),and (3) theprinciples stated are not meantas
mechanical formulae, and a good deal of explanation and analysis of their use is
presented by Rawls as part and parcel of his theory of justice as fairness (for recent
clarifications on the exact claims in this theory, see Rawls 1985,19880,1988ft, !988c,
1990; see also Laden 19910). Notwithstanding these qualifications, it is obvious that
equality of the holdings of primarygoods has an importantplace in Rawls's structure
of political ethics.

would not generate income inequalityin the same way as in Nozick's
system, since the primary goods—on the distribution of which
Rawls's Difference Principle imposes an egalitarian requirement-
include incomes among their constitutive elements. Incomes are,
thus, directly covered in the Rawlsian demands of basal equality.
But the relationship betweenprimary goods (including incomes), on
the one hand, and well-being,on the other, may vary because of
personal diversities in the possibility of converting primary goods
(including incomes) into achievements of well-being. For example,
a pregnant woman may have to overcome disadvantages in living
comfortably and well that a man at the same age need not have, even
when both have exactly the same income and other primary goods,
Similarly, the relationship between primary goods and the
freedom to pursue one's objectives—well-being as well as other
objectives—may also vary." We differ not only in our inherited
wealths, but also in our personal characteristics. Aside from purely
individual variations (e.g. abilities, predispositions, physical
differences), there are also systematic contrasts between groups (for
example between women and men in specific respects such as the
possibility of pregnancy and neonatal care of infants). With the
same bundle of primary goods, a pregnant woman or one with in-
fants to look after has much less freedom to pursue her goals than a
man not thus encumbered would be able to do. The relationship
between primary goods, on the one hand, and freedom as well as
well-being, on the other, can vary with interpersonal and intergroup
variations of specific characteristics.32
Inequalities in different 'spaces* (e.g. incomes, primary goods,
liberties, utilities, other achievements, other freedoms) can be very
different from each other depending on interpersonal variations in
the relationsbetween these distinct—but interconnected—variables.
One consequence of the basic fact of human diversity is to make it
particularly important to be sure of the space in which inequality is
to be evaluated. Person 1can have more utility than 2 and 3, while2
31
On this question, see Sen (J990fr).
32
Rawls (1985, 1987, 19880) himself has emphasized another type of diversity
among the individuals, to wit, differences between their respective conceptions of the
good. This leads to differences in the objectives which they respectively have reasons
to pursue. That heterogeneity has to bedistinguished from the diversity in the ability
to convert resources and primary goods into the fulfltaent of objectives (or into the
ability to fulfil objectives). Neither diversity entails the other, and it is important to
consider both types of interpersonal variations. These issues arc discussed in Ch. 5,

has more income than I and 3, and 3 is free to do many thingsthat I
and 2 cannot. And so on. Even when the rankingsare the same, the
relative distances (i.e. the extent of the superiority of one position
over another) could be very diverse in the different spaces.
Some of the most central issues of egalitarianism arise precisely
because of the contrast between equalityin the different spaces. The
ethics of equality has to take adequate note of our pervasive
diversities that affect the relations between the different spaces. The
plurality of focal variables can make a great difference precisely
because of the diversity of human beings,
1.7. INCOME DISTRIBUTION, WELL-BEING AN0 FREEDOM
Our physical and social characteristics make us immensely diverse
creatures. We differ in age, sex, physical and mental health, bodily
prowess, intellectual abilities, climatic circumstances, epidem-
iological vulnerability, social surroundings, and in many other
respects. Such diversities, however, can be hard to accommodate
adequately in the usual evaluative framework of inequality assess-
ment. As a consequence, this basic issue is often left substantially
unaddressed in the evaluative literature.
Aa important and frequently encountered problem arises from
concentrating on inequalityof incomes as the primary focus of atten-
tion in the analysis of inequality. The extent of real inequality of
opportunities that people face cannot be readily deduced from the
magnitude of inequalityof incomes, since what wecan or cannot do,
can or cannot achieve, do not depend just onour incomesbut alsoon
the variety of physical and social characteristics that affect our lives
and make us what we are.
To take a simple illustration, the extent of comparative depriva-
tion of a physically handicapped person vis-a-vis others cannot be
adequately judged by looking at his or her income, since the person
may begreatly disadvantaged in converting income into the achieve-
ments he or she would value.31
The problem does not arise only
53
The importance of coming to grips withcaws of this kind was discussed fairly
extensively in OEI, ch. I, It was treated there mainly as the basis of a critique of
utilitarianism and its exclusive concern with summing utilities. I have nothing to
withdraw from that critique, but the problem is, in fact, much more pervasive than I
had argued there. It will become clear, as wereviewother standard approaches, that
nearly mil of them tend to fail to do justice to the problem that is illustrated by this
caw.

from the fact that income is just a means to our real ends, but (1)
from the existenceof other important means, and (2) from interper-
sonal variations in the relation between the means and our various
ends.
These issues have on the whole tended to be neglected in the
literature on the measurement of inequality in economics. For
example, consider the approach to constructing Inequality indices*
based on social loss of equivalent income pioneered by Atkinson
(1970ft).34
This approach has been, in many ways, remarkably in-
fluential and productive in integrating considerations of income-
inequality with the overall evaluation of social welfare,35
The ex-
tent of inequality is assessed in this approach by using the same
response function «(y) for all individuals, deined over personal
incomes.3
* This strategy of inequality measurement, thus, incorpor-
ates the restrictive feature of treating everyone's incomessymmetri-
cally no matter what difficulties some people have compared with
others in converting income into well-being and freedom."
34
This welfare-economic approach to inequality evaluation is further discussed in
Ch. 6 below.
ss
The approach is extensively discussed in OEt, eh. 3. For illuminating accounts
acid assessments of the recent literature on inequality evaluation—including the
influence of Atkinson's approach on that literature—sec Blackorby and Donaldson
(1978,1984) and Foster(1985). Atkinson (1983) himself has provided a critical evalu-
ation of that literature and commented on some of the questions that have been
raised. See also Koto (1969, 1976) on related matters.
M
This a function has usually been interpreted as a 'utility function*. But u need
not necessarily be seen as 'utility*; on this see Atkinson (1983: 5-4). Socia! welfare is
taken to be an additively separable function of individual incomes. The bits of social
welfare dependent on the respective persons' incomes are derived from the same
function for everyoneand then added up together to yield aggregate social welfare.If
u is takenas utility (a permissible view, providingperhaps the simplest—certainly the
most common—interpretation), then the assumption of the same u function for all
amounts to that of the same utility function for everyone. But more generally, no
matterwhatinterpretation of «(j>) ischosen, that function must havethis characteristic
of being the same for all. Similarly, in the extension of the Atkinson measure to a
not-necesiarily additively separable format proposed in my OEl (pp. 38-42), the
assumption of a symmetric aggregate W function entails that everyone's income
would have the same overall impact. While formally all this is consistent with many
different underlying stories, the central case is based on the presumption of the same
conversion relation (between income and achievement) for different people. On the
general issue of conversion, sec Fisher and Shell (1972), Sen (1979e), and Fisher
(1987),
31
Taking the same utilityfunction for all, relating utility to income (or to income
and work) isalso quite standard in many other branches of resource allocation, e.g. in
the literature on 'optimum taxation* pioneered by James Mirrlees (1971); Tuomala
(1971) provides a helpful account of that literature. This applies also to the literature
on cost-benefit analysis (see the critical survey by Dreze and Stern, 1987),

It is, of course, true that the object of this approach is to assess
inequality specifically in the distribution of incomes, not in levelsof
well-being. But that assessment is done in the light of what is
achieved from the respective person's income, and these achieve-
ments make up the aggregate 'social welfare'. Income inequality is
assessed by Atkinson in terms of the loss of social welfare (in units
of equivalent aggregate income) as a result of inequality in the
distribution of aggregate income,38
Given this motivation,it will in
general be necessary to bring in the effects of other influences on
people's lives and well-being to assess income inequality itself.39
In
general the measurement of inequality has to bring in information
regarding other spaces—both (1) for the purpose of evaluatingin-
equality in these spaces themselves,and (2) for that of assessing
income inequality in a broader framework, taking note of the pres-
ence of other influences on the objective (in Atkinson's case, social
welfare) in terms of which income inequality is to be ultimately
assessed. These issues will be further examined in Chapter 6,
The tendency to assume away interpersonal diversities can ori-
ginate not only from the pragmatic temptation to maketheanalytics
simple and easy (as in the literature of inequality measurement), but
also, as was discussed earlier, from the rhetoric of equality itself
(e.g. 'all men are created equal*). The warm glow of such rhetoric
can push us in the direction of ignoring these differences, by taking
*no note of them', or by 'assuming them to be absent*. This suggests
an apparently easy transition between one space and another, e.g.
from incomes to utilities, from primary goods to freedoms, from
resources to well-being. They reduce—again only apparently—the
tension between different approaches to equality.
But that comfort is purchased at a heavyprice. As a result of that
assumption, weare made to overlook the substantive inequalities in,
say, well-being and freedom that may directly result from an equal
distribution of incomes (given our variable needs and disparate per-
sonal and social circumstances). Both pragmatic shortcuts and
grand rhetoric can be helpful for some purposes and altogether
unhelpful and misleading for others.
M
The approach (see Atkinson I970h, 1975, 1983) develops a line of analysis
originally explored by Dalton (1920), and revived also by Koto (1969), The main
lines of the approach and the underlyinganalytics are also discussed in OEL
" For insightful remarks on this and related issues, seeAtkinson (1983: Part I).

2
FREEDOM, ACHIEVEMENT AND
RESOURCES
2.1, FREEDOM AND CHOICE
A person's position in a social arrangement can be judged in two
different perspectives, viz. (}) the actual achievement, and (2) the
freedom to achieve. Achievement is concerned with what we manage
to accomplish, and freedom with the realopportunity that wehave to
accomplish what we value. The two need not be congruent. In-
equality can be viewed in terms of achievements and freedoms, and
they need not coincide. That distinction is relevant for judging effi-
ciency as well, which can be seen in terms of individual achieve-
ments orfreedomsto achieve. Thus, the distinction between achieve-
ment and freedom is quite central to social evaluation.1
There are, ofcourse, different ways ofjudgingachievement, e.g.by
utility (such as pleasures achieved or desires fulfilled), or by opu-
lence (such as incomes earned or consumptions enjoyed), or by the
quality of life (such as some measures of living standards).2
As was
discussed in the previous chapter, the choice between these different
spaces is a question of profound significance in evaluating our lives
and attainments, and it is also of central importance in inequality
evaluation (specifically in answering the question 'equality of
what?*). But even with any given way of characterizing achievement,
there is a further issue, viz. the distinction between (1) the extent of
achievement, and (2) thefreedom to achieve.
Some well-known approaches to the evaluation of individual
advantage and to the assessment of good social orders have been
concerned directly with achievement only, treating the importance
1
The relevance of the perspective of freedom in the context of evaluation of
different types of economies has been examined by Kornai (1988), Lindbeck (1988),
and Sen (19880), to a symposium arranged by the European Economic Association.
2
I haveaddressed some of these issues in myTanner Lectures on 'the standard of
living". See Sen et al, (1987), which also include$ comments and further analyses
presented by the discussants, viz.John Muellbauer, Ravi Kanbur, Keith Hart,
Bernard Williams, and Geoffrey Hawthorn, who also edited this volume.

32 Freedom, Achievement and Resources
of the freedom to achieve as being entirely instrumental—as means
to actual achievements. Utilitarianism is an obvious example. The
utilitarian approach is characterized by (1) confining interpersonal
comparisons for social assessment to achievements only, and (2)
identifying achievements with the utilities achieved. The two to-
gether yield the utilitarian informational focus on interpersonally
compared individual utilities for personal and social assessment.
Similarly, the Bergson-Samuelson social-welfare functions, as
presented and explored in their classic formulations (see Bergson
1938; Samuelson 1947), tend to confine direct attention only to
achievements (e.g. preference-fulfilment, consumer satisfaction),
valuing individual freedom only indirectly as means to achieve-
ment. To some extent this applies also to the corresponding
Arrovian social-choice frameworks,3
which too have been mainly
geared towards individual preferences over states of affairs, rather
than taking note ofthe freedom to choose amongstates of affairs. But
much depends on how states are characterized—in particular
whether the choices available to move to other states are taken to be
part of each particular state.4
Recent developments in social-choice
theory have attempted to bring in considerations of freedom into the
framework of evaluation, in the specific context of valuing liberty.5
J
SeeArrow (1951, 1963),
4
Since completing this manuscript, I have tried to examine the possibility of
integrating the perspective of freedom in the structure of social choice theory in my
Kenneth Arrow Lectures at Stanford University in May 1991 ('Freedom and Social
Choke'),
5
An attempt in that direction wasmade in Sen (197Qa, I970c). The large iteratyre
on the treatment of liberty in social choice theory has been well discussed and scruti-
nized by Suzuniura (1983), WrigSesworth (1985), and Riley (1987), among others. In
dealing with the importance of liberty, the social-choice format has both advantages
and limitations compared with (1) traditional deontological statements, (2) standard
welfare-economic formulations, and (3) game-theoretic interpretations. These and
related issues have been discussed, along with examining ways of integrating con-
siderations of liberty in social choice frameworks, in an extensive iteratore, in-
cluding inter alia, Sen (197%, I976c, 1982*. 1982c, 1983o, 19920), Ng (1971, 1979),
Batra and Pattanaik (1972), Peacock and Rowley (1972), Nozick (1973, 1974),
Bernholz (1974, 1980), Gibbard (1974), Blau (1975), Fine (1975ft), Seidl (1975,
19866, 1990), Campbell (1976, 1989), Farrell (1976), Kelly (19760, 1976ft, 1978),
Aldrich (1977), Breyer (1977), Perelli-Minetti (1977), Ferejohn (1978), Kami (1978),
Stevens and Foster (1978), Sreumura (1978,1980, 1983, 1991), Austen-Smith (1979,
1982), Mueller(1979),Barnes (1980), Breyer and Gardner (1980), Breyer and Gigliolti
(1980), Fountain (1980), Gardner (1980), McLean (1980), Weale (1980), Gaertner and
Kruger (1981,1983),G9rdenfors(l98l), P. J. Hammond (1981,1982), Schwartz(1981,
1986), Sugden (1981, 1985), Levi (1982,1986), Wriglesworth (1982,1985), Chapman
(1983), K. Basu (1984), Gaertner (1985, 1986), Kelsey (1985,1988), Schotter (1985),
Coughlin (1986), Barry (1986), Eister and Hylland (1986), Hylland (1986), Webster

Freedom, Achievementand Resources 33
The exclusive focus on achievements has recently been seriously
challenged by arguments for basing political evaluation on the
means to achievement, such as the Rawlsian concern with the dis-
tribution of "primary goods', the Dworkinian concentration on the
distribution of'resources', and so on. Sincethe means in the form of
resources, primarygoods, etc., undoubtedlyenhance the freedom to
achieve (other things remaining the same), it is not unreasonable to
think of these moves as taking us towards freedom-—away from
attention being confined exclusively to evaluating achievement. If
we aim at equalityin the space of resources or of primary goods, this
can beseen as moving the evaluativeexercise towards the assessment
of freedom away from that of achievement as such,
But it must be recognized at the same time that equalizing
ownership of resources or holdings of primary goods need not
equalize the substantive freedoms enjoyed by different persons,since
there can be significant variations in the conversion of resources and
primary goods into freedoms. The conversion problems can involve
some extremely complex social issues, especially when the achieve-
ments in question are influenced by intricate intragroup relations
and interactions.*But, as wasdiscussed earlier, variations in conver-
sion can also arise from simple physical differences. For example
(to illustrate with a simple case referred to earlier), a poor person's
freedom from undernourishment would depend not only on her
resources and primary goods (e.g. through the influence of income
on the ability to buy food), but also on her metabolic rates, gender,
pregnancy, climatic environment, exposure to parasitic diseases,
and so on. Of two persons with identical incomes and other primary
goods and resources (as characterized in the Rawlsian or Dwor-
kinian frameworks), one may beentirely free to avoid undernourish-
ment and the other not at all free to achieve this.7
The move awayfrom achievement to the means of achievement (in
(1986), Harel and Niton (1987), Maelntyre (1987, 1988), Mezwltl (1987), Nuimi
(1987), Riley (1987, I989a, 1989ft), Sonstegaard (1987), Subramanian (1987), Allen
(1988), Oigliotti (1988), Pattanaik (1988), S.O. Hansson (1988), Deb (1989), Glrden-
fors and Pettit (1989), A. Hwnlin (1989), Hurley (1989), VaMentyne (1989), Xu
(1990), Gaertner, Pattanaik, and Suzumura (1992), among other contributions.
* This is a particularly important question in the assessment of poverty involving
the failure to achievesome minimalsocial functionings such as taking part in the life
of the community. On this see Sen (1983*4 and also Ch. 7 below.
' Onthepractical importance ofthis type ofissue, with empirical illustrations, see
Sen (19856), particularly Appendices A and B. See also Arneson (19890, 199Q«,
19906, 1991), and G. A. Cohen (1989, 1990, 1992).

the form of Rawls's focusing on primary goods or Dworkin's con-
centration on resources8
) may we!! have helped to shift the attention
of the literature in the direction of seeing the importance of freedom,
but the shift is not adequate to capture the extent of freedom. If our
concern iswith freedom as such, then there isno escape from looking
for a characterization of freedom in the form of alternative sets of
accomplishments that we have the power to achieve. Before that
more exacting question is addressed (in Section 2.3 and in Chapters
3-5), another aspect of the achievement-freedom distinction related
to standard economic theory may be worth examining,
2.2, REAL INCOME, OPPORTUNITIES AND SELECTION9
The distinction between achievement and freedom can be illustrated
by two different interpretations of real-income analysis that can be
found in the literatureand which are often not clearlydistinguished.
The evaluation of real income can.be seen as the assessment of the
benefit that a person receives from a particular bundle of com-
modities that he or she acquires: 'Is x & better bundle for this person
than yT The comparison is directly of the nature of the bundles
selected, and we may call it the 'selection view'. Alternatively, the
focus can be notjust on the particular bundles purchased, but on the
set of all bundles that the person could have bought with that in-
come: 'Does the budget set A give this person a better set of options
than does the budget set -§?' This may be called the 'options view'.
The rationale of the axiomatic structure of real-income com-
parisons depends on the interpretation of the comparisons, and both
the selection viewand the options view have been extensivelyused—
explicitly or by implication." The use of the options view has taken
the form of asserting the superiority of choosing x from set A over
choosing y from set B if set A offers, inter alia, the possibility of
choosing y as well." This is the 'revealed preference' approach to
* In Dworkin'sproposal much would, in fact, depend on the scope and reach of his
insurance mechanisms against personal handicaps.
* This section invokes somestandardprocedures ineconomictheory.This makes it
a little bit more 'technical* than the rest of the monograph,but no great specialized
knowledge has been presupposed. Also, the text is so organized thata non-economist
reader should be able to follow subsequent sections even if this section is skipped.
10
These uses have been examined and scrutinized in Sen (I979c).
'' Note that this is not itfull comparison of the twosets A and B, but takes the form
of looking at A in the light of the chosen eiemem y from B,

Freedom, Achievement and Resources 35
real-income comparison.12
Its rationale is based on utilizing the
prlee-and-purchase data to make a particular comparison of the
freedom to choose, to wit, whether in A one could choose what was
chosen from B.
The selection view, on the other hand, is focused on comparing
just the bundles x and y, and this is done by assuming a particular
structure of preferences (specifically convexity—essentially, non-
increasing marginal rates of substitution). The data on prices and
purchases are used to ascertain the relative weights attached to the
different commodities locally (i.e. for the respective amounts actu-
ally purchased), and the assumption of non-increasing marginal
rates of substitution is used to determine whether the bundle x
chosen from A can be declared to be superior to y chosen from B,"
The opportunities of choice, as such, are not compared in this
exercise.
The two approaches give similar results in real-income com-
parison under standard assumptions, but (1) they involve very
different strategies, and (2) they lead to different results once the
standard assumptions (e.g. competitive markets) are dropped.14
Here we are concerned with the differences between the two strat-
egies, in particular with the contrast between comparing options and
that of comparing the goodness of the selected bundlesdirectly,
I should, however, add that even in the use of the options view in
the revealed-preference approach—as the analysis has actually
been carried out—no intrinsic importance has, in fact, been attached
to the extent of freedom ofchoice as such. Instead, the opportunities
of choice have been seen only as means to acquiring preferred
bundles of commodities. But it is possible to extend the analyticsof
the logic of options-comparisons used by the revealed-preference
12
See Sarouelson (1938, 1947); also Houthakker (1950) and Little (1950), Com-
parison of options of choice is used in the 'revealed preference' approach to infer the
person's utility function on the basis of his choicesfrom different sets of bundles.That
programme of getting to utility from choices has a number of problematic features,
including the use of some demanding presumptions about the nature of the goals
pursued by the person, and the interpretationof the goal function as the individual's
personal utility function (on this, see Sen 19736, I977c). Here we are not directly
concerned with the strategy of 'recovering* personal utility functions from the
observed choices, but with the illumination provided by the procedure of comparing a
person's options of choice rather thanjust the chosen bundles. Thusweare making use
of a part of the chain of 'revealed preference' reasoning, but not all of it.
13
See Hicks (1939, 1940), Samuelson (1947), and Graaff (1957).
14
On this see Majumdar (1969) and Sen (1979e).

approach in the direction of attaching intrinsic importance to the
freedom of choice itself.IS
While the selection view goes straight at
the goodness of commodity bundles (nothing else is involved), the
options view can also be used to compare the extents of the freedom
of choice. In the revealed-preference literature, no intrinsic value is
attached to freedom as such and it is assessed in purely instrumental
terms (in terms of the value of the chosen bundle), but the same
technique of analysis can be adapted to take note of the intrinsic
value of freedom (if freedom is seen as important in itself).
2.3. FREEDOM DISTINGUISHED FROM RESOURCES
Wecome back now to the distinction between/*-eedom and the means
to freedom (such as primary goods or resources, which help one to
achieve more freedom). We may begin with seeing the distinction in
the context of commodity consumption, though that is not our ulti-
mate focus of attention. The 'budget set* represents the extent of the
person's freedom in this space, i.e. the freedom to achieve the con-
sumption of various alternative commodity bundles. This budget set
is derived on the basis of the person's resources (in this ease, the level
of income and the opportunity to buy commodities at given prices),
The distinction between (1) the resources on which the budget set
depends, and (2) the budget set itself, is a simple illustration of the
general distinction between the means to freedom and the extent of
freedom (in this case all seen in terms of consumption bundles that
can be achieved).
A shift in the attention from achievement to resources (e.g. from
the chosen commodity bundle to the income with which such
bundles could be bought) can indeed be seen (as was stated earlier)
as a move in the direction of paying greater attention to freedom,
since resources tell us about the set of commodity bundles from
which wecan choose."The strategy ofjudging individual advantage
by the person's command over resources, as opposed to what the
15
On the importance of the distinction between Instrumental and intrinsic valu-
ation of freedom, and on its implications for economic theory, sec Sen (t988a).
16
Note, however, that the resources owned or freedoms enjoyed by a person at a
particular point of time may have been the result of achievements in the past. In
considering the contrast between achievements and mains of freedom, there is no
intention to deny that connection. Indeed, a fuller formulation of the relationship
would require a proper intertemporal accounting of the interconnections involved.

Freedom, Achievement and Resources 37
individual actually achieves, is to refocus our vision from achieve-
ment to the means of freedom, and that is, in an obvious sense, a
homage to freedom.
Various important moves in contemporary political and moral
philosophy (e.g. Rawls's focus on primary goods in his theory of
justice, Dworkin's arguments for 'equality of resources'), which
have been partly motivated by a concern for the importance of
freedom, have tended to concentrate on the individual's command
over resources—in one form or another—as the basis of interper-
sonal comparisons of individual advantage,
While this has been a movein the right direction (as far asfreedom
is concerned), the gap between resources that help m to achieve
freedom and the extent of freedom itself is important in principle and
can be crucial in practice. Freedom has to be distinguished not
merely from achievement, but also from resources and means to
freedom,"
This problem of interpersonal variation may look rather un-
likely in the case of converting resources into commodity bundles,
since the assumptions of uniform prices, competitive markets, etc.,
are standardly made in the economic literature on these matters.
But that is partly a quirk of theory, and in actuality non-uniform
prices and other interpersonal variations in conversion are quite
common. But more importantly, once we shift attention from the
commodity space to the space of what a person can, in fact, do or be
(or what kind of a life a person can lead), the sources of inter-
personal variations in conversion can be numerous and power-
ful,18
The resources a person has, or the primary goods that someone
holds, may be very imperfect indicators of the freedom that the
17
Even in the commodity space, the set of commodity bandies over which the
person has command is a better representation of freedom (in that commodity space)
than the resources that can be used to establish command over a set of commodity
bundles. The latter will take us to the former, but the conversion will depend on
certain contingent circumstances, e.g. markets and prices. If these conversions vary
from person to person, then one person can be ahead in the resource space but behind
in terms of the freedom in the commodity space(e.g. with differential rents as a part of
public-housing policy).
" Oneof themore discussed examples of interpersonal variations relates to nutri-
tional 'requirements' of different people. For different ways of dealing with these
variations, see Sukhatme (1977, 1982), Srinivasan (1981, 1992), Gopalan (1983),
Upton (1983), Blaster and Waterlow (1985), Payne (1985), Vaidyanathan (1985),

person really enjoys to do this or be that.n
As was discussed in the
previous chapter, the personal and social characteristics of different
persons, which can differ greatly, can lead to substantial interper-
sonal variations in the conversion of resources and primary goods
into achievements. For exactly the same reason, interpersonal
differences in these personal and social characteristics can make the
conversion of resources and primary goods into the freedom to
achieve similarly variable.
If we are interested in the freedom of choice, then wehave to look
at the choices that the person does in fact have, and we must not
assume that the same results would be obtained by looking at the
resources that he or she commands. The moves towards resource-
based interpersonal comparisons in contemporary political philos-
ophy (such as those of Rawls and Dworkin) can certainly be seen as
taking us in the direction of paying attention to freedom, but the
moves are substantially inadequate.20
In general, comparisons of
resourcesand primary goods cannot serve as the basis for comparing
freedoms. Valuing freedom imposes exacting claims on our atten-
tion—claims that cannot be met by looking at something else.21
Scrimshaw (1987), Payne and Lipton (1988), Anand and Harris (19% 1992), Das.
gupta and Ray (1990), Osmani (1990a, I992«, 19926), Bhargava (1991), among other
contributors.
" For further discussion of this issue, seeSen (I980a, 19916). On related matters,
sec Lehning (1989) and Pogge (1989).
2
* More on this in Ch. 5 when weconsider the problem of inequality evaluation in
the context of theories of justice.
21
In the next chapter we examine and scrutinize the idea of freedom to achieve
well-being as a prelude to examininginequalities in freedoms.

3
FUNCTIONINGS AND
CAPABILITY
3.1. CAPABILITY SETS
This chapter explores the 'capability* perspective on the assessment
of (1) well-being, and (2) the freedom to pursue well-being. The
approach has been discussed in some detail elsewhere.' Here I shall
confine the presentation to only a few elementary aspects of this
perspective.
The well-beingof a person can be seen in terms of the quality (the
'well-ness', as it were) of the person's being. Living may be seen as
consisting of a set of interrelated 'functionings', consisting of beings
and doings. A person's achievement in this respect can be seen as the
vector of his or her functionings.The relevant functionings can vary
from such elementary things as beingadequately nourished, beingin
good health, avoiding escapable morbidity and premature mor-
tality, etc., to more complex achievements such as being happy,
having self-respect, taking part in the life of the community, and so
on.2
The claim is that functionings are constitutive of a person's
being, and an evaluation of well-being has to take the form of an
assessment of these constituent elements.3
Closely related to the notion of functionings is that of the
1
Various aspects of the conceptual foundation and practical problems of
measurement and evaluation have been discussed in Sen (19800,1985a, 19851)). This
chapter draws on Sen (99lh),
2
For helpful discussion of various constitutive elements of quality of life, see
AHardt (1981, 1992), Erijcson and Aberg (1987), Edison (1991), Ysander (1992).
Indeed, the 'Scandinavian studies' on livingconditions have done much to demon-
strate and clarify the empirical possibility of examining diverse functionings as the
basis of quality of life. On related matters, see also Fuohs (1983), Mack and Lansley
(1985), Culyer (1986), A. Williams (1991).
J
The philosophical basis of this approach can be traced to Aristotle's writings,
which include a penetrating investigation of "the good of man* in terms of'life to the
sense of activity' (see particularly The Nicomachean Ethics, i. 7). Aristotle had gone
on to examine—both in Ethics and to Politics—the political and social implications
of concentrating on well-being in this sense, involving'Iranian flourishing'. On the
Aristotelian approach and its connections with the recent explorations of the
capability perspective, see Nussbaurn (1988a, 19886).

40 Functionings and Capability
capability to function. It represents the various combinations of
functioniags (beings and doings) that the person can achieve,
Capability is, thus, a set of vectors of functionings, reflecting the
person's freedom to lead one type of life or another,4
Just as the
so-called 'budget set* in the commodity space represents a person's
freedom to buy commodity bundles,* the 'capability set' in the
functioning space reflectsthe person's freedom to choose from poss-
ible livings.
It is easy to seethat the well-being of a person must bethoroughly
dependent on the nature of his or her being, i.e. on the functionings
achieved. Whether a person is well-nourished, in good health, etc.,
must be intrinsically important for the welSness of that person's
being. But, it may be asked, how do capabilities—as opposed to
achievedfunctionings—relate towell-being?
The relevance of a person's capability to his or her well-being
arises from two distinct but interrelated considerations. First, if the
achieved functionings constitute a person's well-being, then the
capability to achieve functionings (i.e. all the alternative combina-
tions of functionings a person can choose to have)will constitute the
person's freedom—the real opportunities—to have well-being. This
'well-being freedom' may have direct relevance in ethical and
political analysis,* For example, in forming a view of the goodness
of the social state, importance may be attached to the freedoms that
different people respectively enjoy to achieve well-being. Alter-
natively, without taking the route of incorporating well-being
freedom in the 'goodness' of the social state, it may be simply taken
to be 'right' that individuals should have substantial well-being
freedom.1
4
There are several technical problems in the representation and valuation of
functioning vectors (more generally, functioning it-tuples) and the capability sets of
such vectors (or n-tuples); see Sen (19856, 19916).
3
On this see Ch. 2, Sect. 2.2.
* The particular relevance of 'well-being freedom' as opposed to achieved well-
being in social and political ethics is discussed in my Dewey Lectures (Sen 198Sa),
particularly in lectures 2 and 3. Those lecturesalso deal with the distinction between
'well-being freedom* and "agency freedom*. The latter stands for freedom of a more
general kind—the freedom to achieve whatever one's objectives are (possibly going
well beyond the pursuit of one's own well-being). See Ch. 4 below.
1
SeeRawls (1988a) on the importance of the distinction between "the right' arid
'thegood'. Seealso Sugden (1989) for a forceful presentation of thecase for seeing the
duties of the state in terms of ideas of the right rather than being based on maximizing
the good. In Sen (1987) it is argued inter alia that the distinction may be less clear-cut
and perhaps even less fundamental than is frequently assumed.

Functumings and Capability 41
This freedom, reflecting a person's opportunities of well-being,
must be valued at least for instrumental reasons, e.g. in judging
how good a "deal* a person has in the society. But in addition, aswe
have been discussing, freedom may be seen as beingintrinsically
important for a good social structure, A good society, in this view,
is also a society of freedom.8
It is also possible to use the notion of
"rightness* as opposed to 'goodness' of the society to argue for the
same substantive arrangements. Those who see that distinction as
being very fundamental, and argue for 'the priority of right over
ideas of the good' (as Rawls I988a puts it), would have to
approach this question from that end.
The second connection between well-being and capability takes
the direct form of making achieved well-being itself depend on the
capability to function. Choosing may itself be a valuable part of
living, and a life of genuinechoice with serious options may be seen
to be—for that reason—richer.* In this view, at least some types of
capabilities contribute directly to well-being, making one's life
richer with the opportunity of reflective choice. But even when
freedom in the form of capability is valued only instrumentally
(and the level of well-being is not seen as dependent on the extent
of freedom of choice as such), capability to function can neverthe-
less be an important part of social evaluation. The capability set
gives us information on the various functioning vectors that are
* The perspective of freedom is much associated with the important worksof liber-
tarian writers (e.g. Hayek 1960, 1967; Noziek 1974; J. M. Buchanan 1975, 1986;
among recent contributors). But an early argument for concentration on the basic
value of freedom can be found in Marx's political philosophy, with his emphasis on
bringing 'the conditions for the free development and activity of individualstinder
their own control*, to Marx's vision, the liberated future society would 'make it pos-
sible for metodo one thing to-day and another tomorrow, tohuntin the morning, §sh
in theafternoon, rear cattlein the evening,criticizeafter dinner,just as I haveinmind,
without ever becominghunter,fisherman,shepherd or critic' (Marx 1845-6:22). This
is, of course, an overall view of freedom (what one can, everythingconsidered, do) in
contrast with negative rights (what one is not prevented from doing) emphasized in
much of the libertarian literature; on this distinction, see Berlin (1969), On alterna-
tive approaches to the intrinsic value of freedom, see my 'Freedom of Choice: Con-
cept and Content* (Sen 19880). On the Marxian approach to freedom, seeKolakowski
(1978), C, Taylor (1979),Brenkert(1980,1983), A, E, Buchanan (1982), Elst«(1986),
Lukes (1985), G. A. Cohen (1986,1988,1989), Rmmachandran (1990), among other
contributors.
* This is not to saythat everyadditional choice makesa person's well-being go up,
nor that the obligation to choose necessarily adds to one's freedom, These issues will
be taken up in the next chapter.

42 Functioning! and Capability
within reach of a person, and this information is important—no
matter how exactly well-being is characterized.
In either form, the capability approach clearly differs crucially
from the more traditional approaches to individual and social
evaluation, based on such variables as primary goods (as in
Rawlsian evaluative systems), resources (as in Dworkin's social
analysis), or real income (as in the analyses focusing on the GNP,
GDP, named-goods vectors10
). These variables are all concerned
with the instruments of achieving well-being and other objectives,
and can be seen also as the means to freedom. In contrast, function-
ings belong to the constitutive elements of well-being. Capability
reflects freedom to pursue these constitutive elements, and may even
have—as discussed earlier in this section—a direct role in well-
being itself, in so far as deciding and choosing are also parts of
living.''
3.2, VALUE OBIECTS AND EVALUATIVE SPACES
In an evaluative exercise, two distinct questions have to be clearly
distinguished: (1) What are the objects of value? (2) How valuable
are the respective objects? Even thoughformally the former question
is an elementary aspect of the latter (in the sense that the objects of
value are those that havepositive weights), nevertheless the identifi-
cation of the objects of value is substantively the primary exercise
which makes it possible to pursue the second question.
Furthermore, the exercise of identification of the set of value-
objects, with positive weights, itself precipitates a 'dominance
10
For a critical survey of the analytical literature on this, see Sen (1979e).
" A ftiller discussion of the motivational and strategic issues underlying the
capability approach can be found in Sen (I980a, I985a, 19856,19916). For critiques
and extensions of that approach, and also contributions in related but different
traditions, see Roeraer (1982,1986o), Strccten (1984), Beitz (1986), Culyer (1986), P.
Dasgupta (1986, 1988, 1990), de Beus (1986), De Leonardo, Mauris, and Rotelli
(1986), Delbono (1986), Hamlin (1986), Helm (1986), Kakwani (1986), Luker
(1986), O'Neill (1986, 1992), Riley (1986, 1987), Zamagni (1986), Asahi (1987), K.
Basu (1987a), Brannen and Wilson (1987), Erikson and Aberg (1987), Hawthorn
(1987), K. Hart (1987), Kanbur (1987), Kumar (1987), Mueltbauer (1987),
Ringen (1987), B, Williams (1987), Wilson (1987), Gaertner (1988, 1991), Goodin
(1988), Arneson (I989a, 1990*), G. A. Cohen (1989, 1990), Dreze and Sen (1989), K.
Griffin and Knight (1989), Nussbautn (1988a, 1988ft), Suzumura (1988), Stewart
(1988), Pogge (1989), Seabright (1989, 1991), Desai (1990), Hossain (1990), Sterner
(1990), Van Parijs(1990a, 19906), Ahtisaari (1991), D. A. Crocker (1991«, 19914), A.
Williams (1991), Bliss (1992), Brock (1992), A K, S. Kumar (1992) among others.

Functionings and Capability 43
ranking' (x higher than ,y if it yields more of at least one of the valued
objects and at least as much of each). This dominance ranking,
which can be shown to have standard regularity properties such as
transitivity, can indeed take us some distance—often quite a dis-
tance—in the evaluative exercise,'2
The identification of the objects of value specifies what may be
called an evaluative space (briefly discussed in Chapter 1). For
example, in standard utilitarian analysis, the evaluative space con-
sists of the individual utilities (defined in the usual terms of pleas-
ures, happiness, or desire-fulfilment). Indeed, a complete evaluative
approach entailsa class of'informational constraints*in the form of
ruling out direct, evaluative use of various types of information
(those that do not belong to the evaluative space).13
The capability approach is concerned primarily with the identifi-
cation of value-objects, and sees the evaluative space in terms of
functionings and capabilities to function. This is, of course, itself a
deeply evaluative exercise, but answering question (1), on the identi-
fication of the objects of value does not, on its own, yield a particu-
lar answer to question (2), regarding their relative values. But the
selection of space can also have a good deal of discriminating
power, both because of what it includes as potentially valuable and
because of what it excludes from the list of objects to be weighted as
intrinsically important.
For example, the capability approach differs from utilitarian
evaluation (more generally 'welfarist* evaluation14
) in making
room for a variety of doings and beings as important in themselves
(not just because they may yield utility, nor just to the extent that
they yield utility),15
In this sense, the perspective of capabilities
12
On this sec Sen (I970o, 1970ft).
13
An evaluative system can,in fact, be helpfully analysed in terms of theinforma-
tional constraints that it entails—the types of information that it 'rules out' from
being used. On this strategy of "informational analysis of evaluativeprinciples', see
Sen (19771), 1979$.
14
Utilitarianismmay be factorized into three distinct components, viz. (1) const-
quenlialism (decision variables such as acts, rules, etc.,must bejudged by the good-
ness of the consequent states of affairs); (2) welfarism (states of affairs must be judged
by the individual utilities in that state); and (3) sum-ranking(individual utilities must
be judged by simply summing them). On the nature of the factorization, and the
variantsof utilitarianism within thisgeneral structure, seeSen (1979a, i919b) and Sen
and Williams (1982), 'Introduction',
15
There are various ways of deining utility (such as happiness, pleasure, or desire-
fulfiknent) in distinct versions of utilitarianism (see e.g. Gosling 1969). But the
remark here applies to all of them. On the other hand, if 'utility' is defined, as

provides a fuller recognition of the varietyof ways in which lives can
be enriched or impoverished. It also differs from those approaches
that base the evaluation on objects that are not, in any sense, per-
sonal functionings or capabilities, e.g. judging well-being by real
income, wealth, opulence, resources, liberties, or primary goods.
3.3, SELECTION ANDWEIGHTING
There are always elements of real choice regarding the functionings
to be included in the list of relevant functionings and important
capabilities. The general format of 'doings' and 'beings' permits
additional 'achievements* to be defined and included. Some func-
tionings may be easy to describe, but of no great interest in most
contexts (e.g. using a particularwashing powder-—much like other
washing powders16
). There is no escape from the problem of evalu-
ation in selecting a class of functionings—and in the corresponding
description of capabilities. The focus has to be related to the under-
lying concerns and values, in terms of which some definable func-
tionings may be important and others quite trivial and negligible."
The need for selection and discrimination is neither an embarrass-
ment, nor a uniquedifficulty, for the conceptualization of function-
ings and capabilities.'*
In the context of some types of welfare analysis, e.g. in dealing
with extreme poverty in developingeconomies, wemay be able to go
a fairly long distance in terms of a relatively small number of
centrally important functionings (and the corresponding basic
James Griffin (1986) puts it, 'not as a substantive value at all,but instead as a forma!
analysis of what it is for something to be prudentially valuable to some person*
(pp. 31-2), then the whole issue will turn on precisely how that 'formal analysis' is
conducted. Whether the particular reirtterpretetiort of utilitarianism proposed by
Griffin leaves the approach still within the utilitarian fold, in any clearly distinctive
sense, isan issue that I do not address here. Nor do I examinehere the question of the
correspondencebetween Griffin's general strategy as outlinedin the quoted statement
above and his particular use of the strength of prudential desire in substantive ethical
analyses.
16
Bernard Williams (1987) raises this issue in his comments on my Tanner
Lectures on the standard of living; on this, see his discussion (pp.98-101), and ray
response (pp.108-9), in Sen el al. (1987).
17
On the need to relate the evaluational concerns to the underlying motivation,
see Brock (1991), who discusses this general question helpfully in terms of concrete
issues to health care, See also Befeiiie (19830), Verba et'al. (1987), D. A, Crocker
(19916).
18
I have tried to discuss some of the general methodological issues involved in
'description as choice' in Sen (1980ft).

capabilities, e.g. the ability to be well-nourished and well-shel-
tered, the capability of escaping avoidable morbidity and premature
mortality, and so forth).19
In other contexts, includingmore general
problems of economic development, the list may have to be much
longer and much more diverse.20
In his review article of an earlier work of mine, Charles Beitz
(1986) has illuminatingly discussed various features of the
capability approach and has also forcefully raised an important
critical issue (one that has been aired in different forms by several
other critics as well):
The chief theoretical difficulty in the capabilities approach to interpersonal
comparisons arises from the obvious fact that not all capabilities stand on
the same footing. The capacity to move about, for example, has a different
significance than the capability to play basketball.21
This is a natural worry to face, and it is important that the question
be posed and addressed. It is certainly clear that some types of
capabilities, broadly conceived, are of little interest or importance,
and eventhe ones that count have to be weighted vis-a-viseachother.
But these discriminations constitute an integral part of the
capability approach, and the need for selection and weighting
cannot really be, in any sense, an embarrassment (as 'a theoretical
difficulty').22
The varyingimportance of different capabilities is as much a part
of the capability framework as the varying value of different com-
19
See Sen (1984, 19886). The term 'basic capabilities', which I had used in
'Equality of What?' (Sen 1980a), was intended to separateout the ability to satisfy
certain elementary and crucially important functionings up to certain levels. The
term can,of course, be plausibly used in other ways as well, given the ambiguity of
the concept of basieness, e.g.in the sense of referring to a person'spotential capabili-
ties that could be developed, whether or not they are actually realized (this is the
sense in which the terra is used by Nussbaurn 1988fl, 19S8&).
20
The range of functionings and capabilitiesthatmay be ofinterest for the assess-
ment of a person'swell-being or agencyachievementscan bevery wideiadeed; onthis
see my Dewey Lectures (Sen 198S0).
Jl
Beitz(1986:287). See alsoAmeson (1989«, 19906)and G. A.Cohen (1989,1990).
22
The Aristotelian approach,referred to earlier, suggests a more assertive resolu-
tion of the weighting problem in the form of an ordered list of distinct functioning!
and capabilities, even though the ways of achieving specific capabilities may vary; on
this see Nussbaum (1988a, 1988ft). On the merits of the general approachof what is
sometimes called 'perfectionism', seeHaksar (1979). In general, the weighting prob-
lem may not be a$ totally and fully resolved as the Aristotelian approach suggests,
but it is possible to make fruitful use of the capability approacheven without such a
linear ordering.

modities is a part of the real-income framework. Equal valuation of
all constitutive elements is needed for neither. We cannot criticize
the commodity-centred evaluation on the ground that different com-
modities are weighted differently. Exactly the same applies to func-
tionings and capabilities. The capability approach begins with
identifying a relevant space for evaluation, rather than arguing that
everything that can be put into the format of that space must, for that
reason, be important—not to mention, equally significant.
The primary claim is that in evaluating well-being, the value-
objects are the functionings and capabilities. That claim neither
entails that all types of capabilities are equally valuable, nor in-
dicates that any capability whatsoever—even if totally remote from
the person's life—must have some value in assessing that person's
well-being. It is in asserting the need to examine the value of func-
tionings and capabilities as opposed to confining attention to the
means to these achievements and freedoms (such as resources or
primary goods or incomes) that the capability approach has some-
thing to offer. The relative valuation of different functionings and
capabilities has to be an integral part of the exercise.
3.4. INCOMPLETENESS: FUNDAMENTAL AND PRAGMATIC
I move now to a different, but related, issue. The capability
approach can often yield definite answers even when there is no
complete agreement on the relative weights to be attached to
different functionings. First, a particular selection of value-objects
(in this case, the functionings and capabilities that are accepted as
valuable) would yield a 'dominance partial order* even without
specification of relative weights. Having more of each relevant
functioning or capability is a clear improvement, and this is decid-
able without waiting to get agreement on the relative weights to be
attached to the different functionings and capabilities.
More importantly, that dominance-partial ordering can be
extended even without a full agreement on relative values. For
example, if there are four conflicting views claiming respectively
thattherelativewetghttobeattachedtox v/s-a-vu^shouldbe Vi, */j, M,
and B, there is, then, an implicit agreement that the relativeweight
on x should not exceed V4, nor fall below M. But even this agreement
will, in general, permit us to order pairs-—possibly many pairs—not
covered by the dominance ranking. For example, with the weights

FIG. 3,1. Dominance and Intersection
specified, having one unit of x and two of y would be clearly better
than having two units of x and one of y (even though neither pair
dominates the other in the sense of having more of each x and y).
The 'intersection approach', which articulates only those judge-
ments that are shared implications of all of the possible alternative
weights, can indeed take us quite a distance.2
* It makes no further
M
On the mathematical and interpretations! issues underlying the intersection
approach, see Sen (19?0«, 1970ft, 913a, 198SI>, 1986«), Blackorby (1975), Fine
(197Stf), K. Basu (1979), Be»mbinder and van Acker (1979, 1986). See also the
different but rootivationaily related literature on 'fuzzy* sets and measures (e.g.
Zadeh 1965; Gouguen 1967; Bellman and Zadeh 1970), with possible application to
measuring inequality and related social variables (e.g. S. R. Chakravarty and Roy
1985; K. Basu 19876).

demands on agreement than what already exists. In Figure 3.1, the
axes represent the value-objects (e.g. relevant functionings). That
agreement (viz. on the identification of value-objects) already pre-
cipitates a dominance ranking, e.g. a is superior to b. But the dom-
inance ranking is incomplete and cannot rank a vis-a-vis c, d, or e.
Now consider the different 'indifference curves' (more generally,
surfaces, when there are more than two value-objects) that are
regarded as possible, without it being clear which one gives the
correct valuation (or even that there is only one correct valuation).
Let the permissible indifference curves through a be I, II, and III.
Since a lies below c according to all of them, the intersection
approach declares a to be inferior to c. Similarly,since a lies above d
according to all of them, a can be declared to be better than d. Thus
the dominance-partial ordering is extended by the intersection
method. Even after this extension, the result may be still a partial
order, illustrated by the fact that a is above e according to some
indifference curve and below it according to another, so that a and e
cannot be ranked in this case. The intersection approach increases
decisiveness and articulation, but does not eliminate undecidability.
That residual undeeidability—when present—would not be a
reason for embarrassment, since all it does isreflect the fact that with
partially dissonant valuations, agreed completeness cannot be
achieved.
It is important not to see the use of the capability approach as an
'all or nothing* exercise. Indeed, the nature of interpersonal com-
parisons of well-beingas well as the task of inequalityevaluation as
a discipline may admit incompleteness as a regular part of the
respective exercises. An approach that can rank the well-being of
every person against that of every other in a straightforward way, or
one that can compare inequalities without any room for ambiguity
or incompleteness, may well be at odds with the nature of these
ideas. Both well-being and inequality are broad and partly opaque
concepts. Trying to reflect them in the form of totally complete and
clear-cut ordering? can do less than justice to the nature of these
concepts. There is a real danger of overprecision here.
In so far as there is genuine incompleteness, disparity, or ambi-
valence in relative weights, they should be reflected in correspond-
ing ambiguitiesin the characterization of the weighted value of well-
being. This relates to a methodological point, which I have tried to
defend elsewhere, that if an underlying idea has an essential

ambiguity, aprecise formulation of that idea must try to capture that
ambiguity rather than lose it.24
The use of partial ordering has two different types of justification
in interpersonal comparison or in inequality evaluation. First, as
has been just discussed, the ideas of well-beingand inequality may
have enough ambiguity and fuzziness to make it & mistake to look
for a complete ordering of either. This may be called the 'funda-
mental reason for incompleteness'. Second, even if it isnot a mistake
to look for one complete ordering, wemay not be able in practice to
identify it. While there may be disagreements about parts of that
ordering and disputes as to how we should deal with those parts,
there could still be agreement on other parts. The 'pragmatic reason
for incompleteness* is to use whatever parts of the ranking we
manage to sort out unambiguously, rather than maintaining com-
plete silence until everything has been sorted out and the world
shines in dazzling clarity.
The pragmatic course can, of course, work sequentially, and it is
possible to extend partial orderings as and when we sort out the
unresolved parts. But despite acknowledging the possibility of ex-
tension, it may be a mistake to say nothing and make no judgements
whatever even about those parts that are clear, until everything is
resolved. 'Waiting for toto* may not be a cunning strategy in a
practical exercise.
3.5. CAPABILITY OR FUNCTIONINGS?
Capability is primarily a reflection of the freedom to achieve valu-
able functionings. It concentrates directly on freedom as such rather
than on the means to achieve freedom, and it identifies the real
alternatives we have. In this sense it can be read as a reflection of
substantive freedom. In so far as functionings are constitutive of
well-being, capability represents a person's freedom to achieve
well-being.
That connection may seem simple enough, but—as was discussed
in the first section of this chapter—capability may be relevant even
for the level of well-being achieved, and not only for thefreedom to
achieve well-being. The achievement of well-being is not indepen-
24
On this see Sen (19700, 1989ft). This is, of course, not a special problemexclu-
sively for the capabilityapproach.The same issues arise generally in many of the
conceptual frameworks in social, economic, and political theory.

50 Functioningsand Capability
dent of the process through which we achieve variousfunctionings
and the part that our own decisions play in those choices. If this line
of reasoning is seen to be important over a wide domain, there will
be a case for relating capability to achieved well-being—and not
only to the freedom to achieve well-being,
However, this move might seem rather complicated and perhaps
even confusing. For one thing, it would seem to disrupt the neat
structure of (I) functionings achieved being related to the achieve-
ment of well-being, and (2) the capability to function being related to
the freedom to achieve well-being. While—it could be argued—
freedom of choice (and therefore capability) might indeed have some
direct influence on the level of well-being achieved, nevertheless it
would surely be an 'overkill' to see achieved well-being 'just as' a
function of the capability set, rather than of the actual functionings
achieved. In trying to make room for one little influence, aren't we
likely to miss out the simple and important relationship between
achieved functionings and achieved well-being?
In sorting out this tangled issue, the first thing to note is that
capability is defined in terms of the same focal variables as func-
tionings. In the space of functionings, '&ny point represents an «tuple
of functionings. Capability is a set of such functioning « tuples,
representing the various alternative combinations of functionings
from which the person can choose one combination.25
Since an
important part of the force of the capability approach lies inmoving
us awayfrom the space of commodities, incomes, utilities, etc., on to
the space of the constitutive elements of living, it is particularly
important to note that there is no difference asfar as the space is
concerned between focusing on functionings or on capabilities. A
functioning combination is a point in such a space, whereas
capability is a set of such points.
Next, we must note that the capability set contains l«?er alia in-
formation about the actual functioning combination chosen, since it
too is obviously among the feasible combinations. In fact, there is
nothing to prevent us from basing the evaluation of a capability set
exactly on the assessment of the chosen combination of functionings
from that set. Indeed, if freedom had only instrumental importance
for a person's well-being and no intrinsic relevance, then it would be
appropriate—in the evaluation of well-being—to identify the value
of the capability set simply with the value of the chosenfunctioning
25
For formal characterizations, see Sen (19856: chs. 2 and 7).

combination. This will coincide with valuing a capability set by the
value of the best element (or, more generally, a best element, since it
need not be unique) of this set, if the person does in fact choose in a
way that maximizes his or her well-being. These different pro-
cedures, which may or may not yield the same result, can be seen as
examples of what may be called 'elementary evaluation,* i.e. valu-
ing a set by the value of one distinguished element of it (e.g. the
chosen one, the best one, or some such),26
The possibility of elementary evaluation of the capability set
makes it clear that even if we are ultimately concerned only with
achievements, not with freedoms (except instrutnentally as means to
the achievements), the capability set can, in fact, still be used for the
evaluation. The capability set gives us more information than we
need, but the chosen combination is Apart of the capability set.27
In
this sense, there is a case for seeing the theory of evaluation of
well-being in terms of capability in general, whether or not the
elements other than the chosen one are actually invoked (depending
on the importance that is attached to the process of choice}.28
Furthermore, freedom of choice can indeed be of direct impor-
tance for the person's quality of life and well-being. The nature of
this connection may be worth discussing a bit more. Acting freely
and being able to choose are, in this view, directly conducive to
well-being, not just because more freedom makes more alternatives
available. This view is, of course, contrary to the one typically
assumed in standard consumer theory, in which the contribution of a
set of feasible choices is judged exclusively by the value of the best
element available.29
Even the removal of all elements of a feasible
set (e.g. a 'budget set') other than the chosen best elementwould be
seen as 'no real disadvantage', since the freedom to choose does not,
» SeeSen(19856: 60-1).
27
To use this procedure, wedo, of course, need to know what ischosen from each
set, and not just what the set is from which thechoiceis being made.Thiscan be done
through actual observation, or through some behavioural assumption (such as the
maximization of the relevant objective function).
28
In fact, in my first presentation of the capability viewin 'Equality of What?" (Sen
1980a), no real distinction was made between the capability view and the functioning
viewof well-being.On thisissue, s«ealso Cohen (1990), Desai (1990), Ahtisaari(1991).
19
Thus in standard consumer theory set evaluation will take the form of ele-
mentary evaluation. For particular departures from that tradition, see Koopmans
(1964) and Kreps (1979). However,even there-the motivation is not so much to value
the freedom one has as a good thing in itself, but to takenote of uncertaintyregarding
one's ownfuture preferences by instrumentally valuing the advantageof havingmore
options in the future, On the motivational distinctions, see Sen (19856),

52 Functioning! and Capability
in this view, in itself matter. In contrast, if choosing is seen as a part
of living, and 'doing x' is distinguished from 'choosing to do x and
doing it", then even the achievement of well-being must be seen as
being influenced by the freedom relected in the capability set.
It is, in fact, possible to represent functionings in such a way that
they already reflect the alternatives available and thus the choices
enjoyed. For example, Tasting' as a functioning isnotjust starving; it
is choosing to starve when one doeshaveother options. In examining a
starving person's achieved well-being, it is of direct interest to know
whether he is fasting or simply does not have the means to get
enough food. Similarly, choosing a life-style is not exactly the same
as having that life-style no matter how chosen, and one'swell-being
does depend on how that life-style happened to emerge.30
There is, in principle, some real advantage in being able to relate
the analysis of achieved well-being on the wider informational base
of the person's capability set, rather than just on the selected element
of it. This is, however, not to deny that quite often this potential
advantage would have to be forgone given the difficulty of getting
information regarding the capability set as opposed to the observed
functionings. In fact, the capability set is not directly observable,
and has to be constructed on the basis of presumptions (just as the
'budget set' in consumer analysis is also so constructed on the basis
of data regarding income, prices and the presumed possibilities of
exchange). Thus, in practice, one might have to settle often enough
for relating well-being to the achieved—and observed—function-
ings, rather than trying to bring in the capability set (when the
presumptive basis of such a construction would be empirically
dubious).31
But we must distinguish between what becomes acceptable on
grounds of practical difficulties of data availability, and what
would be the right procedure had one not been so limited in terms of
information. In arguing for the importance of the capability set in
39
The importance of choice in the value of living has been emphasized by several
authors, includingAristotle (Nieomachean Ethics,books 11 and V; in Ross 1980) and
Marx (1844, 1845-6). There are many formal as well as conceptual problems in-
volved in valuing both achievements and freedoms, and in makingroom for each in
the evaluative structure. I have tried to discuss these problems elsewhere, and will
not pursue them further here; in Sen (1985<j, I991A).
Jl
Indeed, as will be argued hi Ch. 4, sometimes even the analysis of freedom, and
not just of achieved well-being, may have to be partly based on the observed states of
being (bringing in the perspective of freedom in terms of the power to get what one
would choose, rather than focusingjust on the act of choice).

Functionings and Capability S3
the analysis of achieved well-being,we are not closing our eyes to
the practical problems of informational availability, nor to the
value of the second-best analysis that we can do even with limited
data. But it is also important to be quite clear as to what data, in
principle, can be relevant and useful, even though in many cases we
might not be able to get them. Practical compromises have to be
based with an eye both to (1) the range of our ultimate interests, and
(2) the contingent circumstances of informationalavailability.
Even when the pragmatic acceptance of limitations of data
availability force us to set our sights lower than the full representa-
tion of capability sets, it isimportant to keep the underlyingmotiva-
tions clearly in view and to see practical compromises as the best we
can do under the circumstances. As a matter of fact, the informa-
tional base of functionings is still a much finer basis of evaluation of
the quality of life and economic progress than various alternatives
more commonly recommended, such as individual utilities or com-
modity holdings.
The capability approach can, thus, be used at various levels of
sophistication. How far wecan go would depend much on the prac-
tical considerations regarding what data we can get and what we
cannot. Ideally, the capability approach should take note of the
full extent of freedom to choose between different functioning
bundles, but limits of practicality may often force the analysis to be
confined to examining the achieved functioning bundle only. This is
obviously more of a problem when we use the capability approach
to msses&freedom rather than the actual well-being attained, but even
for the latter, data limitation can be—for reasons already men-
tioned—a substantial drawback.
3.6. UTILITY VIS-A-VIS CAPABILITY
We end this chapter with a brief contrast of the capability approach
with the alternative of relying on utility as a guide to personal
well-being and as the basis of social ethics and the assessment of
equality. The utilitarian notion of value, which is invoked explicitly
or by implication in much of welfare economics, sees value, ulti-
mately, only in individual utility, which is defined in terms of some
mental condition, such as pleasure, happiness, desires.
The 'desire-fulfilment* interpretation of utility is sometimes seen
as quite distinct from a 'mental state' view, on the ground that here

utility is achieved through the objective realization of a desired state
rather than through achievingsome mental state like that of being
pleased (see J. Griffin, 1982,1986). The distinctionis indeed impor-
tant. It is also certainly true that no mental metric is, in fact,
involved in determining the existence of some utility in the sense of
desire-fulfilment—-all we need to check is whether the desired
object has or has not been achieved. However, for a fuller welfarist
evaluation, more is needed than just ascertaining the existence of
utility, and specifically it requires measurement and comparison of
utilities, in some form or other. For this purpose, intensities of desire
would have to be compared, if the approach has to be based on
relating the importance of desire-fulfilment to the strength of the
desire. Indeed, the mental metric of desire as such would have to be
extensively invoked to make use of the desire-fulfilment view of
utility in utilitarian and other utility-basedevaluations,
It might appear that there is no particular reason why the valu-
ation of functionings or capabilities themselves should not be done
by the use of mental metrics, e.g. the strength of desires, and if so
done, why such a utility-based accounting could not be seen as a
possible part of the capability approach. There is indeed no reason
why such desire-based accounting of capabilities and functionings
could not count as a specific version of the capability approach
broadly defined. But there are problems related to motivational
contrasts.
There are different problems with different interpretations of
utility, but they share the programme of getting the evaluation done
indirectly through using some psychological metric like happiness
or desire. This is precisely where the main difficulty lies. White
being happy may count as an important functioning, it cannot
really be taken to be all there is to leading a life (i.e. it can scarcely
be the only valuable functioning). If the utility-based valuation is
done in terms of pleasure or happiness, then in effect the other
functionings would get disenfranchised, and would be valued only
indirectly and only to the extent that they contribute to pleasure or
happiness.
If, on the other hand, desire-fulfilment is taken as the criterion,
then a very particular method of evaluating capabilities and func-
tionings would have been chosen. The adequacy of this particular
perspective for the evaluation of capabilities and functionings is
deeply disputable, since any mechanical use of a metric of desires

rather than facing the problem of reasoned assessment does injustice
to the exercise of normative evaluation,32
The problem is particularly acute in the context of entrenched
inequalities and deprivations, A thoroughly deprived person, lead-
ing a very reduced life, might not appear to be badly off in terms of
the mental metric of desire and its fulfilment, if the hardship is
accepted with non-grumbling resignation. In situations of long-
standing deprivation, the victims do not go on grieving and
lamenting all the time, and very often make great efforts to take
pleasure in small mercies and to cut down personal desires to
modest—'realistic'—proportions. Indeed, in situations of adversity
which the victims cannot individually change, prudential reasoning
would suggest that the victims should concentrate their desires on
those limited things that they can possibly achieve, rather than fruit-
lessly pining for what is unattainable. The extent of a person's
deprivation, then, may not at all show up in the metric of desire-
fulfilment, even though he or she may be quite unable to be ad-
equately nourished, decently clothed, minimally educated, and
properly sheltered.
The problem of entrenched deprivation is particularly serious in
many cases of inequality. It applies particularly to the differenti-
ation of class, community, caste, and gender. While the nature of
these deprivations can be brought out more clearly by concentrating
on socially generated differences in important capabilities, some of
that gain would be wasted if the capabilities themselves were to be
assessed, after all, in the metric of utilities. A return to the old
eonformism as a supplement to the capability perspective would
tend to wipe out—at least partially—the gains from the change,
especially by undervaluing those capabilities which the chronically
deprived dare not covet, The exercise of evaluation of capabilities
cannot be left to the mere totalling of utilities generated by these
capabilities. The difference can be quite far-reaching in the case of
deep-rooted and resilient inequalities.
32
This is a difficult issue, and the summary of a complex argument presented here
is not satisfactory. For » fuller treatment of this question, see my second Dewey
Lecture (Sen 1985«: 185-203). On related matters, seealso Davidson (1986), Oibbard
(1986), and Seanten (1975, J990, 1992).

4
FREEDOM, AGENCY AND
WELL-BEING
4.1, WELL-BEING V1S-A-VIS AGENCY
In the discussion of freedom in the last chapter, our attention was
confined to the freedom to achieve functioning® relevant for one's
own well-being. But a person can—and typically does—also have
goals and values other than the pursuit of one's own well-being. In
the Dewey Lectures (Sen 1985a), I have tried to discuss the distinc-
tion between 'the agency aspect* and 'the well-being aspect' of a
person, and have argued that a person cannot be reduced tojust one
dimension to force these two aspects to coincide.
A person's agency achievement refers to the realization of goals
and values she has reasons to pursue, whether or not they are con-
nected with her own well-being. A person as an agent need not be
guided only by her own well-being, and agency achievement refers
to the person's success in the pursuit of the totality of her considered
goals and objectives. * If a person aims at, say, the independence of
her country, or the prosperity of her community, or some such
general goal, her agency achievement would involve evaluation of
states of affairs in the light of those objects, and not merely in the
light of the extent to which those achievementswould contribute to
her own well-being.
Corresponding to the distinction between agency achievement and
well-being achievement, there is a differentiation also between a
1
The need for reflection, or for havingreasons for one's goals and objectives, is a
qualification that can be quite exacting. Any whim or caprice that a person happens to
have at a certain time need not provide the basis of an 'agency objective* in the sense
used here. On the relevance of thiscoostraint, see Sen(1985a). On some general issues
regarding reasoned defenceof one's objectives, seeRawls (1971), Glover (1977), Hare
(1981), B. Williams (1981, 1985), Hirsehman (1982), Schelling (1984), Parfit (1984),
Nagel (1986), Wiggins (1987), Hurley (1989), among others. In the context of eco-
nomic analysis, the often neglected need for 'rational assessment' has been well
posed by Broome (1978). In this monograph I shall not specifically concentrate on
this aspect of the problem, and will simply assume that a person's 'objectives' and
'goals' refer to those objectivesand goals for whichhe has reasons (rather than refer-
ring to any impulse or whim that he happens to have).

Freedom, Agency and Welt-Being 57
person's 'agency freedom* and 'well-being freedom*. The former is
one's freedom to bringabout the achievements one values and which
one attempts to produce, while the latter is one's freedom to achieve
those things that are constitutive of one's well-being. It is the latter
that is best reflected by a person's capability set, for reasons already
discussed, while the former—agency freedom—would have to be
seen in broader terms, including aspects of states of affairs that
relate to one's agency objectives (whether or not they directly con-
tribute to one's well-being). I have discussed the distinction more
extensively elsewhere,2
and will not pursue it further here.
It should be emphasized that there is no claim here that the two
aspects—agency and well-being—would be independent of each
other. They are distinguishable and separate, but thoroughly inter-
dependent. The pursuit of well-being can be one of the important
goals of the agent. Also, the failure to achieve «o«-well-being goals
can lead to frustration and thus to a loss of well-being. There are
other connections. The point is the recognition of a significant dis-
tinction, not the assertion of any possibility of analysing one in-
dependently of the other.
4,2, AGENCY, INSTRUMENTALITY AND REALIZATION
In analysing the agency objectives, it is possible to make a further
distinction between (1)the occurrence of those things that one values
and one aims at achieving, and (2) the occurrence of such things
brought about by one's .own efforts (or, in the bringing about of
which one has oneselfplayed an active part).3
The former refers to
the realization of one's objectives, regardless of one's own role in
bringing about that realization. The latter, in contrast, is a more
specific notion of agency success, concentrating on one's success
specifically as an agent.4
If my agency objectives include the independence of my country,
or the elimination of famines, the first view of agency achievement
2
In my third Dewey Lecture (Sen 19850).
3
In my Dewey Lectures (Sen 1985o), my primary focus was on the distinction
between the 'well-being aspect* and the 'agency aspect*, and as a consequence, I did
not distinguish between the two different features of the agency aspect sharply. The
motivational discussion included both types of agency considerations. I am most
grateful to Susan Brison for an entirely persuasive discussion on the importanceof
disentangling the two elements,
4
The valuethat people attach (and have reason to attach) to 'involvement' is one
of the important features of this distinction.Oa this see Hirschman (1982).

58 Freedom, Agency and Welt-Being
would be well met if the country does become independent, or if
famines are in fact eliminated, irrespective of the part I personally
manage to play in bringing about that achievement. This idea of
agency success is based on a straightforwardcomparison between
the objectives 1wish to promote—or what I would in fact promote
if I were the effective agent—and the actual realization of those
objectives (no matter who does the actual promoting),
The second view is a more participatory one. I have to look
specifically at my own role in the promotion of those objectives. In
this view, my agency success, in this narrower sense, would depend
precisely on the role I myself play in bringing about the achieve-
ment of those objectives. We can perhaps usefully distinguish
between 'realized agency success' and 'instrumental agency success'
to refer respectively to the two cases,
In some contexts this distinction between the two elements in a
person's agency aspect can be important. To some extent the ques-
tion is closely related to the nature of our values, i.e. whether what
we value is the achievement irrespective of the instrumental pro-
cess, or whether the valuation relates directly to the part we our-
selves play in bringing about the results. Indeed, by specifying the
objectives more fully, e.g. by distinguishing between 'the occur-
rence of A' and 'the occurrence of A through our own efforts', it is
formally possible to embed the particular feature of 'instrumental
agency success' within the general format of 'realized agency
success'. But this is a formal route, which I shall not explore
further here.5
The question of instrumentality relates closely also to the no-
tion of the "control8
that one exercises over the realization of out-
comes. In some views of freedom, definitive and great importance is
attached to a person having the control himself in bringing about
what he wants to achieve. Later on in this chapter, I shall have the
occasion to scrutinize the idea of 'freedom as control' and will
discuss some problems in seeing freedom exclusively in that per-
spective. But before that I would like to investigate further some of
the implications of the general distinction between the agency
aspect and the well-being aspect of a person.
5
The question relates to the waystates of affairs are to be seen—an issue of some
importance in analysing the limits of consequentialism; on this see Sen (19826,
19836).

Freedom, Agency and Well-Being 59
4.3. CAN FREEDOM CONFLICT WITH WELL-BEING?
In arguing for a freedom-based evaluative system, a general pre-
sumption is sometimes made that more freedom is always advan-
tageous—at least not detrimental. Is that supposition correct?6
It
seems clear enough that it cannot be, in general, correct. Indeed,
sometimes more freedom of choice can bemuse and befuddle, and
make one's life more wretched.7
There are costs of decision-taking,
and it may be comfortable to lie back and relax while others make
the detailed choices. It is possible to construct different types of
scenarios in which more freedom makes one certainly less happy,
and possibly even less fulfilled.8
This interesting matter raises two different types of issues. One is
the question as to whether freedom—both agency freedom and well-
being freedom—can conflict with well-being, and if so, in what
sense, and for what reasons? This issue (i.e. possible conflicts
between freedom and well-being) I shall investigate first. But there
is a second—no less important—issue that is also raised by the
possibly contrary effects of the expansion of some types of choices. Is
that contrariness best seen simply as a conflict between freedom, on
the one hand, and advantage, on the other, with some increase in
freedom being disadvantageous! Or is that problem largely a reflec-
tion of an ambiguity regarding the nature of freedom itself which
needs to be sorted out? I shall take up this second issue (related to
the nature of freedom) in the next section, after discussing the poss-
ible conflicts between freedom and well-being.
Given what has already been discussed in this monograph, it
should come as no surprise that freedom and well-being need not
always move in the same way, or even in the same direction. In
pursuing this question, the distinction between well-being freedom
and agency freedom is particularly important. There would be
nothing contrary in the fact that an enhancementof agency freedom
6
For arguments and illustrations showing that more freedom can be disadvan-
tageous, see Elster (1979), G. Dworkin (1982), Schelting (1984), and Frankfurt (1987).
7
Dithering can also kill, as illustrated by the sad case of Buridan's ass,which died
of starvation unable to choose between two haystacks in front of it. A commanding
authority could have saved its life. On the importance of self-command, see Schel-
ling (1984).
8
There is also an interesting issue regarding the intercultural variation of the
importance we attach to taking decisions ourselves and having a range of choice. It
has been argued that in some cultures, more freedom of choice is not viewed with
favour. On related issues, dealing inter alia with cultural variations, see Apffel
Marglin and Margiin (1990).

60 Freedom, Agency and Well-Being
(i.e. an increase in one's ability to promote goals that one has
reasons to promote) can lead to a reduction of well-being freedom
(and correspondingly to a decline in achieved well-being). Indeed, it
is precisely because of such conflicts that the distinction between
agency and well-being is important.*
For example, if instead of being far away from a scene of crime—
a crime that I would like to prevent—I happen to be bang on the
spot, my agency freedom is certainly enhanced (I can now do some-
thing to stop that terrible event which I would much like to pre-
vent), but as a result my well-being may go down (e.g. I may get
wounded in the process of prevention even if my efforts are success-
ful). The fact that, everything considered, I do regard the fight to be
worth it (i.e. my other agency objectives dominate over any dimin-
ution of well-being that I may suffer) does not, of course, entail
that my well-being as such cannot go down as a result of rny
chosen efforts.'0
If freedom is taken to be agency freedom, then it is
quite possible to see contrary movements in which freedom (i.e.
agency freedom) does go up, while achieved well-being goes down.
In fact, not only well-being achievement but also well-being
freedom can often move in the opposite direction to agency freedom,
Indeed, even in the 'crime-prevention case* discussed above, not
only is my well-being reduced as a result of my being at the scene
of the crime rather than far away, there is a possible reduction also
in my freedom to achieve well-being (despite the increase in my
agency freedom). For example, I might not be able to escape get-
ting involved in the event if I am too close (there may be 'no exit'),
so that my increased ability to stop the crime may go hand in hand
with a diminished ability to pursue my own well-being. But per-
haps more importantly, even if I can leave without interference (if I
can choose to "chicken out'), being at that hot spot I mayno longer
be able to be at peace with myself and have the comfort of both
being safe and feeling non-guilty, which I could have enjoyed had I
been far away (not having the need to consciously opt out of help-
* On this, see my third Dewcy Lecture (Sen 1985a: esp. 203-8).
10
A false identification of welt-being success and agency success can, of course,
occur in the model of the 'rational fool', a common behavioural assumption in
economic theory whereby each person is seen as promoting his own well-beingin
every choice that he makes. On the limitations of that behaviouralframework, see Sen
(1973A, I977c, 1987). See also Hirsctanan (1982), McPherson (1982, 1984), Akerlof
(1984), Walsh (1987, 1991), Hausman and McPherson (1991), Weeks (1991).

Freedom, Agency and Well~Being 61
ing the person in peril),11
So this Is a case in which nay well-being
freedom may also go down, and there need not be any conflict
between the directional movements of well-being achievement and
well-being freedom (both down, as opposed to agency freedom and
achievement, which are both up).
Thus, well-being and freedom can move in opposite directions no
matter which interpretation of freedom is chosen. If by freedom we
mean agency freedom, then it is quite possible that an expansion of
(agency) freedom may go hand in hand with a reduction in well-
being freedom as well as well-being achievement. If, on the other
hand, we take freedom to mean well-beingfreedom, then any conflict
between freedom and achieved well-being cannot, of course, arise
from the shrinkage of opportunities of well-being achievement with
increased (well-being) freedom. But there can still be a conflict
between the two (i.e. between well-being freedom and well-being
achievement), since a person's choice is not necessarily guided only
by the pursuit of his or her well-being. If the increase in well-being
freedom is accompanied by other changes that shift one's choices
towards pursuing other—non-well-being—objectives, then well-
being achievement might quite possibly go down while well-being
freedom goes up. This last case may at first sight appear to be a bit
odd (how can an increased freedom to pursue well-being reduce the
achievement of well-being?), but there is no real puzzle here. The
change that brings about the expansion of well-being freedom may
also allow a person to pursue other (non-well-being) goals more
forcefully and this can quite possibly lead to a deterioration of the
extent of well-being the person chooses to achieve.
An illustration may be useful. Consider a doctor who is ready to
sacrifice her own well-being to go and work in some terribly poor
and miserable country, but isprevented from doing that because of a
lack of means and opportunity to go to that far-awayland. Consider
next a rise in her income—it need not matter how this came about—
and in this new economic situation she has both more well-being
11
This kind of problem comes dose to the issue of 'moral luck*, on which see B.
Williams (1973«, 1981, 1985), Nagel (1979), and Nussbaum (1985). Note, however,
that in terms of the duality between well-being and agency explored in this mono-
graph (and in Sen 1985a), a person can be both luckier in the sense of beingbetter able
to promote her agency objectivesand lesslucky at the senseof beingmore constrained
in pursuing her own well-being. In this monograph I do not further explore the
distinctions involved in the general context of 'moral luck'.

62 Freedom, Agency and Weil-Being
freedom (e.g. she can buy lots of nice things for herself) and more
agency freedom (she can also go to the far-away land and sacrifice
her own well-beingfor tireless work in pursuit of the greater good of
suffering humanity). If she chooses the latter, it is quite possible that
she would have less well-being achievement (though greater agency
success). Her well-beingachievement may, thus, go downjust as her
well-being freedom as well as agency freedom both go up.
Freedom and well-being achievement can, thus, move in opposite
directions, no matter whether we interpret freedom as agency
freedom or as well-being freedom. How common such a conflict
will be is, of course, a different question. Greater opportunities of
pursuing well-being—an enhancement of well-being freedom-
may well be frequently seized. Also, as discussed in the earlier
sections of this chapter, increased freedom may itself contribute to
the achievement of well-being, since choosing and deciding may be
valuable parts of livingwell But notwithstanding all this, there can
be a conflict between the achievement of well-being, on the one
hand, and freedom in either form (viz. well-beingfreedom as well as
agency freedom). The reasons for possible conflicts can vary, but
they relate ultimately to the substantive content of the distinction
between the two aspects—well-being and agency- of a person. That
distinction not only indicates two different perspectives in which
inequality may be assessed, but also provides reason, as wehave just
argued, for expecting that well-being may not move in the same
direction as freedom judged in either perspective.
4.4, FREEDOM AND DISADVANTAGEOUS CHOICES
Even when freedom and well-being do not move together, it does
not, of course, follow that an increase in a person's freedom would
be to his or her disadvantage. There are different ways of judging
advantage, related partly to the distinction between well-being and
agency objectives, and partly to the contrast between available
opportunities and resulting states. The realized level of well-being
need not be the only guide to the opportunities that a person values
most. But there is also the second question that was posed earlier
(at the beginning of Section 4.3) concerning the possibility that
increased freedom might be disadvantageous to the person by forcing
on the person the necessity to spend time and effort in making lots of
choices that he or she would rather not have to make. There can be

inconvenience and hassle from having to make more choices, and
this may well be far from advantageous, no matter how exactly we
define advantage.
This question does, in fact, take us in the direction of an important
issue regarding the nature of freedom, to wit,whether any increase in
choices that one can—and has to—make must be seen as an expan-
sion of freedom. This issue is, as will be presently argued, quite
central to the assessment of social structures and public policies
related to enhancing freedom.
Freedom is a complex notion. Facing more alternatives need not
invariably be seen as an expansion of a person's freedom to do the
things she would like to do.n
If a life without hassle is valued (and
the frantic routine of constant decision-taking is not treasured), the
freedom to achieve the preferred form of life is not necessarily
enhanced by a multiplication of trivial choices. What may appar-
ently look like a conflict between freedom and advantage may not,
thus, be anythingof the sort, and may be the result of misspecifying
freedom by overlooking the loss of option of leading a peaceful and
unbothered life (entailed by a proliferation of minor choices that
have to be—willy-nilly—made).
There is an inescapable need for evaluationinjudgingwhether the
forgone opportunity of a hassle-free life may not be more important
than the greater chance of being able to take all these minor de-
cisions. The question really turns on the need to judge what options
are importantand whatare not. The expansion ofchoices to bemade
is both an opportunity (the choices can be made by oneself) and a
burden (the choices have to be made by oneself). It is easy to think of
circumstances when given the choice of having to make theseparticu-
lar choices, one would have good reason to say no. This indicates
that the expansion of those particular choices and obligations need
not be seen as a valued expansion of freedom.
The significant recognition that choosing is a part of living, which
we discussed earlier, cuts both ways. Some types of choosing can be
valuable parts of living,giving us reason to treasure them. But there
are other choices that wemay have no great reason to value, and the
obligatory requirement to face and deal with them may impose on us
losses of time and energy which wemay have good reasons to resent.
12
For different axiomatic approaches to assessing the extents of freedom, see
Koopmans(1964), Kreps(1979),Sen (19Kb, 1991a), Suppes(1987), Pattanaik and XB
(1990),

Thus the expansion of some types of choices can reduce our ability
to choose life-styles that we might treasure.
So the conflict here is not really between our freedom tout court,
on the one hand, and our advantages, on the other, but primarily
between different types of freedoms—the freedom to exercise active
choice over a range of (possibly trivial) options and the freedom to
lead a leisured life without the nuisance of constantly having to
make trivial choices. Some of the perceived conflict between
freedom and advantage arises from an underspecifieation of
freedom—ignoring the fact that we may have good reasons to
choose not to have the necessity of making ail these unwanted
choices. The problem relates to the inescapable requirement of
valuation involved in the assessment of freedom, as has already
been discussed.
4.5, CONTROL AND EFFECTIVE FREEDOM
A related issue concerns the distinction between freedom and direct
control. I have tried to argue elsewhere (Sen 1982e), that the per-
spective of 'freedom as control* is seriously limited. Many
freedoms take the form of our ability to get what we value and
want, without the levers of control being directly operated by us.
The controls are exercised in line with what we value and want (i.e.
in line with our 'counterfaetual decisions'—what we would choose),
and in this sense they give us more power and more freedom to lead
the lives that we would choose to lead. To confuse freedom with
control can drastically reduce the scope and force of that great
idea.
To illustrate, consider first a fairly minor example. When a
proof-reader checks the proofs of this book to weed out the printing
errors and to get the spelling right, he will not be taking away my
freedom to have my book printed the way I would like it to be. The
control will be in other hands, but the proof-reader will be doing
what I would, counterfaetually, have done, if I were to correct all
the proofs myself with eyes as efficient as that of the proof-reader.
Whether he is following my direct instructions, or knows what my
instructions would have been if sought, is not a matter of great
importance here, in judging what freedoms I really do have. As
long as the levers of control are systematically exercised in line
with what I would choose and for that exact reason, my 'effective

freedom* isuncompromised, though my 'freedom as control*may be
limited or absent.'3
In modern society, given the complex nature of social organiz-
ation, it is often very hard, if not impossible, to have a system that
gives each person all the levers of control over her own life. But the
fact that others might exercise control does not imply that there is no
further issue regarding the freedom of the person; it does make a
difference how the controls are, in fact, exercised.
The distinction between 'instrumental agency success' and
'realized agency success* (discussed earlier on in this chapter) is
relevant here. The 'freedom as control*view is concerned only with
the 'instrumental' aspect of the problem. The focus, in this case, is
entirely on the person herself being the active agency in bringing
about the achievement. The broader issue of'realized' success is left
open in this perspective. The idea of'effective freedom' is related to
that broader view of success.
To take a rather more momentous example than getting the spell-
ing right in correcting proofs, the freedom to live in an epidemic-free
atmosphere may be important for us, and given the choice, wewould
choose to achieve that. But the controls of general epidemic prevent-
ion may not be in our hands—it may require national and possibly
even international policies. If we do not have control over the pro-
cess of elimination of epidemics, there is no more to be said, as far as
our 'freedom as control' is concerned, in this field. But in a broader
sense, the issue of freedom is still there. A public policy that elimi-
nates epidemics is enhancing our freedom to lead the life—
unbattered by epidemics—that we would choose to lead.l4
Of course, the elimination of epidemics also increases our well-
being achievement.l5
That is another issue (not unrelated, but not
" What is being called 'effective freedom' here was called 'freedom as power" in
Sen (J982c). G. A. Cohen and Jean Drew have given me good reasons to think that
the word 'power* is not very helpful in makingthis distinction, and that it is hard to
differentiate 'power' from 'control*in ordinary usage. Hencethe attempt here to usea
different term to refer to a person's ability to get systematically what he would choose
no matter who actually controls the levers of operation.
14
If someone does have the odd preferencefor havingone of the epidemic diseases,
then his preference-fulfilment would not be well served by this epidemiologicai
policy. The negative implication may extend, in this case, to freedom as well, if the
preference in question turns out to be reasonably defetidabie,
15
There is, of course, an important question here about the connections between.
the choices of different individuals. In thiscase, it is reasonable to expect a congruence
of our general preferences on the matter of epidemic prevention. It is because of this
congruence that our respective powers to lead lives wewould like to lead do not get

the same either), and the one with which we are directly concerned
here is our freedom to lead a life we would choose to live. The fact
that it would probably also increase our well-being, in this case, is a
different matter. If the levers of control were used by those in charge
of them to, say, promote epidemics, rather thao to eliminate them,
our 'freedom as control' would not be changed (i.e. would continue
to be absent), but our 'effective freedom' (in particular, the freedom
to lead the types of lives we would choose to lead) would be
severely compromised.
In assessing the freedoms that we enjoy and in examining how
unequal we are in that respect, the informational basis of the evalu-
ation has to take into account our counterfactual choices (what we
would choose) and their relation to what is made to happen. An
exclusive concentration on the levers of control would be inade-
quate for analysing our freedoms. While this may look like adding
to the already heavy informational demand for analysing freedoms,
it need not, in fact, make the practical problems of such analyses
more intractable. Sometimes the nature of counterfactual choices
are very easy to guess, e.g. that people would choose to avoid epi-
demics, pestilence, famines, chronic hunger. The elimination of
these unloved things, through public policy aimed at giving people
what they would want, can be seen as an enhancement of people's
real freedom. In this sense, even simple observations of realized
states may have direct relevance to the analysis of freedoms enjoyed,
once we reject relying exclusivelyon the limited view of 'freedom as
control*.
This recognition has some fairly far-reaching relevance to the
nature of empirical analysis of freedoms enjoyed by different people
and of the inequalities in the freedoms that we have. In the next
section, this question is further explored.
4.6. FREEDOM FROM HUNGBE, MALARIA AND OTHERMALADIES
The term 'freedom' is often used to refer to such matters as 'freedom
from hunger* or 'freedom from malaria*. It has sometimes been
suggested that this is a misuse of the concept of freedom, and the
into conflict with each other. The relationship between power and control wouldvary
systematically with the nature of the objectives in question, and that is indeed a
centra! issue in the investigation both of the norms of public policy and of the nature
of'social rationality'. The former issue has been discussed in the literature on public
goods and the allocation of resources in that context.

Freedom, Agency and Welt-Being 67
ideas captured by the expressions have nothing much to do with
freedom in the real sense, '* Not having malaria, or not being hun-
gry, does add to one's well-being, but—so the argument runs—it is
not a matter of having more freedom in any sense.11
This diagnosis can be disputed on the basis of the analysis pre-
sented in the earlier sections of this chapter. In terms of that
analysis, the idea of counter/actual choice—what one would have
chosen if one had the choice—is relevant to one's freedom.l8
Indeed,
one's control over the actual process of choice can scarcely be the
only reference in the ideas of liberty and freedom. When Isaiah
Berlin (1969) talks of 'a man's, or a people's, liberty to choose to
live as they desire' (p. 179), the point of direct reference is the ability
to choose to live as one desires, rather than the mechanism of con-
trol. If people do desire a life without hunger or malaria, the elimin-
ation of these maladies through public policy does enhance their
liberty to choose to live as they desire'. The fact that the levers of
control are not in the hands of most people who are made more free
in this respect does not compromise the importance of the enhance-
ment of their effective freedom to live as they would choose.
However, the rationale of the contrary view (viz, that 'freedom
from malaria*is a misnomer—it is not &freedom as such) need not of
course be based only on concentrating on levers of control. It could
certainly be argued that the elimination of malaria through public
policy does not increasethe range of options that one has, as far as
having or not having malaria isconcerned, (e.g. 'We first had to have
malaria; could not escape it. Now, we don't have to have malaria,
but we cannot, on the other hand, have it, even if we so choose. So
'* The point was forcefully made by several participants in a conference on "qual-
ity of life* arranged at the World Institute of Development Economics Research in
Helsinki in July 1988.
" One must, of course, distinguish between freedom/row malaria as such and the
other freedoms that result from not having malaria (e.g. being able to move about
freely, take on work,etc.). The latter is not in dispute. The question in assessing the
idea of 'freedom from malaria' is whether there is an issue of freedom regarding
malaria itself (i.e. other than the freedom to do various things that are made possible
as a result of the eradication of malaria).
18
The relevance of counterfactual choice in the analysis of personal liberty and
negative freedoms has been extensively discussed in the literature on social-choice
theory. On this and related matters, see Sen (1970«, 1983a), Campbell (1976), Farrell
(1976), Kelly (I976a, 19766,1978), Surumura (1980,1983,1991), Green (1980), Gaert-
ner and Kriiger (1981, 1983), Hammond (1981, 1982, 1985), Basil (1984), Nyrmi
(1984, 1987), Wriglesworth (1985), Elster and Hylland (1986), and Riley (1987,
1989a, 19896), among other contributions(see footnote 5 in Ch. 2, pp. 32-3).

there has been no increase in the freedom to choose?) Indeed, if in
assessing freedom we look only at how many alternatives we have,
then there is no increase in freedom in this respect."
This is precisely where the role of counterfactual choice becomes
relevant—indeed central.20
One values living without malaria, de-
sires such a life, and would have chosen it, given the choice. Being
able to live as one would value, desire and choose is a contribution
to one's freedom (not just to one's well-being or agency achieve-
ment, though it is also that). The fact that the term "freedom* isused
in the expression 'freedom from malaria' is not in itself decisive in
any way, but the relation of the results to what one would have
chosen (and would have had reasons to choose) is a matter of direct
relevance to freedom—the freedom to choose to live as one would
desire.
Indeed, it would be absurd to say we have now achieved freedom
from 'non-malaria'. That is clear enough, but why is it absurd? This
is because non-malaria is not a burden, and we would have no
particular reason to reject it (i.e. to choose malaria instead), if
counterfactually we were given that choice. So the language hap-
pens to be, in fact, in line with the analysis of freedom involving
counterfactual choice. Of course, linguisticoccurrence is not to be
taken as a proof that the counterfactual-choice view of freedom is
correct. Rather, it is the aptness of the counterfactual-choice view,
for reasons already discussed, that shows that the language is not
aberrant here. It fits into a broad generalconcept of freedom, rather
than having to be seen as invoking some peculiarly remote idea of
freedom.
It is not so much that statements such as those of Franklin
Roosevelt on 'the four essential human freedoms* (e.g. 'the third is
freedom from want', "the fourth is freedom from fear') 'prove' the
relevance of counterfactual choice. Rather the converse. Seeing
19
On the axiomatic derivations of that procedure and possible critiques of the
chosen axioms, sec Pattanaik and Xu (1990) and Sen (19900, 199la).
20
Sometimes a bit of fog is added to the idea of freedom by insisting that a person
cannot be said to have the freedom to achieve x if he is not,at the same time, free to
ensure the achievement of not x. Thisis,however, an odd restriction. Ann isfree not to
marry Bill irrespective of what Bill wants, bat that freedom is not conditional on
Ann being able to ensure that she marries Bill (i.e.not 'not marry* Bill) irrespectiveof
what Bill wants. Similarly, Bill may be free to make sure that he dies before
becoming 90 years old (assuming that the anti-suicide laws have limited power), but
that freedom is not conditional on Bill being able to make sure that he does not die
before becoming 90.

freedom, inter alia, in terms of counterfactual choice helps us to
understand and interpret these statements and makes them cogent
(rather than our having to view them as examples of misused—if
moving—rhetoric), Roosevelt was talking about freedom.
Freedom as a value demands that certain things be considered
seriously /or that reason (whether or not it is valued for any other
reason as well). The notion of freedom as effective power to achieve
what one would choose is an important part of the general idea of
freedom.
I end this section with one final observation. I have discussed the
scope of the idea of freedom at some length here partly because
freedom isone of the most powerful social ideas, and its relevance to
the analysis of equality and justice is far-reaching and strong. When
we assess inequalities across the world in being able to avoid pre-
ventable morbidity, or eseapable hunger, or premature mortality,
we are not merely examining differences in well-being, but also in
the basic freedoms that we value and cherish. That viewpoint is
important. And as it happens, the available data regarding the real-
ization of disease, hunger, and early mortality tell us a great deal
about the presence or absence of certain central basic freedoms.
Given the motivation underlying the analysis of inequality, it is
important not to miss this momentous perspective.
4.7. THE RELEVANCE OF WELL-BEING
It ispossible to argue that in an important sense the agency aspect of
a person is more comprehensive than the well-being aspect. A per-
son's goals will, in the case of most 'normal* people, include inter
alia the pursuit of their own well-being. Indeed, the overall balance
of agency objectives might be seen, with some plausibility, as reflec-
ting the weights that the person herself would attach to her own
well-being among the things that she wishes to promote. On this
ground, it could be tempting to argue that in judging the relative
advantages of different people, the proper basis of comparison
should bejust their respective agency freedoms. The information on
agency aspect would inter alia include the value of each person's
own well-being, in the light of whatever importance she herself
gives to it in her agency objectives. Treating the person herself as the
best judge of how shemay beviewed by others, it might look as if the
agency aspect would tell all that is relevant for others to know.

This argument overlooks many things. One issue relates to the
lasting importance of the plurality of our concerns even when for
the purpose of action we choose one compromise or another. The
fact that wemay haveto arbitrate between conflicting demands in the
interest of deciding what we should do need not indicate, especially
when weface deep dilemmas, that this is the end of the story as far as
evaluation is concerned. William Butler Yeats spoke of one such
dilemma:
The intellect of man is forced to choose
Perfection .of the life, or of the work.21
What the person chooses is obviously important, but no matter
which of the two gets chosen by 'the intellect of man', the fact of that
choice does not make the remaining one—life or work—devoid of
importance for that reason.22
The fact that you may have chosen the
'perfection of the work' over that of your 'life*, or have chosen to
give priority to other goals in your agency objectives over your own
well-being is not a reason to think that your life or your well-being
is henceforth of no importance-—to you or to others.21
Another important issue concerns the very different roles that the
welt-being and the agency aspects can have in the use of interper-
sonal comparisons for diverse exercises. Society might accept some
responsibility for a person's well-being, especially when that is in
some danger of being particularly low. But this does not imply that
society must take an equal interest in the promotion of that person's
other agency objectives as well. For example, society may be seen as
having a special responsibility to make sure that no one has to
21
W. B. Yeats, Variorum Edition of the Poemsof W. B, Yeats, ed. P. Allt and R. K.
Alspach (New York: MacmiSlan, 1957), 495, Vivian Walsh, who drew my attention
to these lines, has provided an illuminating and elegant analysis of the scope and
limits of rationality underlying human behaviour and economic choices (see Walsh
1991).
22
On this issue see B, Williams (1973«, 1981), Marcus (1980), and Nussbaum
(1985).
23
Thisis soeven whea the person'schoice isbased on a completeordering in terms
of agencyobjectives. That ordering and the identified 'best'choice reflect yourview of
the relevant 'trade-offs' for that decision, rather than reflecting the unimportance of
the outweighedconsiderations in general, or evenin the specific context of that choice.
But inaddition to that, it must also be recognized thatsometimesour agency decisions
may not be based on a complete ordering (or on the existence of a 'best* alternative),
and the losing side* may not in fact have been "outweighed" by your own reason even
in the specificcontext of that choice. On these issues, seeSen (1970a, 9%2a) and Levi
(1986).

starve, or fail to obtain medical attention for a serious but eminently
treatable ailment. On the other hand, this carries no implication
that the society must take an equally protective attitude about the
person's agency goal of, say, erecting a statue in honour of some
hero he particularlyadmires (even when the person himself attaches
more importance to erecting the statue than to beingwell-nourished
and having medical coverage),
There are, however, other contexts in which the relevant interper-
sonal comparisons may be precisely those of agency aspect—either
of agency achievement or of agency freedom. Wemightwish to know
who has how much power to pursue their own respective goals.24
We
might also have interest in checking how successful they respec-
tively are in bringingabout what they are tryingto achieve. We may
have political and ethical views regarding societies in which some
people can promote all their ends while others have to face great
barriers which they cannot overcome.
Depending on the context, the agency aspect or the well-being
aspect might achieve prominence. It would be a mistake to expect
that one of these aspects would be uniformly more relevant than the
other as a basis of interpersonal comparison for every interesting
exercise. Along with the plurality of spaces (discussed already), we
have to take adequate note of the plurality of purposes. The evalu-
ation of inequality cannot but be purpose-dependent, and the
important need is to provide an appropriate match between (1) the
purposes of inequality evaluation, and (2) the choice of informa-
tional focus. This question will be pursued at a more general level
in the next chapter.
The well-being aspect is especially important in such matters as
social security, poverty alleviation, removal of gross economic in-
equality, and in general in the pursuit of social justice. That case, as
we have discussed, is not conditional on the person himself attach-
ing overwhelmingpriority to his own well-being in his agency objec-
tive. A self-sacrificing idealist who is ready to sacrifice fully his own
well-being for some 'cause' does not thereby make it okay for others
to ignore his well-beingso long as his 'cause' is not harmed. One's
claim to help from others in the society for one's survival and well-
24
The inequality of social power relates to the institutional structureof its organ-
ization and control. For insightful analyses of various aspects of the inequality of
social power and its economic and political implications in the context of an advan-
ced capitalisteconomy, see Galbraith (1952,1958,1967). Seealso Dahrendorf (1988).

being need not be conditional on one's self-eentredness in giving
precedence to one's interests vis-d-vis other goals and values in one's
overall agency objectives.
The well-being aspect of a person has great importance of its own
for the analysis of social inequality and the assessment of public
policy. Problems of social injustice and inequity between different
classes and groups relate strongly to extensive disparities in well-
being—including the freedom that we respectively enjoy to achieve
well-being.25
In some of the analyses that will follow in the rest of
the book, particular attention will be paid to inequalities in well-
being freedom and correspondingly to disparities in well-being
achievement. However, even for that analysis, the relevance of the
agency aspect has to be kept in view, since the person's actual use of
her well-being freedom will depend inter alia on her agency objec-
tives (since these will influence her actual choices).26
Thus, the broader agency concerns remain relevant to the social
or economic analysis of inequality even if the primary focus of the
analysis of inequality is on well-being, or on the freedom to achieve
well-being. The distinction between agency and well-being-
discussed in this chapter—remains, therefore, important even when
our evaluative focus is on inequality of well-being (freedom or
achievement), as will be the case in some of the problems discussed
in the chapters that follow. The specification of our evaluative focus
(to individual well-being) must not be confused with any assumed
narrowing of human beings in terms of motivation.
J$
In anyevaluative structurein which the focus ison thefreedoms that people have
(i.e. in which thefocal variables are freedoms rather than outcomes), no patterningof
outcomes as sach woulddirectly emerge from the equality of the focal variables. The
outcome pattern will depend on the decisions of the persons themselves. This applies,
obviously, to 'well-being freedom" as well. Even when complete equality of well-
being freedom is clearly definable and is entirely realized, this need not lead to an
equality of well-beingachieved, since different individuals may givedifferent priority
to the pursuit of their own well-being (in making use of their respectivewell-being
freedoms).
2
* Also, other people's help to the deprived and the miserable may well relate to
the helpers* agencyobjectives. Thisapplies notjust to charityand other transfers, but
also to political action. For example, it has been observed that a democracy with a
free press tends to eliminate the occurrence of famines through general public pres-
sure on the government to take timely avertive action (threatening its electoral pros-
pects if it fails to do this); on thissee Sen (1984,1988c). Reddy(1988), Dreze and Sen
(1989), Article 19 (1990), D'Souza (1990), Ram (1991), amongothers. The well-being
of the vulnerable famine victims(typically a small minority)may greatly depend on
the agency actions of others, in democratic politics.

5
JUSTICE AND CAPABILITY
5.1. THE INFORMATIONAL BASES OF JUSTICE
Any evaluative judgement depends on the troth of some informa-
tion and is independent of the truth or falsity of others. The 'in-
formational basis of a judgement* identifies the information on
which the judgement is directly dependent and—no less impor-
tantly—asserts that the truth and falsehood of any other type of
information cannot directly influence the correctness of the
judgement.
The informational basis of judgements of justice, thus, specifies
the variables that are directly involved in assessing the justice of
alternative systems or arrangements (the role, if any, of the other
variables being only derivative). For example, in the utilitarian
view of justice, the informational basis consists only of the utilities
of the respective individuals in the states of affairs under evaluation.
I have tried to argue elsewhere that the examination of the informa-
tional basis of each evaluative approach provides a useful way of
investigating and scrutinizing that approach.'
Most theories of justice can also be usefully analysed in terms of
the information used in two different—though interrelated—parts
of the exercise, viz.(1) the selection of relevant personalfeatures, and
(2) the choice of combining characteristics. To illustrate, for the
standard utilitarian theory, the only intrinsically important 'rel-
evant personal features* are individual utilities, and the only usable
'combining characteristic' is summation, yielding the total of those
utilities. The set of 'welfarist' theories, of which utilitarianism is a
particular example, retains the former part (viz. takes utilities as
the only relevant personal features), but can use other combining
characteristics, e.g. utility-based maximin (or lexicographic
1
The diverseroles of theinformational basesof normative choice and judgements
have been discussed in Sen(1974,19776,197941985a). The part played,in particular,
by 'informational constraints', which are typically implicitly imposed, can be both
complex and far-reaching.

74 Justice and Capability
maximin), summation of concave transforms of utilities (such as
summing the logarithms of utilities).2
Examples of selection of 'relevant personal features' other than
utilities include liberties and primary goods (Rawls 1971), rights
(Nozick 1974), resources (R. Dworkin 1981), commodity bundles
(Foley 1967; Pazner and' Schmeidler 1974, Varian 1974, 1975;
Baumol 1986), and various mixed spaces (Suzumura 1983; Wrigles-
worth 1985; Riley 1987). Note that in some cases the personal
features are broadly of the outcome type (e.g. commodity bundles
enjoyed), as they are with welfarist theories (illustrated by utilitari-
anism), whereas in other cases they relate to opportunities, defined
in some way or other (e.g. primary goods, rights, resources).
The selection of personal features has to be supplemented by the
choice of a combining formula, e.g. sum-maximization,3
lexicogra-
phic priorities and maximin,4
equality.,5
or one of various other com-
bining rules.6
The substantive contents of theories of justice have, thus, in-
cluded widely different informational bases and also quite
divergent uses of the respective information. That informational
variation corresponds closely to the question of plurality of focal
variables with which we are concerned in this monograph. As was
argued earlier, each theory of justice includes choosing—explicitly
or by implication—a particular demand for 'basal equality*, which
in its turn influences the choice of the focal variable for assessing
inequality. The respective claims of the different conceptions of
2
Arguments for quite different combiningcharacteristics in even the same utility
space can be found, among others, in Suppes {1966, 1977), Koto (1969, 1976), Sen
(I9?0a, 19776), Mirrtes (1971), Rawls (1971), Phelps (1973), P. J. Hammond
(1976(1), Strasnick (1976), Arrow (1977), Blackorby and Donaldson (1977),
d*Aspremont and Gevers (1977), Maskin (1978), Gevers (1979), Roberts (I980a),
Blackorby, Donaldson, and Weynmrk (1984), d'Aspremont (1985), Thomson and
Varian (1985), While the axioraatics of the combining structuresexplored in these and
related contributions are mostly defined in the utility space, they can, in most cases,
also be readily presented in other spaces, involvingother personal features (such as
indices of primary goods, or of resources, or of capabilities). Thus, the axiomatic
structures, in fact, have a wider interest than the nature of the space might suggest,
' See Harsanyi (1955), d'Aspremont and Gevers (1977), Maskin (1978),
4
See Rawls (1971), Hammond (!97fa), Sirasnick (1976), d'Aspremont and
Gevers (1977), Sen (19776),
1
See Foley (1967), Nozick (1974), R. Dworkin (1981), Van Parijs (19900).
6
SeeVarian (1975), Gevers (1979), Roberts (1980«), Suzumura (1983), Blackorby,
Donaldson, and Weymark (1984), d'Aspremont (1985), WrigSesworth (1985),
Baumol (1986), Riley (1987), Moulin (1989,1990), among manyothercontributions.

Justice and Capability 75
justice have a close connection withthe relevance of the correspond-
ing views of equality.
5.2. RAWLSIAN JUSTICE AND THE POLITICALCONCEPTION
By far the most influential—and I believe the most important—
theory of justice to be presented in this century has been John
Rawls's 'justice as fairness*. The main aspects of that theory are well
known and have been extensively discussed.7
Some features have
received particular attention. This includes Rawls's use of the
device of 'the original position'-—a hypothetical state of primordial
equality in which people (without knowing exactly who they are
going to be) are seen as choosing between alternative principles that
would govern the basic structure of the society. That procedure is
seen as fair, and the principles regarding the basic social structure
that would be picked by that fair procedure are seen as just,
The rules of justice include a pair of principles. The formulation
of these principles has undergone some change since their presenta-
tion in The Theory of Justice (Rawls 1971: 60, 83, 90-5), partly to
clarify what was ambiguous, but also to respond to some early
critiques (e.g. by H. L. A. Hart 1973). In his 1982 Tanner Lectures,
Rawls stated these principles thus:
1. Each person has an equal right to a fully adequate scheme of equal basic
liberties which is compatible with a similar scheme of liberties for all.
2. Social and economic inequalities are to satisfy two conditions. First, they
must be attached to offices and positions open to all under conditions of fair
equality of opportunity; and second,they must be to the great benefit of the
least advantaged members of society,8
The first principle involves a weakening of the condition of liberty
('a fully adequate scheme' demands rather less than the original
requirement of "the most extensive total system' specified in the 1971
version). The second principle continues to include the so-called
'Difference Principle* in whichthe focusis on producing 'the greatest
benefit of the least advantaged", with advantage beingjudged by the
holding of'primary goods' (Rawls 1971: 90-5). But 'fair equality of
opportunity' receives renewed emphasis here.
While these features of Rawls's theory have received wide atten-
1
An eariy set of responses can be found in Norman Daniels's collection Reading
Rawk (Daniels 1975), See also Phdps's (1973) collection on 'economic justice*.
8
Reprinted in Rawls et at. (1987; 5).

tion even among economists, it is important to interpret these
features in the light of some of the political aspects of his approach.
In particular Rawls himself has insisted on the need to seehis theory
as 'a political conception of justice' (see Rawls 1985, 1987, 1988a,
19886,1988c, 1990), I begin with an examination of that feature and
the bearing it might have on the importance of questions of equality
in particular social circumstances.
Two distinct features of Rawls's characteriration of his political
conception of justice can be fruitfully separated out. One feature
relates to the subject-matter of the political conception: 'a political
conception of justice ... is a moral conception worked out for a
specific kind of subject, namely, for political, social, and economic
institutions.'* This does not specify any particular principle that
would have to be used for a moral conception to be political. The
matter turns on whether the subject is 'political', in the sense of
dealing with 'political, social, and economic institutions*.
The other feature, in contrast, relates precisely to a particular
principle to be used, associated with a specific form of social judge-
ment and choice, viz. that of!
a constitutional democracy', inwhich
'the public conception of justice should be, so far as possible, in-
dependent of controversial philosophical and religious doctrines*.
'To formulate such a conception, we apply the principle of toler-
ation to philosophy itself: the public conception of justice is to be
political, not metaphysical.*10
In this characterization, the subject-
matter is not in itself critical, and the crucial 'political* feature is the
"toleration* of possibly divergent comprehensive doctrines (subject
to these ideas of the good themselves satisfying certain features of
toleration, i.e. 'the ideas included must be political ideas').
In Rawls's analysis these two distinct features are closely related,
so much so that he seems to view them inseparably together. It is,
however, possible for an approach to be 'political* in the sense of the
subject-matter (as specified by Rawls) without its endorsing the
feature of'toleration* as a qualifying condition for a theory having
some claims to justice. I make this point here not because I regard
the issue of toleration to be unimportant—quite the contrary, I
regard it to be one of the central'issues in thinkingpolitically about
justice.u
But there can be important issues ofjustice and injustice in
the choice of "political, social, and economic institutions' even when
9
Rawls (1985: 224). !0
Rawls (1985: 223).
11
On this, see Sen (I970a, I985«).

pluralist tolerance of the kind outlined by Rawls simply does not
obtain. While 'toleration', in the sense discussed by Rawls, of
different comprehensiveviews of the good is undoubtedlyone of the
most important political aspects of living together in a society, it is
nevertheless not the only thing that is 'political' in social living.12
The definitional exclusion contained in Rawls's 'political concep-
tion' limits the scope of the concept of justice drastically and
abruptly, and it would often make it hard to identify political rights
and wrongs that a theory of justice should address.
Specifically, on grounds of the absence of toleration a whole lot
of comprehensive doctrines may be ruled out of court (indeed, in
some social situations none of the ones actually championed by
different political groups may remain), and yet there may be very
perspicuous problems of inequality, deprivation, and injustice in the
disputes between the different sides. To be without a theory that can
deal with such problems (when the different sides are intolerant),
and to see the disputes as lying outside the purviewof the so-called
political conception of justice, would appear to be oddly limiting
for the domain of a political conception of justice,
Consider, for example, the well-known dictum of social choice
apparently enunciated by Emperor Haile Selassie during the Eth-
iopian famines of 1973, explaining the absence of famine relief
measures undertaken by his government: 'We have said wealth has
to begained through hard work. Wehave said those who don't work
starve.*13
This is, of course, an old no-nonsense principle, which has
often been articulated, and whichmight even be seen—by stretching
a point—as having some biblical support,14
That 'principle' was, in
fact, very efficiently practised in Selassie's Ethiopia, and at the
height of the 1973 famine, there was little state-arranged relief.15
It is not hard to argue that the Emperor's political ethics regard-
12
There is a related—but larger—issue regarding th« exact role of 'neutrality" in
political liberalism and the feasibility and desirability of imposing neutrality on
theories of justice and fairness. For divergent assessments of that question, see e.g.
Dworkin (1978, W8S), Fishkin (1983), Raz{1986), Larroore (1987), Aekerman (1988),
Rawls (1988a), Pogge (1989), Van Parijs (199!). The discussion here bears on that
issue, but I shall not, here, go on to » fuller treatment of that larger problem.
IJ
Quoted in L. Wiseberg, 'An International Perspective on the African Famines,'
in Glantz (1976: 108).
M
'If any would not work, neither should he eat* (2 Tttess. 3: 10).
11
The Emperor's principles were not, in fact, the only reason for the disastrous
delay in arranging public relief, and there were other factors involved, including a
misdiagnosis of*the nature of the famine and of its causation (on whichsee Sen 198!«,
ch. 7, and also Glantss 1976). But I am not concerned with those issues here.

ing the choice of social and political institutions,as expressed in the
dictum, violates requirements of justice quite robustly. Indeed, at
the substantive level, one can easily point to the drastic inequalities
in capabilities between famine victimsand the rest of the society, and
also—in this case—to great inequalities in the holdings of primary
goods. The argument can proceed by pointing to the wrongness of
denying to out-of-work famine victims—unable to find remunera-
tive work for survival—reasonable claims on support from the
rest of the society.16
There are different ways of working out this
argument in political ethics, and invoking Rawls's device of the
'original position' would be one of the most effective ways.
But neither the Emperor, nor the opponents of his regime, who
eventually overthrew him in a bloody uprising while the famine
raged on, gave any indication of accepting any principle of toler-
ation of the other's view of the good. Indeed, each side pursued their
own objectives with no quarter given to the objectives of the others,
and as far as one canjudge, had no interest in looking for a political
solution based on toleration with the desire of living together. In
terras of a political conception of justice that requires such toler-
ation, it would be hard to pass any judgement about justice in this
case, And yet it would be peculiar to claim that no decidable issue
ofjustice in a political conception was involved in the dispute about
institutional famine relief, and that the principles of social choice
embodied in Haile Selassie's statement (to wit, no state relief for the
out-of-work famine victims) simply lay outside the domain of the
political conception of justice. Justice, in this restrictive political
conception, would seem to have a high admission price,"
All this need not be seen as being at all embarrassing for a theory
ofjustice that self-consciously 'starts from within a certain political
tradition', and that is presented with the 'hope that this political
conception of justice may at least be supported by what we may call
an "overlapping consensus", that is, by a consensus that includes
all the opposing philosophical and religious doctrines likely to
persist and to gain adherents in a more or less just constitutional
16
This case, incidentally, also illustrates well the force of Judith Shklar's (1990)
general argument that the sense of Injustice is a specially cogent starting point for
social analysis and assessment
17
It is, of course, easy to indict Selassie of being intolerant. But the injustice in
question relates not just to that, but to the principles of famine relief—more correctly
non-relief— followed by his government. The insistence on toleration as a common
agreed basis would prevent that question from even being raised.

democratic society' (Rawls 1985:225-6), So there is no real problem
here for Rawls's analysis in terms of his own programme.
It is, however, important to ask whether this particular political
conception gives the idea ofjustice—evenjusticein apolitical sense—
its due. Many blatant injustices in the world take place in social
circumstances in which the invoking of "political liberalism' and the
"principle of tolerance' may be neither easy nor particularly helpful.
And yet to leave these matters out of the scope of a 'political concep-
tion of justice' would be to reduce its domain severely. There are
many conspicuous issues of justice and injustice involved in the
political choice of social institutions all over the world, and it is not
easy to accept the definition of a political conception ofjustice that
rules most of them out of court on grounds of ideological remoteness
from constitutional democracies. The limits of 'the political' need
not beseen to be soclose.'"The pervasive problems of inequality and
injustice in the world call for a less restrictive approach.
While the preceding discussion points to the limiteddomain of the
Rawlsian conception of justice, especially in the light of the em-
phasis hehas recentlyplaced on the aspect of'toleration', it is impor-
tant to recognize that the Rawlsian perspective—particularly the
Difference Principle-—has been widely used in the literature on econ-
omic and social development. The insights that people have ob-
tained from Rawls's analysisof'justice as fairness' seem to have gone
greatly beyond the limits he has himself imposed, and it is not clear
to me that these insights have been all misderived and misdirected.
When it comes to evaluatingRawls'stheory ofjustice as a whole, itis
ofcourse necessary to seeit withinthe specific constraints imposed by
the author himself, but the 'Rawlsian outlook* in a less constrained
form has had a profound impact on a much wider range of contem-
porary political, social, and economic thinking. In particular, the
literature on the evaluation of inequality has not been quite the same
again since Rawls's classic book made its first appearance.
5.3, PRIMARY GOODS ANDCAPABILITIES
There have been various approaches focusing on equality of oppor-
tunities—characterized in different ways—in the recent literature of
18
The point is not about the use of the word 'political', but rather about the
motivation underlyingthe idea of a political conception. It is, however, possible also
to argue that Rawls uses a particularly narrow definition of the term political.

justice. Rawls's concentration on the distribution of 'primary
goods'—including 'rights, liberties and opportunities, income and
wealth, and the social bases of self-respect*19
—in his Difference
Principle can be seen as a move in that direction. This approach can
also be interpreted, as I argued earlier on in this monograph, as
taking us in the direction of the overall freedom actually enjoyed by
people, and this has the effect of reorienting the direction of the
analysis of equality and justice towards freedoms enjoyed rather
than being confined to the outcomes achieved. An important prob-
lem arises from the fact that primary goods are not constitutive of
freedom as such, but are best seen as meansto freedom (an issue that
we discussed in Section 3.3 earlier). Ronald Dworkin's (1981, 1987)
case for 'equality of resources' can also be seen as belonging
broadly in the same general area of substantive accounting, since
resources are also means to freedom, and Dworkin has, in fact,
presented a specific way of accounting resources and adjudicating
'the equality of resources'.
One problem is that of valuation. Since means are ultimately
valued for something else, it is not easy to set up a scheme of valu-
ation of means that would be reallyindependent of the ends. It is by
skilful use of this connection that John Roemer (1986&) has estab-
lished a mathematical result which he has interpreted as 'equality of
resources implies equality of welfare* (the title of his article). The
result has been based on an elaborate set of axioms, but the basic
insight behind the result can be seen as seeking the value of resources
in terms of what the resources yield. Since resources are not valued
for their own sake, such a connection has some obvious plausibility.
By taking a model in which the only ultimate end is welfare, the
result that equality of resources must yield equality of welfare has
emerged in Roemer's theorem.
The congruence of resource valuation with welfarevaluation can
be, in fact, replaced by a similar congruence with whatever is taken
to be the end the promotion of which is the reason for valuing the
resources. The real issue behind this interesting result is the depen-
dence of the valuation of means on the valuation of ends (and not
specifically the interdependence of resources and welfare).
In what follows I shall be concerned primarily with Rawls's
theory of justice as fairness, but some of the comments will apply
also to Dworkin's approach.
19
See Rawls (1971: 60-5).

Themain issueisthe adequacy ofthe informationalbaseof primary
goods for the political conception ofjustice in Rawls's sense, and the
need ifany for focusing on capabilities. Primarygoods are 'things that
every rational man is presumed to want', and include 'income and
wealth', 'the basic liberties', 'freedom of movement and choice of
occupation*, 'powers and prerogatives of offices and positions of
responsibility', and 'the social bases of self-respect*,20
Primary goods
are, thus, generalpurpose meansor resources useful for the pursuitof
different ideas of the good that the individualsmay have.
Earlier in this monograph (especially in Chapter 3), I have dis-
puted the alleged adequacy, for a freedom-oriented assessment of
justice, of this concentration on means to freedom, rather than on the
extent of the freedom that a person actually has. Since the conver-
sion of these primary goods and resources into freedom of choice
over alternative combinations of functionings and other achieve-
ments may vary from person to person, equality of holdings of
primary goods or of resources can go hand in hand with serious
inequalities in actual freedoms enjoyed by different persons. The
central question, in the present context, is whether such inequalities
of freedom are compatible with fulfilling the underlying idea of the
political conception of justice.
In the capability-based assessment of justice, individual claims
are not to be assessed in terms of the resources or primary goods the
persons respectively hold, but by the freedoms they actually enjoy
to choose the lives that they have reason to value,21
It is this actual
freedom that is represented by the person's 'capability' to achieve
various alternative combinations of functionings.
It is important to distinguish capability—representing freedom
actually enjoyed—both (1) from primary goods (and other re-
sources), and (2) from achievements (including combinations of
functionings actually enjoyed, and other realized results). To illus-
trate the first distinction, a person who has a disability can have
more primary goods (in the form of income, wealth, liberties, and so
on) but less capability (due to the handicap). To take another
example, this time from poverty studies, a person may have more
income and more nutritional intake, but less freedom to live a well-
nourished existence because of a higher basal metabolic rate, greater
20
See Rawls (1971; 60-5); Rawls (1982: 162); Rawls (1988a: 256-7),
21
Various aspects of this claim and their diverse implications were discussed in
Chs. 3 and 4.

vulnerability to parasitic diseases, larger body size, or simply be-
cause of pregnancy. Similarly, in dealing with poverty in the richer
countries, wehave to take note of the fact that manyof those who are
poor interms ofincomeand other primarygoods also have character-
istics—age, disability, disease-proneness, etc.—that make it more
difficult for themto convert primarygoods into basiccapabilities, e.g.
being able to move about, to lead a healthy life, to take part in the life
ofthecommunity. Neither primarygoods, nor resources more broad-
ly defined, can represent the capability a person actually enjoys.
To illustrate the second distinction, a person may have the same
capability as another person, but nevertheless choose a different
bundle of functionings in line with his or her particular goals.
Furthermore, two persons with the same actual capabilities and
even the same goals may end up with different outcomes because of
differences in strategies or tactics that they respectively follow in
using their freedoms.
In responding to my critique, Rawls has tended to assume that it
is based on presuming that everyone has the same common ends—
shared objectives pursued by all. This is based on the belief that if
they had distinct objectives, then the differential conversion rates of
primary goods into capabilities could not be ascertained. That
assumption (viz, the same objectives for all), if made, would cer-
tainly go against Rawls's political conception of justice, which
admits interpersonal variation of ends, with each person havinghis
own 'comprehensive view of the good'. Rawls summarizes his inter"
pretation of my objection thus:
... the idea of primary goods must be mistaken. For they are not what,
from within anyone's comprehensive doctrine, can be taken as ultimately
important: they are not, in general, anyone's idea of the basic values of
human life. Therefore, to focus on primary goods, one way object, is to
work for the most part in the wrong space—in the space ofinstitutional
features and material things and not in the space of basic moral values.22
Rawls's response to his interpretation of my objection is the fol-
lowing:
In reply, an index of primary goods is not intended as an approximation to
what is ultimately important as specified by any particular comprehensive
doctrine with its account of moral values.23
21
Rawls (I988a: 256-9).
23
Rawls (lf?88a: 259). Rawls also has a rather different line of answering my
criticism in his 'Reply to Sen' (Rawls 19886). He argues that his full theory ofjustice

The main problem with this reply lies in the misinterpretation
of the nature of the objection, Capability reflects a person's
freedom to choose between alternative lives (functioning combi-
nations), and its valuation need not presuppose unanimity regard-
ing some one specific set of objectives (or, as Rawls calls it, *a
particular comprehensive doctrine'). As was discussed earlier, it is
important to distinguish between freedom (reflected by capability)
and achievement (reflected by actual functionings), and the evalu-
ation of capability need not be based on one particular com-
prehensive doctrine that orders the achievements and the life-
styles.24
The second problem, related to the first, concerns Rawls's claim
that primary goods are 'not intended as an approximation to what
is ultimately important as specified by any particular comprehen-
sive doctrine* (emphasis added). This is a legitimate enough con-
cern for Rawls's 'political conception of justice', but the tack of
correspondence with primary goods does not lie only there. It lies
also in the fact—more important in the present context—that a
disadvantaged person may get less from primary goods than others
no matter what comprehensive doctrine he or she has.
To illustrate the point, consider two persons 1 and 2, with 2
disadvantaged in some respect (e.g. physical disability, mental
handicap, greater disease proneness). They do not have the same
ends or objectives, or the same conception of the good. Person 1
values A more than B, while 2 has the opposite valuation. Each
values 1A more than A> and 2J more than B, and the orderings of
the two (representing the relevant parts of their respective "compre-
hensive doctrines') are the following:
has more 'flexibility' than I recognize, and some of the interpersonal variations 1 am
concerned with can be taken note of at later stages, such as 'legislative' and
'judicial' ones. It is not altogether easy to be sure what overall procedures and
altocational principles would be in fact satisfied by such a complex stagewise
structure, but if it is indeed the case that all the relevant interpersonal variations
will be effectively dealt with at some stage or other, then that would certainly
reduce the force of the criticism. Some of the issues raised by interpersonal vari-
ations in the conversion of primary goods into capabilities would then end up
receiving attention after all.
24
In the capability format, achievement is given by an n-tupte of relevant func-
tionings, whilecapabilityitself is a set of such it-tuples from which one can be chosen.
On some alternative forms of representation, and their relevance to the analysis of
individual advantage and thus to the study of inequality, see also Chs. 3 and 4.

Person 1 Person 2
2A 2B
26 2A
A S
B A
With the given set of primary goods, person 1can achieve 1A or
2B,also—though there may beno great merit in this—A or B. On the
other hand, given 2's disadvantage, with the very same primary
goods, she can achieve only A or B. Person 1proceeds to achieve 2A
(the best feasible outcome for him), while 2 settles for B (the best
feasible outcome for her). The problem is not just that 2 is at a
disadvantage in terms of onsparticular comprehensivedoctrine (her
own or that of person 1),but she has a worse deal than 1nomatter
which comprehensive doctrine we take. Equality of primary goods
has given 2 lessfreedom to achieve and not just less achievement with
respect to some one comprehensive doctrine.
If the comparisons were made not in terms of primarygoods, but
in terms of capabilities, 2's worse deal would be obvious. Person 1 *s
capability set consists of (A, B, 2A, 2B), whereas 2's capability isjust
a proper subset of it, to wit (At B), with the best elements-—no matter
which comprehensive doctrine is considered—lost. Capability rep-
re$ents/re«fww, whereas primarygoods tell us only about themeans
to freedom, with an interpersonally variable relation between the
means and the actual freedom to achieve. Rawts is right to think that
my objection did relate to primarygoods being meansonly, but that
problem is not disposed of by saying that they are not meant as an
approximation of 'any particular comprehensive doctrine*.2
*
25
Dominance in the space of capabilities does not require agreement on any one
comprehensive doctrine, since one capability set can bea proper subset of another {as
in the example given). Furthermore, even when the capability sets ate not subsets of
each other, for agreement to exist on their ranking,wedo not need the acceptance of
any one comprehensive doctrine. Partial rankings of capabilities can be based on
superiority in terms of each of the relevant comprehensive doctrines. However, to
insist on a complete ordering can be problematic. (There is a similar problem of
complete indexation of the holdings of primary goods for Rawls's Difference Prin-
ciple, since different primary goods can be disparately effective in the pursuit of
different comprehensive ends; on this and related problems, see Plott 1978, Gibbard
1979, Blair 1988, Sen 1992c.) But partial ordering*can be an adequate basis for many
evaluative judgements, especially in dealing with serious problems of inequality.
How extensive the rankings of capability sets turn out to be will depend on (I) the
divergence between the relevant comprehensive views,and (2) the differences between
the sets to be ranked. The analytical problems involved are discussed in Sen (1970a,
1970ft, 1985*). See also Ch. 3 above.

5,4, DIVERSITIES: ENDS AND PERSONAL CHARACTERISTICS
There are in fact two sources of variation in the relation between
a person's means in the form of primary goods (or resources) and
achievement of ends. One is inter-end variation—different concep-
tions of the good that different people may have. The other is
inter-individual variation in the relationship between resources
(such as primary goods) and the freedom to pursue ends. Rawls
shows great sensitivity to the first variation, and is keen on pre-
serving respect for this diversity (rightly so, in line with his plur-
alist political conception). To deal with this problem, Rawls
assumes that the same primary goods serve all the different ends.24
As far as the second inter-individual variation is concerned (i.e. in
the relation between resources and freedoms), the problem created
by this is in no way reduced by the existence of the first variation (i.e.
over ends and objectives). A person's actual freedom to pursue her
ends depends both on (1)what ends shehas, and (2)what power she
has to convert primary goods into the fulfilment of those ends. The
latter problem can be serious even with given ends, but it is not the
case that it can be serious only with given ends. The reach and rele-
vance of the second problem is not reduced by the existence of the
first.
To conclude, human beings are diverse, but diverse in different
ways. One variation relates to the differences in ends and objec-
tives. The ethical and political implications of this diversity we
now understand much better as a result of Rawlsian analysis of
justice as fairness. But there is another important diversity—
variations in our ability to convert resources into actual
freedoms. Variations related to sex, age, genetic endowments, and
many other features, give us very divergent powers to build
26
Presumably for the sake of fairness it must not be the case that some people's
ends are so Imperfectly served by the primary goods (compared with the ends of
others) that the first group may have a legitimate complaintaboutjudging individual
deals in terms of primary goods, Rawls's comprehensiveassertion that 'approxi-
mation* of primary goods to 'no other space of values' is needed (indeed has to be
shunned) would seem to overlook the nature of this particular problem. If every
possible list of primary goods (and every wayof doing an index) makessome people's
ends very well served and others* very poorly indeed, then the importantfeature of
'neutrality' will be lost, and the entire line of reasoningof "justice as fairness" may be
significantly undermined. Thus, some strong requirements are imposed on the rela-
tion between primary goods and the spaces of other values. I shall not discuss this
issae further in this essay. See also Ruth Anna Putnam (1991).

freedom in our lives even when we have the same bundle of primary
goods.27
The Rawlsian approach to justice has transformed the way we
think about that issue, and his theory has had the effect of shifting
our concerns from inequalities only in outcomes and achievements
to those in opportunities and freedoms. But by concentrating on the
means to freedom rather than on the extent of freedom, his theory of
a just basic structure of the society has stopped short of paying
adequate attention to freedom as such. While the motivation for
focusing on the means of freedom might have, it would appear,
rested on Rawls's belief that the only alternative would be to
choose one particular comprehensive view of outcomes and achieve-
ments, that presumption is, as shown above, not quite correct.
Freedom can bedistinguished both from the meansthat sustain it and
from the achievements that it sustains.
Rawls's theory ofjustice has many distinct features, and the ques-
tions raised here must not be seen as an attempt to undermine the
entire approach. Indeed, it would be difficult to try to construct a
theory of justice today that would not have been powerfully influ-
enced by the illumination provided by Rawls's deep and penetrating
analysis.28
The point of criticism relates specifically to the tension
between Rawls's concentration on primary goods and his concern
for the freedoms we enjoy to pursue our ends. In so far as freedoms
are what we are concerned with, I have tried to argue that there is a
different—more accurate—way of examining the distributive issue.
Rawls is, in fact, concerned with many other things as well, in-
cluding the importance of certain liberal institutions and processes,
and the need to restrain public policy when personal liberty is
threatened. The discussion on equality of effective freedoms pre-
sented here does not dispute these aspects of Rawls's concerns.
27
Some of the empirical issues involved are discussed in Sen (1984, I985A, 1988c)
and Kyneh and Sen (1983).
'* In proposing his alternative political theory, Robert Nozick says: 'Political
philosophers now must either work within Rawls' theory or explain why not. The
considerations and distinctions we have developed are illuminated by, and help
illuminate, Rawls" masterful presentation of an alternativeconception. Even those
who remain unconvinced after wrestling with Rawis* systematic vision will learn
much from closely studying it' {Nozick 1974: 183). Needless to say, the last remark
will apply even more to those who remain unconvinced only of particular parts of
Rawls's overall conception. But here we are concerned specifically with such a part,
and hence the concentration on the differences with Rawls, rather than on the many
points of agreement and on the great debt owed to Rawlsfor teaching uswhat it is like
to examinejustice.

The point may be illustrated by taking up the role that liberty is
given in Rawtsian theory ofjustice. Rawls givescomplete priority to
the principle of liberty over other principles of justice, and this
rather extreme formulation has been cogently disputed by Herbert
Hart (1973).2
* On the other hand, it can be argued (on this, see Sen
1970a, 1983a) that some additional recognition is indeed needed for
liberty over and above the atteetion it may receive as one primary
good, or as one influenceon well-being, or even as one of the causal
determinants of a person's capability. Indeed, a person's capability
may be reduced in exactly the same way in two cases: (1) through a
violation of his 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 twocases. In this sense, thecapability per-
spective, central as it is for a theory of justice, cannot be entirely
adequate for it. There is a real need to bring in the demands ofliberty
as an additional principle (even if that principle is not given the total
priority that Rawls recommends). The importance of the over-all
freedom to achieve cannot eliminate the special significance of nega-
tive freedom.30
Our focus in the present discussion has been only on a specific part
of Rawls's theory of justice and the relation between one of his
concerns and his proposed way of dealing with it. But in that
specific—and I believe crucial—part of Rawlsian theory of justice,
the point to emerge from our analysis is-—I would argue—of some
conceptual and practical importance. Equality of freedom to pur-
sue our ends cannot be generated by equality in the distribution of
primary goods. We have to examine interpersonal variations-in the
transformation of primary goods (and resources, more generally)
into respective capabilities to pursue our ends and objectives.
If our concern is with equality of freedom, it is no more adequate
to ask for equality of its meansthan it is to seek equality of itsresults,
Freedom relates to both, but does not coincide with either.
29
Note, however, that the principle of liberty that receives this priority is itself
made less demanding in the later formulationchosen by Rawls, quoted earlier in this
chapter, compared with the 1971version.The change is largely in response to Hart's
(1973) forceful critique.
30
On this sec Sen (19700, 1976e, 1985«: Lecture 3).

6
WELFARE ECONOMICS AND
INEQUALITY
6,1. SPACE CHOICE AND EVALUATIVE PUSPOSE
As was discussed in the earlier chapters, the evaluation of inequality
has to take note of both the plurality of spaces in which inequality
can be assessed, and the diversity of individuals. The relative advan-
tages and disadvantages that people have, compared with each
other, can be seen in many different perspectives, involving different
concentrations, e.g. liberties, rights, incomes, wealths, resources,
primary goods, utilities, capabilities, and so on, and the question of
inequality assessment turns on the selection of the space in which
equality is to be assessed. While the pictures of inequality in the
different spaces are not unrelated to each other, the pervasive
diversities of human beings make them non-congruent—indeed, fre-
quently far apart.
That much has already been discussed in some detail. It is, how-
ever, also worth emphasizing that the relative appropriateness of the
different spaces depends ultimately on the motivation underlying
the exercise of inequality evaluation. Inequality is measured for
some purpose, and the choice of space as well as the selection of
particular inequality measures in that space would have to be made
in the light of that purpose. There is, of course, nothing at all
surprising in the recognition that the nature of interpersonal com-
parisons and the assessment of inequality should depend on whatwe
are after.1
Sometimes we may be interested in knowing how disparate the
levels of well-being of different people happen to be, e.g. between
different classes or distinct communities, or between women and
men in a particular group. In the light of the discussion in the last
chapter, one could argue that comparisons in thefunctioning space
may be more relevant for the analysis of well-being than in the
spaces of incomes, primary goods, or resources. And yet in some
1
On this see Broome (1987).

Welfare Economicsand Inequality 89
other context wemay be interested precisely in the relative positions
of different individuals (or groups) in terms of incomes, or primary
goods, or resources. Even if the distribution of incomes gives us little
idea of the inequalities in well-being as such, it does not follow that
income distribution is of no direct interest, since there are possible
motivations other than well-being comparison.
For example, we may be interested in knowingwhat effect income
inequality has, say, on crime, or on social discontent, or for that
matter—inter alia—on the distribution of well-being. Similarly, we
may wishto check what kind of a distribution of primary goods (or of
resources) a state or a political system is working towards, and this
may bean interesting issue to investigatein understanding the role of
public policy—not in promoting well-being or freedom as such, but
in making the means of freedom available to all.2
The argument for
paying greater attention to functionings (or capabilities) in assessing
inequalities of well-being (or of freedom) must not be seen as an
all-purpose preference for those variables.
6.2. SHORTFALLS, ATTAINMENTS ANDPOTENTIALS
Before I go into some specific questions that have been much dis-
cussed in the recent welfare-economic literature on inequality, I
would like to address a distinction that has some general relevance
in judging individual advantage. Should a person's position be
judged, positively, in terms of the level of achievement, or nega-
tively, in terms of the short/all vis-&~vi$ what she could have
maximally achieved? The two general approaches in assessing
achievement are closely linked but nevertheless distinct.3
2
This is,in fact, a part of Rawis's(1971)and Ronald Dworkia's(1981)motivation
in placing the emphases they do oa distributions in these spaces. SeealsoDahrendorf
(1988).
3
Shortfalls have been used as the basis of evaluation in welfare-economic litera-
ture in various forms. For example, FrankRamsey's (1928) formulation of the prob-
lem of optimum saving was in terms of the minimization of the shortfall from the
maximum level of aggregate utility (the so-called 'bliss'). In fact, Ramsey's shortfall
minimization problem is well defined over infinite time in a way that the direct
fornrolation of the problem (as the maximization of the utility-stun over an ininite
horizon) would not have been (on which, seeSukhamoy Chakravarty 1969). To take
another type of case, Dalton's (1920) measure of inequality looks for the shortfall
vb-d-vis the maximal value of social welfare that could be achieved through equal
distribution. Similarly, Atkinson's (1970Ji) measure of inequality looks for the
income equivalent of the social-welfare shortfall. Dalton'sand Atkinson's measures
of in equality are discussedlater on in this chapter. See also Musgrave (1959) on
different conceptsof 'sacrifice*.

90 Welfare Economicsand Inequality
The two approaches can yield different results for two distinct
reasons. First, if the maximal value from which shortfall is
calculated is varied between the persons (e.g. on the pound that
what is maximally achievable by that specific person may not be the
same as the achievement that another person could have maximally
attained), then the ordering of absolute attainmentscan differ from
that of the respective shortfalls. Second, even when the same
maximal value is taken for all, a difference may arise from the fact
that the comparisons may be done not in absolute terms, but in the
scale of 'proportionate* attainments and shortfalls. There may be a
case for that shift in perspective when the extent of 'difficulty' in
achievement is seen as becoming relatively more acute with the rise
in absolute attainment.4
The second issue may or may not be an immediate one in the
particular context of assessing the level of interpersonal inequality;
the formulation of achievements in proportionate terms would
often be rather remote. But the first issue—that of interpersonal
variations in maximal potentials—is peculiarly relevant when in-
equality assessment is addressed after a clear recognition of human
diversity.
Equality between persons can be defined either in terms of attain-
ments, or in terms of the shortfalls from the maximal valuesthat each
can respectively attain. For 'attainment equality' of achievements,
we compare the actual levels of achievement.5
For 'shortfall
equality*, what are compared are the shortfalls of actual achieve-
ments from the respective maximal achievements.6
Each of the two
4
For example, in assessing the performance of countries in terms of average life
expectancy at birth, increasingthe fife expectancyfigurefrom, say, 40yearsto 50years
is a 25%increase, whereas raising it from 60 years to 70years is only a 17% increase. It
is, nevertheless, arguable that the latter is a much harder task, since further improve-
ments get increasingly more difficult as we get closer to the maximal possible value,
To alter the perspective, if we take, as our working assumption, 80 years (say) as the
maximal value of average life expectancy at birth, then a movefrom 40 to 50years isa
reduction of the shortfall by 25%, whereas raising life expectancy from 60 to 70years
would involve a reduction of the shortfall by 50%, It can be argued that the latter
ranking is relatively more in line with the recognition of the difficulty in raising life
expectancy further, as the absolute achievements become higher. The underlying
issues are discussed in Sen (IfSli). Empirical examples of the use of proportionate
shortfall comparisons for intercountry contrasts can be found there, and also in
UNDP (1990). See also Desai, Boltvinnik, and Sen (1991),
5
In the case of freedoms, attainment equality compares the levels of alternative
actual achievements from which the person can choose.
* Correspondingly, in the case of freedoms, we have to look at the differences in
shortfalls from the respective maximal freedoms to achieve,

Welfare Economics and Inequality 91
views has some considerable interest of its own. Shortfall equality
takes us in the direction of equal use of the respective potentials,
whereas attainment equality is concerned with equal absolute
levels of achievement (no matter what the maximal potentials are).
If human diversity is so powerful that it makes it impossible to
equalize what ispotentially achievable, then there is a basic ambigu-
ity in assessing achievement, and injudging equality of achievement
or of the freedom to achieve. If the maximal achievement that person
1 can have—under the most favourable circumstances—is, say, x,
while person 2 can maximally manage 2x, then equality of attain-
ment would leave person 2 invariably below his or her potential
achievement. It is partly to address such issues that Aristotle had
incorporated a parametric consideration of what a person's 'circum-
stances admit', and had seen his 'distributive conception* in that
light. 'For it is appropriate, if people are governed best that they
should do best, in sofar as (heir circumstances admit—unless some-
thing catastrophic happens,"
In the case of serious disabilities, attainmentequality may be hard
to achieve, and it may be particularly tempting to opt for shortfall
equality. There might well be a good argument in that direction, but
I would like to argue that it is not the case that the choice is made
clear-cut simply by the non-feasibilityof attainment equality. It can
be argued that even when a disabled person cannot, in any way, be
given the freedom to enjoy the same level of the functioning in
question (e.g. the same ability to move about freely as others), there
is nevertheless a good case—based on fairness—for trying to
maximize his below-par functioningability, rather than settling for
the same shortfall (absolute or proportionate) as others have from
their-—much higher—maximal functioning(as would be demanded
by shortfall equality).
Indeed, that is the direction in which the Rawlsian logic of
'maxirnin' or 'lexicographic maximin' should take us-—'to make the
worst off as well off as possible*.8
This is, in fact, an illustration of
7
Aristotle, Politics, VII. 1 (1323*17-19), emphasis added. The translation istaken
from Nussbaum (1988a), who also discusses the precise role that this qualification
plays in Aristotle's "distributiveconception' (pp. 146-50).
* There is a more complex problem of fairness when the different maximal
achievements are not particular to the individual,but apply to a whole'natural type*,
e.g. children vis-d-vis adults. Indeed, Aristotle's reasoning, presented earlier, seems
to relate to inter-type comparisons, rather than to inter-individual variations (on this
see Nussbaum 1988o). The lower ability of children to acquire some capabilities (e.g.
in activities for which fuller bodily development or greater maturity is important)

the point made earlier in this monograph that a good deal of
Rawlsian reasoning can be applied even beyond his own constrained
framework. Rawls's own principles of justice will not directly
admit the claim of the disabled in the domain of 'maximin' reason-
ing, since his Difference Principle restricts the use of 'maximin* to
holdings of primary goods, whereas the deprivation of the disabled
occurs in the field of capability (because of his unfavourable trans-
formation possibility of primary goods into capability). But
Rawls's general argument for focusing on the least advantaged does
apply here (for reasons of 'fairness* that he has analysed so power-
fully). That certainly provides reason enough to take seriously the
claims of attainment equality even when the maximal achievements
are quite diverse.
It could, of course, be objected that a policy of attainment
equality would lead to a very 'low level equality' for all. In trying
to achieve equality when person I can only achieve x, person 2 will
have to be dragged down from his maximal achievement of 2x, and
each will have, at most, x. There is some force in that objection, but
it would have been a more telling counterargument had equality
been the only principle to be used. However, as was argued earlier
(and will be discussed further in the last chapter), equality would
typically be one consideration among many, and this could be com-
bined with aggregative considerations including efficiency. These
latter influences would work against choosing "low-level equality',
and against pulling person 2 down to the low level of person 1just
for the sake of achieving equality of attainments.
The real question isnot about the kind of equality to ask for /fthat
were the only principle to beused, but whether in a mixedframework
in which aggregative considerations as well as equality are taken
into account, the demands of equality as such are best represented by
'shortfall equality' rather than by 'attainment equality', I have not
argued for unqualified pursuit of attainment equality, but have
argued against opting for shortfall equality on the simple ground
might not be seen as a definitive argument for giving priority to expanding these
capabilities for the children at the cost of the larger capabilitiesof the adults. This
problem would not arise in that form if the advantagesof individuals are seen in
terms of lifetime opportunities (as Rawls 1971himself suggests), but the question is
not entirely dismissible, since thereis a further issue of inter-agedistribution that the
lifetime focus tends to avoid. There are also complexproblemsof fairness related to
the biological advantagesthat women seem to have vis-a-vM men in longevity, even
though women often live hiss long in many countries because of unequal treatment
(on this see Sect. 8.2).

that attainment equality may be unfeasible or inefficient (involving
the levelling down' of all to the condition of the lowest achiever),
The case for retaining an interest in moving towards attainment
equality can survive the difficulties in achieving it fully.
The issues involved in the distinction between attainment and
shortfall will reappear in specificcontexts in some of the arguments
in the rest of the monograph (particularly in Chapters 8 and 9). But
since they do not come in much in the standard welfare-economic
discussion of inequalities, with which the rest of this chapter is con-
cerned, the distinction will not figure further in this chapter. The
literature on welfare-economic theory dealing with inequality has
typically overlooked human diversities and has tended to take
everyone as being exactly similar (including inter alia having the
same maximal potentials).
6,3, INEQUALITY, WELFARE AND JUSTICE
The literature on the measurement and evaluation of inequality has
expanded very rapidly over the last decade or two, and remarkable
progress has been made in the understanding of many analytical
aspects of the evaluative problems.5
While the analytical contri-
* Recently, Blackorby and Donaldson (1978) and Foster (1985) have provided
critical surveys of the developments in the contemporary literature. See also Cowell
(1977), Nygrad and Sandstrom (1981), Eichhoro and Gehrig (1982), ChipnwB (1985),
and Lambert (1989). The literature has, by now, what seems like uncountabty many
contributions, but here is a list to get one started: Algner and Heins (1967), Theil
(1967), Koto (1969, 1976), Atkinson (19706, 1975, 1983), Bentzei (1970), Newbery
(1970), Ttebergen (1970), Pen (1971), Sheshinski (1972), Dasgupta, Sen, and Starrett
(1973), Rothschildand Stiglitz (1973), Pazner and Sehmeidler (1974), Blackorby and
Donaldson (1977, 1978), Muellbauer (19744 1978), Wolfson (1974), Gastwirth
(1975), Hammond (19766, 1977, 1978), Mehran (1976), Pyatt (1976, 1987), Bhatta-
charya and Chatterjee(l988), Cowell (1977,1980,1985,1988), Graaff(1977), B.Hans-
son (1977), Fields and Fei (1978), Kern (1978), Osmani (1978, 1982), Archibald and
Donaldson (1979), Bourgwigaon (1979), Donaldson and Weymark (1980), Deaton
and Mwellbauer (1980), Dutta (1980), Fields (1980a), Kakwani (19806, 1981, 1986),
Roberts (19806), Shorrocks (1980, 1982, 1983, 1984, 1988), Blackorby, Donaldson,
and Auetsperg (1981), S. R. Chakravarty (1981, 1988, 1990), Cowell and Kuga
(1981), Jasso (1981), Nygrad and Sandstrom (1981),Weymark(1981), Atkinson and
Bourguignon (1982), Kanbur (19826), Mookherjee and Shorrocks (1982), Then
(1982), Anand (1983), Broder and Morris (1983), Blackorby, Donaldson, and Wey-
mark (1984), Foster, Greet, and Thorbecke (1984), Jorgenson and Slesnick (1984a,
19846), Le Grand (1984), Slottje (1984), Chakrsvarty, Dutta, and Weymark (198S),
Chakravarty and Roy (1985), Fine(1985), Lambert(1985,1989), Le Breton, Trannoy,
and Uriarte (1985), Baumol (1984), Hutchens (1986), Kanbur and Stromberg (1986),
Maasoumi (1986), Ternkin (1986, 1989), K. Basu (19876), Chakravarty and Dutta
(1987), Ebert (1987, 1988), Lc Breton and Trannoy (1987), Meyer (1987), Shorrocks

94 Welfare Economics and Inequality
butions are not alwaysexplicit about the purpose of the exercise, it is
clear that many different purposes have motivated the analytical
literature.
The connection between inequality and social welfare has, how-
ever, been invoked more often than any other. That connection can
be pursued in different ways, depending on what we take to be the
'argument' of the social-welfare function. For example, social wel-
fare may be seen as a function of individualutilities, as in the wel-
farist framework, of which utilitarianism is a distinguished case,
involving 'sum-ranking* (i.e. simply adding up the utilities).10
Or
alternatively, social welfare may be seen as a function directly of the
vector of incomes (without being intermediated by the utilities re-
lated to those incomes), or of the combination of multiple-attribute
characteristics of individual economic status or opulence."
A particular approach to this class of problems is to see social
welfare as a function of the person-specific distributionof each com-
modity (i.e. of the 'named-goods vector', as it is sometimes called12
}.
Social welfare can, of course, also be seen as a function of the
combination of everyone's functioning vectors (or of everyone's
capability sets). The nature of the presumed social-welfare function,
thus, influences the type of interpersonal comparability that may be
correspondingly sought.
Therefore, even for the purpose of social-welfare assessment,
there are several possible leads to the problem of choice of space in
inequality evaluation. And as other purposes (i.e. purposes other
than the assessment of aggregate social welfare) are taken up, the
space-choice problem will also alter in conformity with the respec-
and Foster (1987), Bhattacharya, Chatterjee, and Pal (1988), Eichhorn (1988a,
1988ft), Foster, Majumdar, and Mitra (1988), Foster and Shorrocks (1988a, 19886),
Dutta and Ray (1989). There are, of course, many other contributions.
19
On the distinction between welfarism in general and utilitarianism in particular,
see Sen (1979a, 1979A) and Sen and Williams (1982), 'Introduction',
11
On this see Kota (1977), Atkinson and Bourguignon (1982), Maasoumi (1986,
1989), Foster, Majiimdar, and Mitra (1988).
a
Theconcept of a "named good'—good i going to person/' being a named good
{/—was introduced by Hahn (1971), The named-goods vector contains the same in-
formation as the "commodity matrix* as defined by Fisher (1956). The informational
enormity of a named-goods vector can be systematically handled to get practically
usable results, given some regularizing axioms, on which see Sen (19766, 1979e),
Hammond (1978), Roberts (19806). On related matters, sec Graaff (1977), Das-
gupta and Heal (1979), Atkinson and Bourguignon (1982), Broder and Morris (1983),
Osmani (1982), Atkinson (1983), Atkinson and Bourguignon (1987), Bhattacharya,
Chatterjee, and Pal (1988).

Welfare Economicsand Inequality 95
tive objectives. In this sense, the question 'equality of what?' cannot
lead to anything like a clear answer until the purpose and the
motivation behind the question have been specified,
There are, in fact, two implicit programmes that have been fre-
quently invoked as background to the exercise of inequality evalu-
ation,13
The first is that of using inequality evaluation for
analysing social justice, and specifically for the choice of the 'basic
structure* of the society in a general framework of political and
social ethics, such as that of Rawls's (1971) 'justice as fairness*.
This programme was already examined in Chapter 5, and it has
obvious implications for normative analysis of public policy,
In contrast with this 'justice-based inequalityevaluation' (much
discussed in modern ethics and political philosophy), the second
frequently invoked programme—more commonly used in welfare
economics—is that of assessing inequality in the context of social-
welfare analysis, assuming that the vector of incomes determines
the level of social welfare. This is the exercise that was specifically
investigated in the pioneering contributions of Dalton (1920),
Kolm (1969), and Atkinson (19706), and the literature that it led to
(including substantial parts of my OEI). I shall call this the prob-
lem of 'welfare-based inequality evaluation', and I shall have a
few methodological comments to make on the nature of that exer-
cise in the next section of this chapter.
6.4. WELFARE-BASED INEQUALITY EVALUATION
In his pioneering contribution to measuring inequality in terms of
social-welfare loss, Hugh Dalton (1920) used a simple utilitarian
social-welfare function. Social welfare was taken to be the sum-
total of individual utilities, and each individual utility was taken
to be a function of the income of that individual,The same utility
function was taken to apply to all individuals, and this fact, along
with diminishing marginal utility from income, ensured that for
any given total income to be distributed among the people, an
equal distribution would maximize social welfare (i.e. would
a
Another approach that has been used with illuminating results is that of survey-.
ing pople's general attitude to inequality. On this, see Yaari and Bar-Hillef (1984),
Amiel and Cowell (1989), Fields (1990).For a systematic use of common intuitions
to judge inequality, see Temkin (1986),

96 Welfare Economics and Inequality
generate the highest utility sum-total for that total income).14
Dalton identified the level of inequality for a given total income
with the percentage shortfall of the actual sum-total of utilities
from the maximal value, i.e. the sum-total that would be generated
through equally distributing the given total income over all the in-
dividuals.
Since Dalton's measure of inequality operates on utilities as
such, it is very exacting on the measurability and interpersonal
comparability of individual utilities. It is, in fact, not easy to talk
about percentage shortfalls of utility sum-totals from the maximal
sum-total (e.g. 'the sum of utilities is reduced by 17 per cent'),15
Atkinson's (197(M») index of inequality, in contrast, operates on
incomes, and measures the social loss involved in unequal income
distribution in terms of shortfalls of equivalent incomes. Atkinson
measures the inequality of a distribution of incomes by the percent-
age reduction of total income that can be sustained without
reducing social welfare, by distributing the new reduced total ex-
actly equally. It requires judgements of the kind: 'A 22 per cent
smaller total income, if equally distributed, would be just as good
for the society as the present [higher] income distributed as [un-
equally as] it in fact is.* The more unequal the distribution of
present incomes, the more of a reduction of total income is
sustainable, without social-welfare loss, by distributing the new
total equally. The level of the "equally distributed equivalent
income' (in this case 22 per cent lower than the actual income) is
the crucial concept in Atkinson's approach, and the difference with
the actual income (in this case 22 per cent) gives us Atkinson's
measure of inequality.
Atkinson's and related approaches to inequality measurement
have several technical features which I have discussed elsewhere,
14
This was also the basis of the famous 'Dalton's principle of transfers' (discussed
earlier), which asserts that given other things a transfer of a unit of income from a
richer to a poorer person must increase social, welfare and be seen as reducing the
normative measure of inequality. The principle has been extensively used in the
literature on the normative measurement of inequality (see Atkinson 1970*, 1975,
1983), See also Dasgupta, Sen, and Starrett (1973) and Rothschild and Stiglitz
(1973).
15
This requires, in fact, that the utility functions have ratio-scale measurability
and comparability, which is a very demanding assumption. On this, see Sen (1977ft,
1986«), Gevers (1979), Roberts (19800), Blackorby, Donaldson, and Weymark
(1984), d'Aspremont (1985), among others.

Welfare Economicsand Inequality 9?
and on which I shall not comment further here,1
* Atkinson's
approach has the advantage, over the Dalton formulation, of not
demanding any more measurability or comparability of utilities
than is already involved in the characterization of social welfare—-
the indexingoperations being all in the space of incomes. TheAtkin-
son approach is also more general in permitting non-utilitarian
social-welfare functions, even though (as we discussed earlier) he
actually chose to see social welfare as the sum-total of individual
M-values (usually taken to be individual utilities, but open also to
other interpretations). Indeed, there isno difficulty withinthe Atkin-
son approach in changing the form of the social-welfare function, as
long as it is ensured that (I) social welfare is ultimately a function of
the vector of incomes only, and (2) for any total income, an equal
distribution must be the social-welfare maximizing distribution of
that total income,11
The merits of this approach are clear enough, and the class of
Atkinson indices (with various assumptions about the relation
between individual incomes and social welfare) has been widely
used not only in the literature of normative measurement, but also
in public economics in general. In particular, the idea of an 'equally
distributed equivalent income* links inequality measurement
directly to the evaluation of public policy in a most usable way.
There are, however, several serious problems as well. First, since
inequality has descriptive as well as normative content, a purely
normative approach to inequality can go against certain obvious
intuitions regarding inequality.For example, if individual utility is
a linear function of individual income (i.e. if marginal utility is
constant), then with a utilitarian social-welfare function, 'the
equally distributed equivalent income* would be the same as the
actual income (since there would be no aggregate-utility loss from
inequalities in income distribution). In this case the Atkinson index
of inequality, which identifies the social-welfare loss from in-
equality with inequality itself, would declare that every distribution
of incomes—no matter how unequal—in fact has a zero level of
'* They were explored particularly in ch. 3 of my OEI. See also Kota (1976),
Blackorby and Donaldson (1978, 1984), Atkinson (1983), Foster (1985). The prob-
lems discussed in what follows apply generally to the whole class of normative
measures of inequality; I specificallyconcentrate on Atkinson's analysis because of
its clarity and its pre-eminent position in the literature on the subject,
" In OEI, such a generalized form was presented. See also Kolm (1976) and
Blackorby and Donaldson (1978, 1984).

inequality, A distribution of incomes (1,99) would be, then, seen as
perfectly equal, in the same way as a (50, 50) distribution would be
viewed. This would run somewhat contrary to the fact that both in
terms of incomes and in terms of utilities, a (I, 99) distribution is
quite unequal,
More importantly, the Atkinson index of inequality can move in
the opposite direction to the actual inequality of individual utilities,
as we take less and less 'concave' utility functions, i.e. take marginal
utility to diminish more slowly with increased income.18
When the
marginal utility goes down comparatively more slowly, the utility
gap related to an income gap is more than it would otherwise have
been, so that the directly observed level of utility inequality related
to a given pattern of income inequality is larger than otherwise. On
the other hand, with slowly diminishing marginal utility the loss of
aggregate utility as a result of the income inequality is less, and thus
the Atkinson measure of inequality (and the Dalton measure too) is
lower, rather than higher. This has the somewhat perverse feature
that—for any given income distribution—the more the inequality in
the utilities that people enjoy, the lower is the index of inequality,"
The Atkinson index moves (goes down) in a way contrary to the
directly observed income inequality (which is stationary) and in a
way contradictory to the directly observed utility inequality (which
goes up).
The 'perversity' may or may not be in itself particularly disturb-
ing, but it draws our attention to a featureof the normativeapproach
to the measurement of inequality that is important to appreciate.
The Atkinson index and other normativemeasures of inequality are,
in fact, measures of the distributional badness—according to the
chosen social-welfare function—of the particular configuration of
personal incomes; they are not specifically measures of inequality
per se—neither of incomes, nor of utilities. With a given configura-
'* This interpretation is possible if the 'utility function' is seen as giving the values
of individual utilities, rather than as a purely analyticaldevice to express the social
welfare as an additively separable function of individual incomes (y-), with the utility
«(FJ) not standing for any particular feature of the person i. On these issues, see
Atkinson (1983),
" This discussion maybe a little too compressed, and it certainly could be aided
by the diagrammatic version presented in Sen (19786), and also in Sen (1984). The
philosophical issues are best discussed in Bengt Hansson (1977); see also Atkinson
(1983), Since this is not a central issue in the context of this monograph, I am resisting
a fuller discussion, since it might defect attention towards a side-issue that is related
and relevant but by no means crucial to the main themes of this monograph.

tion of personal incomes (and thus, in one obvious sense, a given
inequality of income distribution), a less concave utility function
can make the Atkinson index go down—precisely when different
persons' utilities move further apart and the inequality of utilities
goes up.
There is nothing really perverse in all this, if the Atkinson index is
seen as a measure of'how bad' the income inequality is, according to
the chosen social-welfare function (i.e. how much loss of aggregate
welfare, or of equally distributed equivalentincome, there is). In the
example considered, the utility inequality may be larger, but the
utility sum-total is closer to the maximal value, thanks to the slowly
diminishing marginal utility. And the Atkinson index would be
right to declare this to be a case of low index value of the loss of
aggregate welfare.
The immediate context of the Atkinson approach, which has been
widely used in welfare economics and in public-policy analysis, is
well-defined, viz. the assessment of social-welfare implications of
income inequality. The interest in income inequality in the Atkinson
approach, is mainly in that context. When there is slowly dimin-
ishing marginal utility, there is less inefficiency from unequal dis-
tribution of incomes, and it is this inefficiency in generating social
welfare that the Atkinson index really measures, While it may be
somewhat misleading to call this an 'inequality index' as such, the
index relates very well to Atkinson's purpose of measuring the
social-welfare loss from inequality,
The second issue concerns the characterization of the social-
welfare function. Atkinson chose an additive form, and earlier on
Dalton had chosen specifically a utilitarian social-welfare func-
tion, with no concern whatever for the distributional inequality of
utilities. These features can be changed, and it is in fact even possible
to drop the additive form, in addition to eschewing utilitarianism,20
But the framework requires, as we stated earlier, that (1) social
welfare must be a function, ultimately, of incomes alone, and (2)
social welfare must be maximized by an equal distribution of in-
comes, given the total income, I have already discussed why these
assumptions are disputable. Indeed, this monograph is much con-
cerned withgetting away from the limiting perspective oninequality
that isprovided by these—and related—assumptions, widelyused in
welfare economics.
20
See Sen (1973a), Atkinson (1983), Blackorby and Donaldson (1984),

If social welfare is seen as a function of individualwell-beings,
then variations in the conversion of incomes into well-beingmust be
taken into account, with adequate attention being paid to the vari-
able relationship between incomes, on the one hand, and function-
ings and capabilities, on the other. These conversion factors in-
luencing the relation of income to well-being must also be brought
in. This can raise questions about assumption (1). Even if the para-
meters of the conversion rates are respectively given, an equal dis-
tribution of income may yield very unequal well-being levels—with
differences related to group-specific parameters (such as gender, age,
environments) and individual parameters (such as genetic charac-
teristics). It is easy to envisage circumstances with wide interper-
sonal variations in conversion rates, in which an equal distribution
of income would be quite bad for social welfare, so that assumption
(2) may be quite unacceptable,21
These problems arise even with retaining the basic assumption
that social welfare is a function of individual well-beings. When
that assumption itself is dropped, and direct attention is paid to the
freedoms enjoyed by the people (in line with the reasoning presented
earlier in this monograph), both the assumptions needed for the
approach would face more radical difficulties.
So the conclusion wemust come to is perhaps something like this.
The Atkinson approach of inequality measurement is indeed very
useful within a fairly limited format in which individual diversities
are left out of account. The format makes no room for substantial
interpersonal variations in the conversion of individual incomes
into personal well-beings, and furthermore it does not accommo-
date the importance of freedom as a constitutive element of a good
society (and thus as a determinant of what is called social welfare).
However, since that very format is standardly used in mainstream
welfare economics and in standard public-policy analysis, the use of
the Atkinson index does not add to the limitations already present in
the tradition. In fact, it makes the accounting more systematic and
efficient within that standard structure. So the contingent usefulness
of the approach is not in doubt. The fact that it may be somewhat
misleading to call the Atkinson index a measure of inequality as
11
In these circumstances 'Dalton's principle of transfers' would also be, in
general, uasupportable in the context of social-welfare maximization. For example,
the transfer of some income from an ill person to one who isin robust health but with
a little lower income may be far from a social improvement.

such (as opposed to an indexof 'distributional badness' taking note
also of efficiency considerations involved in generating social wel-
fare) does not detract from the usefulness of that index. But we have
to be clear as to what exactly we are doing, and why.
If the fundamental fact of human diversity and its far-reaching
implications come to be recognized more widely in welfare-
economic analysis and in public-policy assessment, then the
approach would certainly need some radical transformation. The
operations would have to move from the income space to the space
of the constitutive elements of well-beingand also of freedom, if the
intrinsic importance of freedom, discussed earlier, is accepted.
Social-welfare analysis would then take a different form, and the
evaluation of inequality and of distributional badness would then
have to reflect that foundational transformation.
In the next chapter, when we examine poverty measures, we shall
see that something similarlycritical can be said also about the cur-
rently used poverty indices.22
These standard measures are all basic-
ally parasitic on the traditional concentration on the income space
and ultimatelyon ignoring the fundamental fact of humandiversity
and the foundational importance of human freedom.
21
Including, I fear, the so-called Sen index (Sen !9?3c, 1976a).

7
POVERTY AND AFFLUENCE
7.1. INEQUALITY AND POVERTY
The mainstream approach to identifying poverty specifies a cut-off
'poverty line*, defined as the level of income below which people are
diagnosed as poor. The conventional measure of poverty, still
widely used, proceeds from here to count the number of people
below the poverty line—the so-called 'head count'—and defines the
index of poverty as the proportion of the total population that
happens to be below the poverty line (i.e. the fraction of the popu-
lation identified as poor). This gives a neat and well-defined
measure, and it is not hard to see why it has been so widely used in
the empirical literature on poverty and deprivation.
The measurement of poverty can be seen as consisting of two
distinct—though interrelated—exercises, viz, (1) identification of the
poor, and (2) aggregation of the statistics regarding the identified
poor to derive an overall index of poverty. In the traditional 'head
count' approach, the identification exercise is done through the use
of the 'poverty line' income as a cut-off. Then, the aggregation is
done simply through counting the number of the poor, and
calculating the proportion H—the 'head-count ratio'—of people
below the poverty line. Both the exercises are, in this approach,
thoroughly dependent on seeing deprivation in terms of low income
as such.'
But, in addition, the aggregation exercise done through simple
head counting pays no attention to the fact that people could be a
little below the line, or a lot, and also the distribution of income
among the poor may or may not be itself very unequal. It is this
lacuna in using H as a measure of poverty that has received a tre-
mendous amount of attention in the recent literature on formal
measures of poverty. I shall first discuss that development in the
'aggregation* exercise, before turning, in the next section, to the first
1
In principle, the 'head count" measure can be used with some other way of
identifying the poor, rather than relying only on low income. In practice, however, it
has been used almost invariably only with a low-income dividing line.

Poverty and Affluence 103
question, viz. the use of low income as the main characteristic of
poverty, which influences both 'identification' and 'aggregation*.
There is, in fact, another aggregate statistic, like the 'head count*
ratio, that has also been used—though not quite so widely—in the
traditional literature.2
This is the so-called 'income gap', which
measures the additional income that would be needed to bring all
the poor up to the level of the poverty line, i.e. the minimal extra
income that would be sufficient to wipe out poverty—in the form of
low income—altogether. This 'gap* can be expressed in per capita
terms, viz. the average shortfall / of income of the identified poor
from the poverty line,3
Note that just as the head-count ratio H is completely insensitive
to the extent to which the incomes of the poor fall short of the
poverty line and takes note only of the number below the line, the
income-gap ratio / iscompletely insensitive to the number of heads
involved and takes note only of the average gap of the income of the
poor from the poverty line. It is natural to think that the two must
complement each other, since they address different aspects of
poverty. The need to put H and / (or similar indices) together is,
thus, obvious enough,
It can now be asked whether the two together would provide an
adequate informational base for poverty measurement (still
sticking to the idea that poverty is best seen as low income)! The
answer, briefly, is: no. H and / together still cannot be adequate,
since neither pays any attention to the distribution of incomeamong
the poor. For example, a transfer of income from a poor person to
one less poor but also below the poverty line (before and after the
transfer) would leave both the values of If and / completely un-
changed. But it can certainly be argued that aggregate poverty is
increasedby this transfer, since thepoorer person isevenpoorernow,
and this intensification of the more acute deprivation cannot be
outweighed by the increase in the income of the person who was less
poor even to start with. Hence we need some other statistic, pre-
sumably some measure of inequality in the distribution of income
among the poor, Let us call such a measure of inequality among the
poor D,
2
See Beckennan (1979), Anand (1983), Beckennan and Clark (1982).
5
There are also other ways of normalizing this, such as a proportion of national
income needed to wipe out the poverty gaps of every identified poor person. See
Anand (1977), Beckennan (1979).

104 Poverty and Affluence
It Is easy to construct an axiomatic derivation of a poverty
measure that is sensitive to all these three related but distinct con-
siderations (still sticking to the idea that poverty is best seen as low
income). This would make the axiomatically derived aggregate
poverty measure P a function of H, I, and D, The axioms that I had,
in fact, presented in Sen (1973c, 1976a) led to the identification of
the inequalitymeasure D of the distribution of income among the
poor as the Gini coefficient Or, and it thus led to a poverty measure P
that depended on H, I, and Of.
The axioms, postulated in a specified format of measurement,
included accepting the informational sufficiency of H and /together
in the special case in which all the poor had the same income (so
that the question of inequality among the poor would not, then,
arise),4
In fact, in this special case the poverty measure was simply
taken as the product of the two, i.e. HI. The axioms also demanded
that whensome of the poor are poorer than others, then the per-unit
weighting of the income shortfall of each poor person must increase
with his or her poverty rank, i.e. the poorest poor would have the
highest weight and the richest poor the least. Taking the special case
of 'rank-order weighting' giving a weight of « to the income
shortfall of the «-th richest among the poor, a very specificmeasure
of aggregate poverty emerges, which takes account of the inequality
of the distribution of income among the poor through the use of the
well-known Gini coefficient.5
Rank-order weighting has been extensivelyused in social-choice
theory beginning with the classic use of this 'ordinal' approach to
voting decisions by the French mathematician Borda (1781), and
this procedure—and the general 'positional' approach underlying
it—provide useful ways of 'weighting' competing claims with
simple ordinal information.* It is interesting—and rather useful
from a practical point of view—that the Borda method of weighting
leads to the Gini coefficient in the special case of inequality
4
This is not at all a demanding restriction; on this see Foster (1984; n. 27),
J
The exact formula isthe following; P = H [/ + (1 -I) G. For proper statements
of the axioms and the proof of the theorem deriving this exact measure, sec Sen
(1976o). Note that in the special case in which all the poor have the same income, G
would be zero, and thus P would equal HI,
6
On the use of rank-order method in social choice theory, see Arrow (1951),Sen
(]970a), Suzuiuura (1983). In the context of measurementproblems to economics and
development studies, the possibility of using the positional approach in general and
the Borda rank-order method in particular has been explored in Sen (1974, 19?6a,
19766, 19816).

measurement, which is one of the most common measures of income
inequality,7
The so-called Sen measure of poverty and related distribution-
sensitive indicators have been used inter alia in many practical exer-
cises in evaluating poverty, and these empirical works—related to
Bangladesh, India, Iran, Malaysia, United States, Brazil, and
several other countries—have been done with much care and skill,"
Despite my scepticism of the appropriateness of the so-called Sen
measure and other indicators that diagnose poverty in terms oflow
income, I would not dispute the analytical and practical contri-
bution made by these empirical works through the use of distribu-
tion-sensitivity in evaluating poverty.
The concentration on the incomespace isoften hard to avoid given
the comparatively greater availability of income statistics rather
than other types of data.* Within that informational format, the
traditional use of the head-count ratio as a measure of poverty can
deflect anti-poverty policy by ignoring the greater misery of the
poorer among the poor. Indeed, with the head-count ratio as the
measure of poverty, any government faces a strong temptation to
concentrate on the richest among the poor, since that is the way that
the number of the poor—and the head-count ratio H—can be most
easily reduced.I0
Recasting the empirical measurementof poverty in
7
On the characteristics of the Gini coefficient, see OEI, ch. 2. On the properties of
theOini coefficient, Lorenz-curve comparisons, and related matters, seeGraaff (1946,
1977, 1985), Koto (1969), Atkinson (1970ft), Newtay (1970), Sheshinski (1972), P.
Dasgupta, Sen, and Starrett (1973), Sen (1974, 19766), Pyatt (1976, 1987), Cowell
(1977), Btaekorby and Donaldson (1978, 1980), P. J. Hammond (1978), Yitehaki
(1979), Kakwani (1980*), Roberts (19806), P. K. Sen (1986), SeidI (19860).
8
SeeAhluwalia(1978), Alamgir(1978), Anand (J977,1983), Bhatty (1974), Sastry
(1977), Sesstrand and Diwan (1975), Clark and Hemming (1981), Szal (1977),
Dntta (1978), Fields (1979, 1980a), Fishlow (1980), Gaiha and Kazrni (1981), van
Gitineken (1980), Kakwani (19806, 1981, 1986, 1988), Sundaram and Tendulkar
(1981), Gsmani (1982), Pantelu (1980), Sastry (19800, 1980*), Hemming(1984), Ray
(19846), Gaiha (1985), Babu (1986), and others. There have been a number of other
important empirical studies in similar lines in more recent years.
* This applies even to the use of income as opposed to expenditure statistics. The
case for using the latter on grounds of greater relevance is often very strong, but
actual ute is nevertheless restricted because of the difficulty of obtaining reliable
expenditure information. On this see Atkinson (1989: ch. I).
" On therelevance of this type ofconsideration for the Indian debates on poverty,
see Sen (1973c), Ahluwalia(1978), Dutta (1978), S. R. Chakravarty (1981), I, Chak-
ravarty (1986). The policy implications of the neglect of distributional issues in
insensitive measures of deprivation is, in fact, a pervasive problem which crops np in
different contexts. For example, if the extent of homelessness isjudged just by the

a distribution-sensitive way has the effect of making comparatively
better use of income data (despite the overall limitations of that
income-based informational base).
The need for having distribution-sensitivity in measuring poverty
seems to be fairly widely accepted by now, and various otherdis-
tribution-sensitive measures of poverty have also been suggested in
the theoretical literature, taking some other inequality measure D
among the poor, and using other mathematical forms of combina-
tion." I shall not comment on the relative merits of these different
measures and assess the distinct ways of taking note of the three
aspects of the problem of deriving an aggregate measure of
poverty.l2
The major issue in the context of measuring poverty in the
income space seems to me to be the need to pay attention to all the
three aspects—particularly to incorporate distribution-sensitivity
(rather than insisting that this incorporating be done in some very
specific way).l3
And the primary issue in the more general context,to
be addressedin the presentstudy, is the questioning of the relevance
of the income space itself for the measurement of poverty.That
number of people without a proper home, the temptation to deal with more easily
remediable cases—irrespective of the extent of homelessness or the intensity of the
misery resulting from it—can be very strong.
11
See particularly Anand (1977, 1983), L. Taylor (1977), Drewnowski (1978),
Hamada and Takayama (1978), Takayatna (1979), Thon (1979), Blackorby and Don-
aldson (1980), Fields (1980a), Kakwani (1980a. 1980*, 1981), Sastry (I9Wa, 19806),
S. R. Chakravarty (1981, 1983«, 19836), S. Clark, Hemming, and Ulph (1981),
Qsmani (1982), Kundu and Smith(1983), Foster, Greer, and Thorbeeke (1984), Foster
(1984), Ray (19840), Bigman (1985, 1986), Upton (1985), Coweil (1986), Donaldson
and Weymark (1986), Jorgenson and Slesnick (1986), Seidl (I986o), Atkinson (1987,
1989), Lewisand Ulph (1987), Pyatt (1987), R, N. Vaughan (1987), Besley and Kanbur
(1988), Buhmannirta/. (1988), Foster and Shorrocks (19880,19886,1991), Ravallion
and van de Walk (1988), Srneeding, Rainwater, and O'Higgins (1988), Bourguignon
and Fields (1990), Pattanaik and Sengupta (1991), among other contributions,
12
One issue that has received much attention is the insistence on "separability* in
poverty measures, Separability is certainly convenient property, and permits us to
build up the overall poverty picture from the poverty measures applied to subgroups.
The requirement has much cutting power. There is a helpful technical literature on
this and related issues, including their implications (see Anand 1983; Foster 1984;
Foster, Greer, and Thorbecke 1984; Coweil 1986; Foster and Shorroeks 1991). There
remains a more general question as to whether it is sensible to assume that poverty
indicators should be combinable in this way,which requires that the view of poverty
for particular groups be, in some specific ways, insensitive to what happens to other
groups, and that the whole picture does not introduce anything other than what is
already there in the parts,
IJ
See, however, the critical and comparative analyses presented by Foster (1984)
and Seid!(I986«), and also the general methodological critique presented by Atkin-
son (1987, 1989). See also Sen (1981a, 1983o%

critical examination would apply to all the different measures that
share this concentration on seeing poverty in terms of low income,
7.2, THE NATURE OF POVERTY
Consider two persons 1and 2—person 1has an income level some-
what lower than that of 2, But 2 has a kidney problem and needs to
use a dialysismachine which costs him a lot, and he has also a much
more impoverished life than person 1, Who is the poorer of the
two—person 1because his income is lower, or person 2 because his
capability set is more restricted?
The question may sound like one of pure semantics. It mighteven
be tempting to take the view that it does not matter at all whomwe
call 'poorer', as long as we define our terms clearly enough. That bit
of philosophical 'nominalism* does make some sense, but the fact
also remains that 'poverty* is a major evaluative concern in most
societies, and how we identify poverty is a matter of some practical
moment in the contexts in which questions of this kind are posed. So
here wedo have a substantive issue. Whilethe term poverty isused in
rather different ways, there are some clear associations that con-
strain the nature of the concept, and we are not entirely free to
characterize poverty in any way we like,
The question can be given both descriptive and policy forms. In the
first view, the identification of poverty is an acknowledgement of
deprivation. It may lead to a policy recommendation also, but that
is a derived feature, and the first exercise is one of deciding who are
truly deprived as these things are judged in the society in question.
The second viewsimply identifiespoverty with a policy recommend-
ation, viz. an assertion that something should be done by society to
encounter these deficiencies.14
In the second view poverty is pri-
marily a matter of identifying the focus of public action, and its
descriptive meaning is only derivative. In contrast, the first view
makes the description primary and the policy conclusion derivative.
Again, it may be a mistake to spend much time on sorting out
which of the two views to take. I have tried to argue elsewhere that
there is a case for sticking to the first—primarily descriptive—view,
so that diagnosis precedes policy choice.15
This is also important in
14
See Beckerman (1979) for a discussion ef the policy-related view of poverty.
ls
The arguments are presented in Sen(I979e, 19810). Ihave to confess thatwhile I
haven't changed my position on this, I don't now attach as much importanceto the
contrast as I then evidently did.

order to make sure that the non-availabilityof public resources to
help eliminate severe deprivations should not make us redefine
poverty itself. For example, if the state and the society lack the
means to alleviate extreme economic hardship, that would be a
conclusive reason against a policy recommendation to counter that
deprivation through using the necessary—but non-available—
funds.16
But that fact in itself should not make us decide that there
isn't much poverty around (as wewould be obliged to say, ifwe were
to define poverty entirely in terms of the recommended choice of
policy).
A policy recommendation is contingent on feasibility, but the
recognition of poverty has to go beyond that. One can argue that the
firststep is to diagnose deprivation, and related to that, to determine
what we should do if we had the means. And then the next step is to
make actual policy choices in line with our means. In this sense, the
descriptive analysis of poverty has to beprior to the policy choice.
But how should that descriptiveanalysis of poverty proceed? Just
because it isa primarilydescriptive exercise, weshould not make the
mistake of thinkingthat the analysis must be somehow independent
of the society in whichpoverty is beingassessed. Eventhe demandof
'objectivity' in description does not really require social invariance,
as it is sometimes supposed. What is seen as terrible deprivationcan,
of course, varyfrom society to society, but from the point of view of
the social analyst these variations are matters of objective study.17
We could, of course, debate about the exact ways in which norma-
tive judgements should take note of such social variations, but the
primary exerciseof diagnosingdeprivation cannot but besensitive to
the way various types of hardships are viewed in the society in ques-
tion. To deny that connection is not so much to be super-objective,
but to be super-dense.
The existence of social variations does not, of course, rule out
various agreements on what is to count as serious deprivation. In-
deed, it can be argued that ifweconcentrate on certain basic general
functionings and the corresponding capabilities, there may bemuch
more agreement on their importance, than there would be if we
concentrated on particular commodity bundlesand particular ways
of achieving those functionings. For example, there is likely to be
14
This claim relates to the old dictum: 'ought implies can'. On this philosophical
issue, see Hare (1952).
11
This issue has been discussed in Sen,(19806).

more intercultural—and also interpersonal—agreement on the im-
portance of having the capability to avoid acute hunger or severe
undernourishment, than on the significance of having an adequate
supply of particular food items (e.g. some specific types of meat or
fishor grains or pulses) to serve those funetionings. To take another
type of examples, there may be more agreement on the need to be
entertained, or to have the capability to take part in the life of the
community, than on the form that entertainment must take, or on
the particular way the life of the communitymay be shared.l8
This is, in fact, one reason why poverty Is better seen in terms of
capability failure than in terms of the failure to meet the 'basic
needs* of specified commodities. The 'basic needs' literature and the
related studies on the 'quality of life' have been enormously helpful
in drawing attention to deprivations of essential goods and services,
and their crucial role in human living.19
The underlying motivation
can perhaps be more directly addressed in terms of achieving certain
basic funetionings and acquiring the corresponding capabilities.20
In
so far as the underlying reasoning of the basic-needs approach re-
lates to giving people the means of achievingcertain basic function-
ings, the problem of interpersonal variations in 'transforming* com-
modities into funetionings—discussed earlier—can also be avoided
by directly looking at the functioningspace rather than at the com-
modity space.21
7.3. LOWNESS VIS-A-VIS INADEQUACY OF INCOMES
In linewith the preceding reasoning, it is possible to argue for seeing
poverty as the failure of basic capabilities to reach certain
minimally acceptable levels.22
The funetionings relevant to this
18
These issues are discussed fa Sen (1980&, 1981a, 1983d). The basic point of social
variation of form related to the same general functioning goes back, in fact, to Adam
Smith (1776). There is some similarity here with the Aristotelian position on 'non-
relative virtues', on which see Nussbaum (1988ft).
19
For an excellent discussion of the general approach, see Streeten gt al. (1981).
On related issues, see also Pant et at, (1962), Adelmanand Morris (1973), Sen (19734
1981o), P. Bardhan (1974ft, 1984), Adelman (1975), Grant (1978), Morris (1979),
Chichilnisky (1980), P. Dasgupta (1986), Dreze and Sen (1989), UNDP (1990, 1991),
Desai, Boltvinnik and Sen (1991),
» On this general question, seeSen(1984,1985ft), Streeten (1984), Stewart (1988),
Griffin and Knight (1989).
JI
On these issues, see Griffin and Knight (1989).
22
On this see Desai (1990) and Hossain (1990). See also Sen (19800) and Griffin
and Knight (1989).

analysis can vary from such elementaryphysicalones as being well-
nourished, being adequately clothed and sheltered, avoiding pre-
ventable morbidity, etc., to more complex social achievementssuch
as taking part in the life of the community, being able to appear in
public without shame, and so on. These are rather 'general' func-
tionings, but—as was discussed earlier—the specific form that their
fulfilments may take would tend to vary from society to society,
This capability-based approach to povertycan becontrasted both
(1) with the view of poverty as low utility,and (2) with seeing poverty
in terms of low income (or, more generally, low holdingofprimary
goods or of resources).23
The analyses presented in the earlier
chapters indicatewhyneither utilitynor income (nor primary goods,
nor resources) can be identified withwell-being as such. But that fact
is not decisive in determiningthe appropriate approach to poverty,
In particular, since the concept of poverty has a well-recognized
link with deprivation caused by economic problems, the approach
of poverty as 'low income* needs further consideration,
It can be argued that poverty is not a matter of low well-being,but
of the inabilityto pursue well-being precisely because of the lack of
economic means. If Mr Riehmao has a high income and can buy
whatever he needs, and stillsquanders the opportunities and ends up
rather miserable, it would be odd to call him 'poor*. He had the
means to live well and to lead a life without deprivation, and the
fact that he managed neverthelessto generate some deprivationdoes
not place him among the poor. This way of analysingthe problem
would seem to get one some distance towards seeing povertyin terms
of income deprivation after all.
That lineof reasoning certainlyhas some merit. It does indeedget
us 'towards' seeing poverty in terms of incomedeprivation, but does
not take us quite there. There are other distinctions to be considered.
Perhaps the most important point to note is that the adequacy of the
economic means cannot be judged independently of the actual
possibilities of'converting' incomes and resources into capability to
function. The person with the kidney problem needing dialysis (in
the examplediscussed earlier in thischapter) may have more income
than the other person, but he is stillshort of economicmeans (indeed
of income), given his problem in converting income and resources
into functionings.If we want to identify poverty in terms of income,
it cannot be adequate to look only at incomes (i.e. whether it is
23
On related matters, see Goodin (1985, 1988).

generally low or high), independently of the capability to function
derivable from those incomes. Income adequacy to escape poverty
varies parametrically with personal characteristics and circum-
stances,24
The basic failure that poverty implies is one of havingminimally
adequate capabilities, even though poverty is also inter alia a matter
of inadequacy of the person's economic means (the means to prevent
the capability failure), Consider the example touched on earlier of
the person with a high metabolic rate, or a large body size, or a
parasitic disease that wastes nutrients. He is less able to meet
minimal nutritional norms with the same level of income, corn-
pared with another person without those disadvantages. If he is to be
seen aspoorer than the second person, despite the fact that both have
the same income, the reason for this lies in his greater capability
failure (the focus of our concern). The same set of facts can also be
seen as indicating the greater inadequacy of his income given his
personal characteristics and circumstances. To have inadequate in-
come is not a matter of havingan income level below an externally
fixedpoverty line, but to have an income below what is adequate for
generating the specified levels of capabilities for the person in
question,
In the income space, the relevant concept of poverty has to be
inadequacy (for generating minimally acceptable capabilities),
rather than fewness (independently of personal characteristics).25
A
'poverty line' that ignores individual characteristics altogether
cannot do justice to our real concerns underlying poverty, viz,
capability failure because of inadequate economic means. Often it
will make sense to group individualsinto particular categories (re-
lated to class, gender, occupational group, employment status, and
so on).26
If wechoose to express poverty in the income space, then the
incomes needed would have to be linked to the causal requirements
of minimal capabilities.
This question relates to the fact that the primary concern in
poverty analysis is with capability to function, rather than with
achieved functionings. The example, discussed earlier, of the person
with means whofasts out ofchoice, as opposed to another who hasto
24
On this, sec also Hinuneifarb (1984).
25
Rowntree (1901, 1941) himself was much concerned with the issue of income
inadequacy—and not just with its fewness—in his pioneering studies of poverty in
Britain.
16
The use of such categories is discussed in Ch, S,

starve because of lack of means, is relevant here. Both may end up
starving and fail to be adequately nourished, but the person without
the means—and thus without the capability to be adequately nour-
ished—is poor Jo a way that the fasting person isnot. So the focus of
attention of poverty analysis has to be capability as opposed to
achievement (even though wemay sometimes use information about
achievement to try to surmisethe capability enjoyed by a person).27
AM this is close to the issue of the distinction between freedom
and resources examined earlier (in Chapters 2 and 5), Resources are
important for freedom, and income is crucial for avoiding poverty,
But if our concern is ultimately with freedom, we cannot—given
human diversity—treat resources as the same thing as freedom.
Similarly, if our concern is with the failure of certain minimal capa-
bilities because of lack of economic means, we cannot identify
poverty simply as low income, dissociated from the interpersonally-
variabte connection between income and capability. It is in terms of
capability that the adequacy of particular income levels has to be
judged.
7,4 DOCONCEPTS MATTER?
The idea of Income inadequacy*, as discussed in the last section,
goes well beyond that of 'low income' as such, since the former is
sensitive to the conversion of incomeinto capability in a waythat the
size of income as such cannot be. When the ranking of incomes goes
opposite to the relative advantages in converting income into
capability, the ordering of poverty and the identification of the poor
may be very different if it is done entirely in terms of the size of
income (as is the standard practice in most countries) compared with
what it would be if the focus is on capability failure.
The problem is particularly serious with specific types of depriva-
tion. For example, both because of biological reasons and social
27
The fact that weare not classifying as 'poor" a person who has the capabilityto
achieve good nourishment, but chooses not to, should not, of course, be taken to
imply that such a person's deprivationshould not command any sympathyor atten-
tion. As wasdiscussed earlier (in Ch. 4), a person may givepriority to non-well-being
goals in his agency objectives, but that does not entail that others cannot have good
reason to try to raise his low level of achieved well-being. Poverty is not the only
reason for sympathetic regard from others. For example, while Mahatma Gandhi
clearly had excellent reasons for fasting indefinitely in protest against thecommunal
riots of 1947, his friends and well-wishers also had good reasons for tryingto see that
this did not lead to his fate! debilitation.

factors (especially as they operate with a resilient tradition of—
explicit or implicit—sexism), women may have special disadvan-
tages in converting income into particular functioning^. To consider
a varietyof different types of cases, such disadvantages may apply to
the capability of being nourished (e.g. because of the demands of
pregnancy and neonatal care), achieving security (e.g. in single-
parent families),having fulfilling work (e.g. because of stereotyping
of'women's jobs'), establishing one's professional reputation early
on in one's career (e.g. because of the asymmetricdemands of family
life). The extent of deprivation may be underjudged if we concen-
trate only on the size of incomes, and the need to bring incapability
failures explicitly can be particularly acute in such cases.
Similarly, the relationship between income and capability would
be strongly affected by age (e.g. by the specific needs of the old and
the very young), by location (e.g. by the special challenges to safety
and security in urban living), by epidemiological atmosphere (e.g.
vulnerability to diseases endemic in a region), and by many other
parameters.28
By focusing poverty study specifically on incomes as
such, crucial aspects of deprivation may be entirely lost.2
*
Sometimes the same handicaps, such as age or disability or ill-
ness, that reduce one's ability to earn an income, can also make it
harder to convert income into capability. Often, a high proportion
of the poor in the advanced countries have such handicaps,30
and the
extent of poverty in such countries is substantially underestimated,
since it overlooks the 'coupling* of 'mcome-earning handicap and
income-using handicap in generating capability. For example, an
old person has a much harder time in being free from disease, in
leading a healthy life, in achievingmobility,in taking part in the life
of the community,in seeing friends, and so on.31
And these income-
usirtg disadvantages can tremendouslycompound the feature of low
M
For analyses of some of the factors involved in the context of the UnitedStates,
see the papers included in Danziger and Weinberg (1986)', Illuminating analyses of
different types of non-income variables influencing the extent of deprivation of the
vulnerable in the USA can be found in Palmer, Smeeding, and Torrey (W88), Case
and Kate (1990), and several other recent studies.
'* In the "Scandinavian studies on living conditions' the informational base has
been widened to include important functionings (rather than just examining income
and opulence), and the empirical works demonstrate how big a difference is madeby
this dcpartyre. On this, sec Altardt (1981) and Eriksoa and Aberg (1987), and the
literature cited there, and also Allardt (1992), Erikson (1992) and Ysander (1992).
30
See Atkinson (19700) and Townsend (1979).
31
See Wedderbutn (1961), Townsend (1979), Palmer, Smeeding, and Torrey
(1988), Lasiett (1991), among others.

earning power which is the only one that the traditional focus on
income-based poverty analysis manages to capture,
7.5. POVERTY IN KJCH COUNTRIES
Even the causes of the persistence of hunger in rich societies cannot
be fully understood if we confine our attention only to the size of
incomes. Hunger in the United States is associated with many para-
meters of which low income is only one.32
The health aspects relate
to the social environment, to the provision of medical care, to the
pattern of family life, and a variety of other factors, and a purely
income-based analysis of poverty cannot but leave that story half
told.33
The extent of capability deprivation can be quite remarkably high
in the world's most affluent countries. For example, a study by
McCord and Freeman (1990), presented in The New England Journal
of Medicine, indicates that men in the Harlem region of the pros-
perous city of New York have lesschance of reaching the age of 40 or
more than Bangladeshi men have. This is not because the residents
of Harlem have lower incomes than the average Bangladeshi does.
The phenomenon is more connected with problems of health care,
inadequacy of medical attention, the prevalence of urban crime, and
other such factors that affect the basic capabilities of the Harlem
resident.
The problem is not confined only to 'pockets* of deprivation in a
small number of places. There are systematic patterns of intense
inequality in non-income features between different groups. For
example, in an article in The Journal of the American Medical
Association,Otten et a/, (1990) show that in the age group between 35
and 55, African-Americans have 2.3 times the mortality rate as do
whites in the United States, and that only about half their excess
mortality can be explained by income differences. The need to go
beyond the information on incomes to the pervasive diversities of
social circumstances and characteristics is well illustrated by the
nature of these terrible problems. The social environment is deeply
33
On this, sec School of Public Health, Harvard University (1985).
53
For interesting empirical analyses related to this genera! issue in the contextof
the richer countries,seevan Praag, Hagenaars, and van Weeren (1982) and Mack and
Lansiey (1985). Seealso the conceptual contributions to poverty study and its welfare-
economic antecedents madeby the Leiden School, e.g. van Praag (1968,1978,1991),
Kapteyn and van Praag(1976).

influenced by the inadequacy of health facilities, the violent modes
of inner-city living, the absence of social care, and such other fac-
tors. The tewness of income is only one factor among many that
influence poverty in the United States,
The issue of food deprivation in rich America also raises a ques-
tion of profound importance in understanding the nature of Ameri-
can poverty. Surprise is sometimes expressed at the fact there could
be any actual hunger in a country as rich as the United States, where
even the poorest groups tend to have much higher incomes than the
middle-classes in many poorer countries who may not be particu-
larly bothered by hunger as such. To some extent the difference may
be due to the fact that money buysless of some types of commodities
in the richer countries.M
But evenafter corrections are made for these
price differences, the paradoxical feature is still retained. Also, as it
happens, food is not one of the items that are typically very much
cheaper in the poorer countries than in the United States.
In explaining the apparent paradox, the capability perspective
can help in two different ways. First, hunger and undernutrition are
related both to food intakeand to the ability to make nutritive useof
that intake, The latter is deeply affected by general health con-
ditions, and that in turn depends much on communal health care
and public health provisions (a subject that will be further
examined in the next section).35
This is precisely where the civic
problems of health delivery and inequalities in health care can
precipitate capability failures in health and nutrition even when
personal incomes are not that low in international standards.
Second, being poor in a rich society itself is a capability handicap
for reasons that I have tried to discuss elsewhere.36
Relativedepriva-
tion in the space of incomes can yield absolute deprivation in the
space of capabilities. In a country that is generally rich, more income
may be needed to buy enough commodities to achievethe same social
functioning, such as 'appearing in publicwithoutshame'.37
The same
applies to the capability of 'taking part in the life of the com-
munity'.38
These general social functionings impose commodity
34
This relates to sharp differences in relative prices, on which see Usher (1968).
35
On this, sec Preze and Sen (1989),
34
See Sen (1983(/), and also the exchange with Peter Townsend on this subject:
Townscnd (1985), Sen (1985e).
Jl
This is an issue that Adam Smith had investigated with great clarity in the
context of discussing the idea of'necessary goods*; see Smith (1776: 351-2).
38
On the importance of this achievement, see Townscnd (1979).

requirements that vary with what others in the community stan-
dardly have.
While the rural Indian may have little problem in appearing in
public without shame with relatively modest clothing and can take
part in the life of the community without a telephone or atelevision,
the commodity requirements of these general functionings are much
more demanding in a country where people standardly use a bigger
basket of diverse commodities,39
Not only does this make it more
expensive to achieve these socialfunctionings themselves, but the
deflection of resources involved in pursuing these social function-
ings also drains the financial means that are potentially usable for
health and nutrition. The apparent paradox of hunger in the rich
countries is not hard to explain once our attention is shifted from
exclusive concentration on the space of incomes, so that wecan take
note of the conversion of income and other resources into capabilities
of various types.
The distinction between low income* and 'capability failure* does
matter, A poverty analysis that concentrates only on incomes can be
quite remote from the main motivation behind our concern with
poverty (viz,the limitation of the lives that some people are forced to
live). It may also fail to provide empirical guidance regarding the
genesis and prevalence of deprivation.40
Concentrating on the right
space is no less important for poverty study than it is for the general
investigation of social inequality,
39
SeeTownsend (1979), Townsend interprets this aspect of poverty as requiring a
'thoroughgoing relativist*approach. On the other hand, it can be argued that while
this variability of commodity requirement shows the relativity of poverty in the space
of commodities and incomes, we are still concerned here with absolute deprivation in
the space of capabilities (e.g. not being able to take part in the life of thecommunity).
This claim (see Sen 1983^) has led to a certain amount of heated—and not invariably
enlightening—debate (see Townsend 1985; Sen 1985rf, Seidl 198&r, Desai and Shah
1988), but the main issue is fairly simple. The absolute-relative correspondence
relates to the variable commodity requirement for the same functioning (e.g. a much
larger need for commodities in richer countries to achieve thesamefunctionings, such
as taking part in the life of the community, or—to go back to Adam Smith's
example—appearing in public without shame). While the minimally acceptable
capabilities to functionmay also vary from society to society, the variable commodity
requirement for the same capabilities does not, to itself, require that we take a
basically "relativist* approach to poverty, provided we see poverty as capability
failure.
48
On this question, see the results of the 'Scandinavian studies in living standards*
presented in Allardt (1981) and Erikson and Abcrg (1987), and the literature cited
there. See also Allardt (1992), Erikson (1992) and Ysander (1992).

8
CLASS,GENDER AND OTHER
GROUPS
8,1, CLASS AND CLASSIFICATION
As was discussed in the very first chapter of this monograph, the
importance of the distinction between seeking equality in different
spaces relates ultimately to the nature of human diversity. It is
because we are so deeply diverse, that equality in one space fre-
quently leads to inequality in other spaces. The force of the question
'equality ofwhafl', thus, rests to a great extent on the empirical fact
of our dissimilarity—in physical and mental abilities and dis-
abilities, in epidentiologieal vulnerability, in age, in gender, and of
course, in the social and economic bases of our well-being and
freedom.
There are diversities of many different kinds. It is not unreason-
able to think that ifwetry to take note of all the diversities, wemight
end up in a total mess of empirical confusion. The demands of
practice indicate discretion and suggest that we disregard some
diversities whileconcentrating on the more important ones. That bit
of worldly wisdom is not to be scoffed at, and indeed, no serious
study of inequality that is geared to practical reasoning and action
can ignore the need to overlook a great deal of our immenserange of
diversities. The question in each context is: What are the significant
diversities in this context?
In fact, general analyses of inequality must, in many cases, pro-
ceed in terms of groups—rather than specific individuals—and
would tend to confine attention to intergroup variations,' In doing
groupanalysis, wehave to pick and choose between different waysof
classifying people, and the classifications themselves select partieu-
1
A distinction has to be made between (!) an intrinsic interest in the inequality
between different groups (¥iewed as groups), and (2) the derivative interest in group
inequality because of what it says about inequality among individuals placed in
different groups. Our focus here is on the latter. On the distinction between the two
approaches and the ethical status of any intrinsic interest in group inequality, see
among others Beteille (I983a, 1983ft) and Loury (1987).

118 Class, Gender and Other Groups
lar types of diversities rather than others. In the literature on in-
equality, the classification that has been, traditionally,most widely
used has been that of economic class—either defined in terms of
Marxian or some similar categories (mainly, concentrating on
ownership of means of production and occupation), or seen in terms
of income groups or wealth categories,2
The importance of this type of class-based classifications is
obvious enough in most contexts. They also indicate why it is the
case that equality in the space of, say, libertarian rights does not
yield anythinglike equality of well-being, or equality of the overall
freedoms to lead the lives that people may respectively value. They
also draw attention to the importance of inequalities in wealth and
income in generating unequal well-beings and living conditions,
even when there is equality in formal procedures and in the allo-
cation of some specific facilities'—which are sometimes called, some-
what euphemistically, 'equality of opportunities*.3
The crucial rele-
vance of such class-based classifications is altogether undeniable in
the context of general political, social, and economic analysis.
The class analysis is also central to the Marxian theory of 'exploi-
tation'. The contrast between some people working hard and getting
little income while others toil little and enjoy high income is one
that has moved social critics to theorize the dichotomy in different
ways. While Marx rejected Proudhon's diagnosis that property was
'theft*, he did outline a system of accounting in terms of effective
labour time that captured the contrast in a descriptively rich way.4
But going further into production analysis, the theory of exploita-
tion involved an identification of who is 'producing* what. Exploita-
2
For examples of recent economic and social analyses making use of the ideas of
class, income, and ownership, see Dalton (1925), K»jzn«ts (1961, 1966, 1973), Lydall
(1966), Atkinson (1972, 1975), Thurow (1975), Edwards, Reich, and Weisskopf
(1986), Dahrendorf (1988). Recent analyses of Marxian class categories include such
diverge contributors as Hobsbawm (1964), Miliband (1977), G, A. Cohen (1978,
1988), Kolakowski (1978), M. Cohen, Nagel, and Scantea (1980), A. E, Buchanan
(1982), Roemer (1982), Marglin (1984), Elster (1986), among many others.
J
On the ambiguities of theconcept of'equal opportunities', seeThurow (1975), Le
Grand (1982), Bayer, Caplan, and Daniels (1983), Beteitle (1983a), Verbs et at.
(1987), Goodin (1988), Van Parijs (19906,1991). On the limitsof what can be achieved
by the standard imperatives of "equality of opportunities' even in a country like the
United States, see Jencks (1972).
* This was one of the uses to which Marx put the "labour theory of value*.
Indeed, both the major theories of value explored in the 19th cent.—the labour
theory andtheutilitytheory—paid much attention to descriptive richness, as opposed
to just predictive usability. See Dobb (1937), Sen (1980ft), Roemer (1982, 19860),

Class, Gender and Other Groups 119
tion wasseen as the enjoymentof one person of the fruits of another's
labour,
Diagnosis of who is producing what in an integrated system of
production is not an easy task, and this is a difficulty that applies
even to the later attempts—on the basis of neo-classical economic
analysis—toattribute to each factor of production a definite share of
the product. Such attribution plays a major part in normative
theories of production-based 'desert', as reflected—in one form or
another—in the writings of J. B, Clark (1902) and Peter Bauer
(1981). Peter Bauer's attack on 'the unholy grail of equality* has
many distinct features, but it turns crucially on the right of the
'producers' to enjoy the fruits of their production (as he puts it, 'it is
by no means obvious whyit should be unjust that those who produce
more should enjoy higher income*).5
The identification of who has produced what is, in fact, quite
arbitrary in any integrated production structure. Production is an
interdependent process involving the joint use of many resources,
and there is in general no clear way of deciding which resource has
produced what. The concept of the 'marginal product* of a resource
is not really concerned with who has 'actually produced*what, but
with guiding the allocation of resources by examining what would
happen if one more unit of a resource were to be used (given all the
other resources). To read in that counter/actual marginal story (what
would happen if one more unit were applied, giveneverything else) an
identification of who has 'in fact' produced what in the total output
is to take the marginal calculus entirely beyond its purpose and
depth.6
This problem of identification (who has produced what) applies
to Marx's theory of exploitation as well—perhaps even more
strongly, since the non-labour resources are treated in a verylimited
way. If that theory is to be seen not mainly as a significant descrip-
tion of the production process in terms of human work,7
but as the
usurping of one person's product by another, there remain manyun-
answered questions. Marx himself was fairly sceptical of this more
3
Bauer (1981:17). It was argued earlier—in Ch. I—that despite Bauer's explicitly
anti-egalitarian stand, his own theory insists on the equal right of the producers to
enjoy what they have produced. We are not concerned with that particular issue here,
but with the substantive content of the type of justice Bauer seeks.
6
I have discussed this issue in Sen (J985e).
7
SeeDobb (1937) for the classic exposition of the viewof labour theory—and also
of utility theory—-as rich description.

assertive diagnosis. Even though he did invoke it in many contexts,
he refused to see it as the central distributive concern in his
evaluative system (discussed most clearly in his 'Critique of the
Gotha Program*, Marx 1875),
In analysing the relation between economic opportunities and
freedoms, the tradition of classification based on the so-called
Marxian classes can be quite inadequate. There are many other
diversities, and an approach to equality related to the fulfilment of
needs or to ensuring freedoms has to go beyond purely class-based
analysis. For example, even if inequalities based on property
ownership are eliminated altogether, there can be serious inequal-
ities arising from diversities in productive abilities, needs, and other
personal variations,
The case for going beyond class analysis was, in fact, persuasively
made by Marx (1875) himself in chastising the German Workers
Party for taking for granted that equality in rewards for work would
not conflict with equality in satisfying needs.
But one man is superior to another physically or mentally and so supplies
more labour in the same time, or can labour for a longer time; and labour,
to serveas a measure, must be defined by its duration or intensity, otherwise
it ceases to be a standard of measurement.This equal right is an unequal
right for unequal labour,8
Marx saw the insistence on equal reward for equal work-
irrespective of needs—as an extension of a 'bourgeois right' seeing
human beings only as producers (Marx 1875: 9). The diversities
within the category of the working class made Marx insist on the
need to seek other classifications. In fact, productivity differences
constituted only one of Marx's concerns. He also focused attention
on the necessity to address our manifold diversities, including
differences in needs, and this led him to the well-known slogan
'from each according to his ability to each according to his needs'.
An essential part of Marx's complaint was about the mistake of
seeing human beings 'from one definite side only1
, in particular
seeing people 'only as workers, and nothing more seen in them,
everything else being ignored*.'
As an illustration Marx had referred specifically to the fact that
different workers have families of different size,I0
While it could be
* Marx (1875: 9), 9
Marx (1875; 9), On this distinction, see also OE1, ch. 4.
to 'Further, one worker is married, and another not; one has more children than
another and so on and so forth* (Mara 1875: 9),

argued that as far as the parents are concerned, the number of
children are, at least partly, within their control (and therefore a
subject on which their own responsibility must be accepted), that
argument would not reduce at all the force of the claims of the
children themselves." The differential needs arising from unequal
family size can be, to some extent, accommodated within the in-
come-based approach by suitable normalization and by the use of
'equivalence scales'.'2
But the more general problem of need vari-
ations cannot be similarly handled. Equality of incomes, or—more
generally—of primary goods or resources, can fail to yield equal
satisfaction of needs, when needs vary interpersonally and so does
the transformation of resources into need-fulfilment. In pursuing
the demands of equality in the space of well-being, or need-fulfil-
ment, we have to go beyond the income-based categories and also
the so-called Marxian classes (indeed, as Marx himself had argued).
This argument is exactly parallel to the one, discussed earlier, of
the variable conversion rates of incomes, primary goods, and re-
sources into freedom to do, to be, and to live the way one would like,
Pervasive human diversityis the source of both the problems and the
reason for their respective importance. After acknowledging and
taking on board the widespread relevance of the diversity of class,
ownership, and occupation, we also have to go well beyond that to
other diversities that influence the lives that we can lead and the
freedoms we can enjoy.
Sometimes the other sources of disparity may bepartly associated
with class, but diverge from it in specific respects with influences of
their own. For example, race and colour may have good statistical
correlation with class in the United States or the United Kingdom,
but the deprivation associated with being black isnotjust a matter of
its class correlates. The way a person is viewed in a society with
racial disparity may be deeply influenced by his or her visible racial
characteristics, and that can act as a barrier to functioning possibili-
" This is not to deny that there is an 'incentive issue* also involved in offering
larger social help for larger families, and problems of population policy cannot be
ignored in thiscontext. But that incentiveargumenthas to be tempered by the concern
for the well-being and capabilities of the disadvantage! members of the larger
families among the poor.
11
See Barten (1964), Muellbauer (1974a, 19746, 1987), Potlak and Wales (1979,
1981), Deaton (1980,1988), Deaton and Muellbauer (1980,1986), Btackorby, Don-
aldson, and Aucrsperg (1981), Atkinson(1983,1989), Jorgenson and Slesnick (1983,
1984a, 19846, 1987), Bhckorby and Donaldson (1984, 1988), Atkinson and Bour-
guignon (1987), Buhmann et al, (1988), among othercontributions.

ties in many circumstances. Distinctions of caste similarly have
influences of their own, despite being frequently correlated with
class.u
Race or caste can be a factor with far-reaching influence on
many aspects of day-to-day living—varying from securing
employment and receiving medical attention to being fairly
treated by the police. Inequalities in the distribution of income
and ownership will typically be part of the story, but by no means
the whole of it.
8,2. GENDER AND INEQUALITY
One basis of classification that is particularly relevant in this con-
text is gender. There are systematic disparities in the freedoms that
men and women enjoy in different societies, and these disparities are
often not reducible to differences in income or resources. While
differential wages or payment rates constitute an important part of
gender inequality in most societies, there are many other spheres of
differential benefits, e.g. in the division of labour within the house-
hold, in the extent of care or education received, in liberties that
different members are permitted to enjoy.l4
Indeed, in the context of intrahousehold divisions, it is not easy to
split up the total household income into the incomes going respec-
tively to different members of the family. The sharing may take an
unequal form, especially in relation to needs, but this is hard to
translate into income differentials, which would be an odd concept
to use in examining intrahousehold divisions." Inequality inside the
13
See Srinivas (1962) and Beteille (1981, 1983a, 1987, 1990),
14
On different aspects of'the gender gap', see Amsden (1980), Okin (1987,1989),
Bergmana (1986), Goidin (1989), Folbre et at, (1991), Nussbaum (1991a, I99lb),
R. A, Putnam (1991), Annas (1992), O'Neill (1992), among other contributions,
15
The use of 'household-equivalence scales' provides a way of making inter-
household comparisons, typically assuming no differential treatment within the
household, even though that feature can to some extent be altered by discriminating
analysis (sec Mueiibauer 1987; Deaton 1988; Blackorby and Donaldson 1988). But it
is, fa general, informationally limited to the extent that the observations do not
directly include actual functionings (e.g. individual undcrnutriiion or morbidity) and
concentrates instead only on aggregate consumption patterns and commodity com-
positions of the household. However, sometimes the limitations of data regarding
functioniags can make this approach the best that can, in those circumstances, be
used.

household is one of resource-cue and of the transformation of the
used resources into capability to function, and neither class of in-
formation is well captured by any devised notion of 'income dis-
tribution* within the family.
There is a lot of indirect evidence of differential treatment of
women and men, and particularly of girls vis-a-vis boys, in many
parts of the world, e.g. among rural families in Asia and North
Africa,1
* The observed morbidity and mortality rates frequently
reflect differential female deprivation of extraordinary proportions,
Even the crude ratio of women to men in the total population
varies between only 0.93 and 0.96 in South Asia, West Asia, North
Africa, and China. In contrast, partly because of the biological
advantages that women seem to have over men (given symmetric
care), the female-male ratio in the population tends to be much
higher than unity(around 1.05 or so) in Europe and North America.
It is, of course, quite possible that a part of the higher mortality rate
of the males in the richer countries reflect social rather than bio-
logical factors.17
For example, there is considerable evidence of
greater incidence of death from violence for men in many societies,
such as the USA. There are also some effects of higher male mor-
tality due to war deaths. But there does seem to be a substantial
biological component in the advantages in favour of women, given
similar treatment.18
The mortality differential against women in
Asia and North Africa, thus, reflects quite a remarkable departure
from what could be expected on the basis of biological potentials,
given symmetric care (on this see Sen, 1989«). The highermortality
and morbidity of women vis-it-vis men in these countries
reflect serious 'attainment inequality*, in addition to exhibiting
'* On this see, among many other contributions, Boserup (1970, 1987, 1990),
Tinker and Bramsen (1976), A. Mitre (1990), Miller (1981), L. C. Chen, Huq, and
D'Souza (1981), Rosenzweig and Sehultz (1982), Buvinic, Lycette, and McOreevey
(1983), Kynch and Sen (1983),Sen and Sengwpta (1983), P. Bardhan (1984,1987),Sen
(1984,198Srf, 198Se, 1990c), Jain and Banerjee (1985),Kynch(1985), M. Chen (S986a,
19866), Banister (1987), Harriss and Watson (J987), Das Gupta (1987), M. Vaughan
(1987), A. Ba§u (1988), Behrman (1988, 1992), Behrman and Deolalikar (1988),
Kumar (1989), Sen (!989«, 19906), Tinker (I990a), Kmnbur and Haddad (1990),
Harriss (1990), Whitehead (1990).
17
On the generalissue of cultural influences on gender differences in demographic
factors, see Johansson (1991).
18
On this, see I. Waldron (1976,1983). The biological advantages seem to apply
even in the womb, with female foetuses havinga lower rate of miscarriage than their
male counterparts.

extraordinary extents of 'shortfall inequality*, given the biological
potential in the opposite direction.l9
In the context of many developing countries, these are elementary
and important aspects of gender inequality, and their assessment
need not be derivative on any constructed concept of income in-
equality within the family. They reflect functioning differences and
the corresponding disparities in the elementary capabilities to avoid
escapable morbidity and preventable mortality. We are not con-
cerned here with the causal factors underlyingthese gender inequal-
ities,20
but with the prior exercise of identifying the nature of the
problem of gender inequality. Here a departure from the traditional
perspective of income distribution towards direct accounting of
functionings and capabilities can be an important step,21
Even when the gender differentials in morbidity and mortality are
not so acute, there can be disparities in other important functionings
and capabilities yielding substantial inequalities in freedoms.
w por example, if India had the African ratio (1.02) of females to males (not to
take up the very high ratio of 1.05 or so of the long-lived European or North Ameri-
can population), rather than the ratio it does actually have (0.93), then—given the
number of males in the country—there would have been nearly 30 million more
women in India in the mid-1980s. The corresponding number of'missing women* in
China (vis-a-vis the African ratio) is close to 40 million (see Dreze and Sen 1989). To
provide serious quantitative estimates of 'missing women', proper demographic
models of births and deaths would have to be considered (with clear specification of
the possible counteifactual scenarios), but we get some idea of the enormity of the
problem from even these crude estimates of the millions involved(based on using the
ratios obtaining in sub-Saharan Africa). For analyses of the different economic,
social, and cultural factors underlying the problem of 'missing women', see Sea
(!988c, 1989a) and Dreze and Sen (1989).
2B
I have tried to discuss them elsewhere, in Sen (1984, 19854 19&9a, 1990c). For
analyses of various aspects of this question, see also Boserap (1970, 1987, 1990),
Gardiner, Himroelweit, and Mackintosh (1975), Banerjee (1979,1982), Loutfi (1980),
Manser and Brown (1980), McElroy and Homey (1981), Miller (1981, 1984),
Rochford (1981), Young, Wolkowite, and McCullagh (1981), Beneria (1982), Dton
(1982, 1983), Rosenzweig and Sctaite (1982), Ahmed (1983), Buvinie, Lycette, and
McGreevey (1983), Kynch and Sen (1983), Sen and Sengupta (1983), P. Bardhan
(1984,1987), Folbre(1984), K. Bardhan (1985), Jain and Banerjee (1985), Mazumdar
(1985), Agarwal (1986, 1991), M. Chen (19860, 19866), Banister (1987), Behrman
(1988), Dm Gupta (1987), Harriss and Watson (1987), A, Basil (1988), Okin (1989),
Harriss (1990), Papanek (1990), Tinker (1990a), Allen (1991), Folbre el at, (1991),
among other contributions,
11
This does not deny that differential earning power of women vit*£-vb men may
affect the status of women (on this, see Boserup 1970; P. Bardhan 1984; Sen 1984,
I985rf, I990c, K. Bardhan 1985), or influence economic calculations underlying child
care (on this see Rosenzweig and Schultz 1982). That is a separate issue requiring
cmaal analysis of the role of income-earning power, and must be distinguished from
the problem of diagnosis of gender inequality.

While anti-femalebias in nutrition,morbidity, or mortality ismuch
less present in sub-Saharan Africa,22
there are often big gender
differences in many other capabilities, such as being able to read and
write, avoiding bodily mutilation, being free to pursue independent
careers, or being in positions of leadership.
In terms of many social functionings, gender differences can be
important also in the rich countries of Europe and North America,
even though in terms of survival and mortality, women do have a
relative advantage (at least in terms of attainment, if not in shortfall
avoidance). I shall not have the opportunity of pursuing this ques-
tion further here,23
but I would argue that the question of gender
inequality in the advanced societies—no less than in developing
countries—can be understood much better by comparing those
things that intrinsically matter (such as functionings and capabili-
ties), rather than just the means likeprimarygoods or resources. The
issue of gender inequality is ultimately one of disparate freedoms.
8.3. INTERREGIONAL CONTRASTS
Before closing this chapter, I would like to discuss some empirical
examples of interregional contrasts to illustrate the importance of
the distinction between judging poverty by income and judging it by
the capability to achieve some basic functionings.
Some of the most important functionings for living standard,
including the most elementary one of being able to live long (with-
out being grabbed by premature mortality), often diverge from real
income per head in a really spectacular way. This is easily seen in
making international comparisons of gross national product (GNP)
per head and life expectancy at birth,2
'1
In terms of per capita GNP,
South Africa ($2,470), Brazil ($2,540), Gabon ($2,960), and Oman
22
On the relatively betternutritional situation of girls vis-it-vis boys in manyparts
of Africa, see Svedberg (1988, 1990), Sec also Deaton (1988).
23
I have, elsewhere, tried to discuss the general question of gender inequality as
unequal outcomes of'co-operative conflicts' in the broadperspective of inequality or
entitlements and capabilities; see Sen (198la, 19S5J, I990c). See alsoKyncb and Sen
(1983), Sen and Sengupta (1983), Bryceson (1985), Jain and Banerjee (1985),Kyneh
(1985), Tilly (1985), Vaughan (1985, 1987), Brannen and Wilson (1987), Wilson
(1987), Asianbeigui and Summer-field (1989),Drfae and Sen (1989), Papanek (1990),
Tinker (I990o), UNDP (1990),Agarwal (1991), M. Chen (1991), Ahmad et a!.(1991).
24
The data on international comparisons are taken from the World Development
Report 1991 (World Bank 1991). Seealso UNICEF (1987,1992), UNDP(1990,1991)
on related matters. The figures are in US dollars.

($5,220) have six or more times the per capita GNP of China ($350)
and Sri Lanka ($430). But these relativelyricher countries givetheir
people significantly lower ability to survive premature mortality
(with life expectancies varying between 53 and 66 years) than do the
two lower-income countries (with life expectancies around 70 years
or more), Costa Rica, which is also considerably poorer than the
first four countries, offers not only a much higher life expectancy
than those four (and other 'upper-middle-income countries'), but a
life expectancy that is not significantly below those obtaining in the
richest countries of Europe and North America (with ten or more
times Costa Rica's GNP per head). For example, the USA with a
GNP per head of $20,910 has a life expectancy at birth of 76 years,
whereas Costa Rica with a GNP per head of only $1,780 has already
achieved a life expectancy of 75 years,
As we move our attention from commodities and income to func-
tionings and capabilities, the relative picture can change radically.
The difference seems to relate to a great extent to differences in the
social, educational, and epidemiological conditions. The achieve-
ments of China, Sri Lanka, and Costa Rica in quality of life has
much to do with policies regarding communal health services,
medical care, and basic education.25
Thus, this distinction between
deprivation of income and that of the capability to achieve ele-
mentary functionings, also has some relevance for public policy—
both for development and for the removal of poverty and in-
equality.26
Another interesting exercise relates to contrasts within a large
country, e.g. India. Among the Indian states, Kerala has one of the
lower real incomes per head, but by a long margin the highest life
expectancy at birth—over 70 years (compared with around 57 years
for India as a whole).27
Its infant mortality rate is, correspondingly,
25
In the case of China, the big surge in life expectancy and decline in mortality
rates took place before the economic reforms of 1979, and occurred actually in a
period of very moderate economic growth and virtually stagnant food output per
head. In contrast, in the post-reform period growth of GNP has been fast, but progress
in life expectancy seems to have faltered a little. On this and related matters, see C.
Riskin (1987) and Dreze and Sen (1989).
2
* Of course, life expectancy, literacy, and other common indicators of 'quality of
life* are not, by themselves, anywhere near adequate in reflecting the overall picture
of the capability to achieve valuablefunctionings, but they are importantparts of the
overall picture.
21
The Sample Registration Survey for 1986-8 yields a figure of 73.2 years for
females and 67.0 years for males for Kerala as a whole.

much lower than the Indian average.28
Kerala also has a much
higher level of general literacy (91 per cent, as opposed to the Indian
average of 52per cent), and particularly female literacy (87 per cent,
compared with the national average of 39 per cent).29
Indeed,
Kerala's achievements for many crucially important functioning
are not only very much better than those of the rest of India, but they
have an edge in some fields—especially with respect to women—
even over China and Sri Lanka. For example, the low female-male
ratio that characterizes China as well as India as a whole (around
0.93), in contrast with the substantial excess of females over males
in Europe, North America, and sub-Saharan Africa (discussed
earlier in this chapter), does not apply to Kerala. The female-male
ratio for Kerala is 1.04, which is very similar to the ratios around
1.05 in Europe and North America,*
If the value of average GNP is 'corrected* by taking note of dis-
tributional inequality, the income perspective is made somewhat
more articulate. But even with this adjustment (i.e. even when dis-
tribution-corrected measures of real income are used), Kerala still
remains one of the poorer Indian states.31
Distributional corrections
do not seem to eliminate adequately the deficiency of the income
approach to explain the high capability levels in Kerala to escape
premature mortality. The deficiency of the income approach is not
adequately remedied by supplementing the average income figures
21
However,Kerala does have a much higher self-reported morbidity rate than the
rest of India (on this questionand related issues, see Panikar and Soman 1984; Kumar
1987, 1989; Vaidyanathan 1987), This may, to some extent, reflect the low-income
level of the Kerala population and possible nutritional deficiencies resultingthere-
from. But to a great extent the higher self-reported morbidity seems to be a conse-
quence of more awareness of health status on the part of the Kerala population,
largely related to greater literacy and higher use of health services. Indeed, the
self-reported morbidityrates are lowest in the least literate states of Bihar and Uttar
Pradesh, which have very high mortality rates (combined with illiteracy). And as
Murray and Chen (1990) have shown in a recent paper, using similar criteria of
self-reporting of morbidity, the United States has even higher rates of reported ill-
health than Kerala. Incidentally,this inverse connection between self-perception of
illness and observed mortality rates also illustrates the pitfalls of going only by
self-perception in judgingwell-being(discussed in Ch. 3).
** The Indian comparative data are taken from publications and working papers of
the office of the Registrar General and Census Commissioner of India, including
Census of India 1991: Provisional Population Totals (New Delhi: Government of
India, 1991). The literacy data relate to the population aged ? years or more.
39
In fact, the ratios would possibly not be materially different if the effects of the
differentially higher male mortalityof Europeans and North Americans due to past
wars were to be factored out.
31
See Sen (19766) and Bhattacharya, Chatterjee, and Pal (1988).

by considerations of inequality of incomes and commodity
holdings.
The explanation of Kerala's success in the important space of
basic capabilities has to be sought in the history of public policy
involving education (including female literacy) and health services
(including communal medical care), and to some extent, food dis-
tribution (including the use of public support of food consumption
of the rural as well as the urban population), in contrast with the
rest of India. There are also other factors involved, including a more
favourable position of women in property rights and in inheritance
among a substantial and influential section of Kerala's population,
and the greater public activism connected with educational cam-
paigns as well as politics in general. The history of public action in
Kerala goes back a long time, with remarkable literacy campaigns
in the native states of Travancore and Cochin in the nineteenth
century.
This monograph does not, of course, provide the occasion to go
into details of policy issues,32
but it is important to emphasize that
the evaluativeperspective of the capability approach does draw our
attention forcefully to examining and scrutinizing such policy ques-
tions. It also suggests the need to take a broad view of development
efforts, goingfar beyond the focus on improvingthe national output
and the distribution of incomes,
Jl
These diverse matters (including the international comparisons and the regional
contrasts within in India) havebeen discussed, with a focus on policy issues, in Dreze
and Sen (1989).

9
THE DEMANDS OF EQUALITY
9,1. QUESTIONS OFEQUALITY
This monograph has had two rather different—though inter-
related—objectives. The first aim is primarily methodological, the
second mainly substantive.
The §rst set of issues is concerned with understanding the rele-
vance and reach of the questions that can be legitimately asked
about egalitarianism, in particular, 'why equality?' and 'equality of
what?". In that context, it isimportant to come to grips both (1)with
the diversityof human beings (the fact that wediffer from each other
in personal characteristics as well as external circumstances), and
(2) with the plurality of relevant 'spaces' in which equality can be
judged (the multiplicity of variables—incomes, wealths, utilities,
liberties, primary goods, capabilities—which can respectively serve
as the sphere of comparison). The demands of equality in the
different spaces do not coincide with each other precisely because
human beings are so diverse. Equality in one space goes with sub-
stantial inequalities in others.
The second set of exercises relates to exploring a particular sub-
stantive approach to equality. While I began with methodological
issues (Chapter I), the bulk of the book has been concerned with
substantive matters. The particular approach to equality that Ihave
explored involves judging individual advantage by thefreedom to
achieve, incorporating (but going beyond) actual achievements. In
many contexts, particularly in the assessment of individual well-
being, these conditions can, I have argued, be fruitfully seen in terms
of the capability to function, incorporating (but going beyond) the
actual functionings that a person can achieve. The 'capability
approach' builds on a general concern with freedoms to achieve
(including the capabilities to function).
Thecapability approach points to the need to examine freedom to
achieve in general and capabilities to function in particular. In addi-
tion to discussing the motivation underlying the capability
approach, I have also tried to consider how we might address the

130 The Demands of Equality
difficulties in interpreting and defending this view (Chapters 2 and
3). I have gone on to examine the implications of this approach for
assessing freedom and advantage (Chapter 4), for theories of justice
(Chapter 5), for the welfare economics of inequality evaluation
(Chapter 6), for the assessment of poverty in rich as wel! as poor
countries (Chapter 7), and for analysing inequalitiesassociated with
categories such as class, gender, and other groups (Chapter 8).
In this final chapter I shall try to review and assess some of the
general points that have emerged from the preceding analysis (even
though there will be no attempt to 'summarize' the discussions pre-
sented, or list* the specific conclusions reached). I shall be particu-
larly concerned with the interrelations between the methodological
and substantive issues.
9.2. EQUALITY, SPACE AND DIVERSITY
It was argued, in Chapter 1, that the often-asked question 'why
equality?* can be quite misleading. Every plausibly defendable ethi-
cal theory of social arrangements tends to demand equality in some
'space*, requiring equal treatment of individuals in some significant
respect—in terms of some variable that is important in that particu-
lar theory. The 'space* that is invoked does differ from theory to
theory. For example, 'libertarians' are concerned with equal
liberties; 'economic egalitarians' argue for equal incomes or
wealths; utilitarians insist on equal weight on everyone's utilities in
a consequentialist maxirnand; and so on. But in each system a de-
mand for equality—in its own form—-is incorporated as a founda-
tional feature of that system. What reallydistinguishes the different
approaches is the variation in their respective answersto the question
'equality of what?'. That question is truly central in understanding
the distinctions between the diverse ethical approaches to social
arrangements.
I have also argued that there are good reasons why all the major
ethical theories of social organization tend to demand equality in
some space—a space that has some basic importance in that theory.
There is a connection here with the foundational need for the plau-
sibility of a theory to extend equal consideration to all, in some
crucial way (related to the structure of that theory).1
This diagnosis
does not, of course, suggest that we should 'do away' with the ques-
1
I have examined some of the issues involved in this diagnosis in Sen (I985«).

The Demands of Equality 131
tion 'why equality?', Each specific theory of equality in any space
does require a defence. But the nature, content, and demands of the
query 'why equality?' can be understood and examined only by
relating it to the central question 'equality of what?'.
Each approach has its own interpretation of what we have been
calling 'basal equality5
—equality in some individual feature that is
taken to be basic in that particular conception of social justice and
political ethics. That foundational concern leads to a particular
way of interpreting the demands of symmetry and impartiality, and
this in its turn disputes the case for insisting on equality in other
spaces—seen as less basic in the light of that theory. For example, to
demand equal entitlements to an extensive set of libertarian rights
as a basic requirement of social organization (as is done by Nozick
1974) has the consequence of undermining any insistence on the
equality of incomes, wealths, or utilities, as well. Similarly, if equal
incomes are to be achieved, then wecould not insist also onequality
in utilities, or rights, or freedoms. An immovable object leaves little
room, in principle, for any irresistible force.
Conflicts that can arise in principle need not, however, arise in
practice. The need to determine conceptual priorities—important in
principle—becomes momentous in practice because of the far-
reaching diversities of human beings (the reach of these diversities
was extensivelydiscussed in the previous chapters). The demands of
equality in different spaces tend to conflict, infact (not just in prin-
ciple), with each other. Thus, the choice of basal equality has tre-
mendous practical importance in asserting some claims and denying
others. The need for ensuring the fulfilment of basal demands, in-
cluding basal equality, necessitates the tolerance of inequality in
what are seen as the outlying'peripheries'.
9,3. PLURALITY, INCOMPLETENESS AND EVALUATION
The demands of equality can take various forms, and a certain
amount of plurality of concerns is inescapable in the evaluation of
basal equality itself. This can introduce some ambiguities in charac-
terizing the conditions under which equality can be taken to be
complete (or 'full'), and—perhaps more importantly—in ranking
alternative possibilities in terms of their extents of inequality (i.e. fa
judging the respective 'distances* from full equality). The source of
these problems lies in what may be called 'internal plurality',

relating to: (1)the heterogeneity of the space in terms of which basal
equality is defined, and (2) the different ways in which 'distances'
may be measured and inequalities compared in the distributions
even of a homogeneous variable.
Some of the basal variables certainly have much heterogeneity
within their own respective categories.2
For example, the liberty that
is championed by libertarians would inevitably involve different
types of rights, covering different spheres and taking various forms
(such as 'claims', 'immunities', 'powers', etc.),3
Similarly, primary
goods are of different types. Different capabilities are distinct from
each other. Even utilities are diverse,4
It is also possible for an ethical theory to include more than one
type of variables in the category of basal significance. We may be
concerned both with liberties and with levels of well-being. We may
attach intrinsic importance both to well-being and to agency. We
may value freedom as well as achievement.5
Indeed, pluralist pro-
posals make up much of practical ethics, even though descriptive
homogeneity evidently appeals to many moral philosophers (utili-
tarians among them).6
There are also different ways of evaluating equality in the same
homogeneous space, using distinct methods of measuring inequality.
Variations in inequality indicators in a.givenspace (e.g. coefficient of
variation, Gini coefficient, standard deviation of logarithms,
measures of entropy) have been extensively discussed in the litera-
ture,7
The ideas underlying the discipline of measurement vary
greatly among the different measures, and whilemany of these ideas
have good reasons behind them, they often conflict with each other.
Different features of basal equality can, therefore, suggest
different rankings of particular situations. Sometimes the plurality
2
Thisquestion was discussed in the specific contextof the heterogeneityof fune-
tionings in ON. 3,
3
See Ranger (1957, 1972, 1985), Lindabl (1977), Raz (1986).
4
The claim that different types of utilities have different status in human well-
being and in socialrelevancecan be traced back to Aristotle aod wasemphasizedalso
by John Stuart Mill (1859, 1861).
5
In the previouschapters I have discussed the force of these considerations wen
severally and jointly (Chs. 2-4).
* The belief that ethical pluralism lacks sophisticationis of respectableancestry
(going back at least to John Stuart Mill), but it is not particularly sound for that
reason (or any other). I have tried to discuss this questionin Sen (198Sa). On related
matters, see B. Williams(1973a, 1985), Nagel (1979), Hurley (1989).
7
It was one of the main subjects of OEl. Helpful accountsof the literature can be
found in Cowell (1977), Foster (1985), and Lambert (1989).

arising from these sources may be reduced through scrutinized ex-
clusion of the claims of particular elements in the initialplurality,
At other times the pluralitycan be lessened through 'combining' or
'uniting' the considerations by some procedure of evaluativeweight-
ing. But even after all these reductions are carried out, there may
remain some residual plurality, with consequent ambiguities in the
ordering of equalities and inequalities.
The presence of ambiguities of this type is, in fact, a central prob-
lem in decision theory and in social choice theory, and can arise in
many different contexts. Several lines of analysis have, in fact, been
proposed to address the demands of reasoned decision despite
residual ambiguity(or "unresolved conflicts'),8
One simple line of—-
relatively unambitious—reasoning, which was much used in my
earlier book on inequality (OEI), involves separating out a shared
partial ranking in which all the desirable featuresmove together. An
'intersection' partial ordering places x above y if and only if x is
better than y according to all the desirable features.
In the context of the capability approach, the problems of selec-
tion and weighting of different functionings and capabilities were
discussed earlier (in Chapter 3), There are various analytical and
logistic issues involved in evaluating functioning vectors (and
capability sets consisting of such vectors), but the techniques of
'dominance* and of'intersection* are not only consistent and cogent,
they can also, typically, take us quite a substantial distance.9
The use of'intersections' does not, of course, obviate the necessity
of thoroughly scrutinizing the claims of each allegedly 'desirable
feature'. But when the different criteria that survive scrutiny actually
conflict in rankingtwo alternatives, the pair has to be left unranked.
Sometimes further analysis or more information may permit us to
cut down the 'pluralities* that haveto be accommodated. When that
occurs, the intersection partial ordering can be correspondingly
extended— ordering pairs that could not be ranked earlier.w
In this
sense an intersection partial ordering would typically be tentative,
and always open to extension if and when reasons are found to cut
down the relevant pluralities (through scrutinized elimination of
some features, or weighted combining of distinct aspects). The
8
Some of the problemsinvolved are discussed in Sen (I970o, 1982<t), Suzumura
(1983), and Levi (1986). See also the literatureon 'fuzzy* sets and preferences.
* I have tried to discuss the scope and range of the 'intersection* analysis in Sen
(19700, 19706, 1985*).
19
On this general question, see Sen (1970a, 1970*. 19820).

distinction between 'foundational* and 'pragmatic' incompleteness
discussed in Chapter 3 is relevant in this context."
The need to admit incompleteness in inequality evaluation is
inescapable, and there is much to be said for addressing that ques-
tion explicitly rather than in grudgingly implicit ways. The incom-
pleteness may be due to the nature of the concept itself (e.g. the idea
of equality may incorporate substantial ambiguities), or because of
the absence of information (e.g. data may be lacking that would
permit some comparisons to be made), or due to the need to respect
residual disagreements among the parties involved (e.g. this would
relate to the acceptance of plurality as a part of an approach to
fairness and justice). To 'complete' partial orderings arbitrarily for
the sake of decisiveness, or convenience, or order, or some other
worthy concern, may be a very misleadingstep to take. Even when
the partial ranking is quite extensively incomplete, the case against
'forcing' completeness may be quite strong. Babbling is not, in
general, superior to being silent on matters that are genuinely
unclear or undecided.
However, one consequence of adopting such a 'partial ranking*
view of interpersonal comparisons and of the assessment of equality
is to admit the possibility that in many situations no clear judgement
can be made as to whether there is more equality in situation a than
in situation b.n
Those who see equality as a clear, articulate, and
decisive arbitrator of every social or political dispute would find
this position particularly unattractive. I don't share that disappoint-
ment, for reasons that I have discussed (in Chapters 1~4),13
Nor do I
see that arbitrarily completing a partial order does much justice to
the idea of equality, or—for that matter—helps to ensure that the
demands of equality would receive attention when it does have
something clear and compelling to say. There is even a real danger
of undermining the concern for reducing inequality by an over-
ambitious programme of trying to catch all the little gaps that may
be detectable in some very particular scaling but which would be
hard to see in terms of other plausible views. A more solid founda-
1
' See Sect. 3.4, Different reasons for acceptingsome foundational incompleteness
has been discussed in Chs. 1-4.
12
It must, however,be emphasized that in identifying poverty and social depriva-
tion muchcan be fruitfully said even with a list oflevels of capabilities that are taken
to be minimally necessary, and this does not require anything like an ordering of
different bundles of capabilities. On this, see Sen (1984) and Hossain (1990).
» Also in Sen (I970a, 1973a).

tion can be built for the rejection of substantial inequalities—visible
from different angles—by not tumping them together with more
fine-tuned presumptions,
9.4, DATA, OBSERVATIONS AND EFFECTIVE FREEDOMS
The limits of practical calculations are also set by data restrictions,
and this can be particularly hard on the representation of capability
sets, as opposed to observed functioning achievements. When the
data simply do not exist to calculate the extents of the respective
capability sets, there is no option but to settle for the chosen func-
tioning combination as the basis for forming a view of the oppor-
tunities that were actually enjoyed.14
In many situations, practical
compromises would have to be made, at least partially. But even
when the pragmatic acceptance of limitations of data availability
force us to set our sights lower than the full representation of
capability sets, it is important to keep the underlying motivations
clearly in view and to see practical compromises as the best we can
do under the circumstances.
It is also relevant to note here that even for assessing freedom, the
perspective of "effective freedom' as opposed to 'freedom as control*
can permit some limited comparisons of freedoms (in the form of
counterfactual choice) on the basis of observed outcomes. The argu-
ment was outlined in Chapter 4, and it involved defending references
to freedom in such expressions as 'freedom from hunger*or 'freedom
from malaria'. I have argued that these uses of the term freedom
need not illustrate just loose talk (as they are often taken to be),
since an important concept of freedom is, in fact, involved in. these
outcome-based judgements of freedom. If that argument isaccepted,
then the observed functioningscan, in particular circumstances, tell
us something not merely about well-beings achieved, but also about
freedoms enjoyed. For the main theses of this monograph, it is not,
of course, necessary that there be general acceptance of the rele-
vance of this view of effective freedom and of counterfactual choice
(giving it the limited but important role for which I have tried to
argue). But if that view is accepted, then there would be a bit more
articulation on some aspects of freedom and inequality than would
otherwise be possible on the basis of available data.
14
There may, of course, be various degrees of data unavailability, and the choice
need no! take an 'either-or' form.

9.5. AGGREGATION, EGALITARIANISM AND EFFICIENCY
The understanding of 'internal plurality* within the demands of
basal equality has to be supplemented by the recognition of claims
other than those of basal equality in general. No matter which space
is chosen for the assessment of equality, a conflict can arise between
aggregative considerations (e.g. generally enhancing individual ad-
vantages, no matter how distributed) and distributive ones (e.g.
reducing disparities in the distribution of advantages). Considera-
tions of 'efficiency'—much discussed in economics—reflect a
common element in aggregative concerns.l5
When it is expressed in
the form of a demand that no dominant improvement (enhancing the
position of each) should be left undone, there is an appeal to a
particularly non-controversial part of our aggregative moral
sentiments.
In economics, the demands of efficiency frequently take the special
form of 'Pareto optimality", which is defined in the particular space
of utilities. It is a weak condition of unimprovability, demanding
that there remains no possibility of any change that raises the utilities
of all.'* Pareto efficiency is almost certainly the most widely used
criterion in modern welfare economics, and its acceptability is
typically taken to be entirely non-controversial. Disputations on
the acceptability of Pareto efficiency as a necessary condition of
social optimality relates to the status of the utility space itself.17
They rarely take the form of questioning the case for carrying out
dominant improvements in what would be accepted as the relevant
space (enhancing the position of each in that space). The question is
whether utilities constitute the relevant space. What is disputed in
ls
Note that thispart of the aggregativeconsideration can be used even when there
is interpersonal ineomparability in the variable in question (e.g. individual utilities
are non-comparable).
'* A stronger version ('strong Pareto optimality*) is also used, particularly in
welfare economics, and it demands that no possible improvement could raise the
utility of one, without reducing the utility of any.
17
The acceptability of Pareto optimality as a necessary condition has to be distin-
guished from any claim regarding its sufficiency as a criterion of social optimality.
Even though Pareto optimally is sometimes treated implicitly as if it were sufficient,
the more common—and certainly tic more explicit—claim is in favour of its neces-
sity for social optimality. Given its complete neglect of distributive issues, sufficiency
would be a most implausible claim for Pareto optimality (except in specialized
contexts, such as consensual constitutionality, on which see J. M. Buchanan and
Tullock 1962), Whatisbeingdiscussed in the text here is not the obviouslimitation of
the sufficiency claims in favour of Pareto optimality, but the inadequacy of that
criterion viewed even as a necessity condition.

particular is the alleged priority of the space of utilities (or of
fulfilment of desires) on which Pareto optimally concentrates, and
this involves arguing for the need to accommodate the rival claims
of other considerations such as liberties or freedoms.a
Aggregative considerations can makeus move in a different direc-
tion from equality in general. In concentrating on the programme of
explaining the far-reaching role of basal equality, we must not
overtook the plurality of ethical concerns that take us beyond
equality altogether," The distinction between aggregative and dis-
tributive considerations has often been discussed in the specificcon-
text of assessing results (e.g. the conflict between increasing total
income and reducing distributional inequalities in incomes, or
between raising aggregate utility and decreasing interpersonal
utility differences). But similar contrasts hold in other spaces as
well, which may not involve any particular concentration on results
as such. For example, there may be a conflict between promoting
some rights in general (irrespective of distribution) and seeking a
more equal distribution of those rights. Indeed, the aggregative-dis-
tributive dichotomy is one of the more pervasive issues in social
evaluation. Equality—no matter how broadly defined—can hardly
be the only concern in any basal space, and aggregative considera-
tions (including the demands of efficiency) tend to have an ir-
reducible status.
However, in integrating these distinct considerations, we also
have to note that the aggregating concerns themselves may give a
crucial role to equality in their formulation. In carrying out aggre-
gation, there are questions of what to include and what weights to
attach, and here equality is often invoked as a demand on the disci-
pline of aggregation. Indeed, it is precisely in giving equal weight to
each person in the utilitarian aggregative maximand that utilitar-
ianism adopts a firmly egalitarian stand—a stand the significanceof
which has been particularly emphasized by such utilitarian analysts
as Harsanyi (1955,1982) and Hare (1963,1981,1982). Giving equal
weight on the utilities of each person in the aggregate rnaximand
makes utilitarianism an egalitarian approach in a very specific
way—specialized not just through the choice of a particular space
(viz. utilities), but also through the form that the requirement of
18
On this and related issues, see Sen (19700, I979a, l9S3a, 19920).
19
This plurality goes beyond the multiplicityof focus that may be part and parcel
of basal equality itself (Internal plurality*), discussed in Sect. 9.3.

equality takes in that space (viz, giving equal weight to all in the
aggregate objective, rather than, say, promoting the equality of
utility levels).20
The demands of equality can be imposed in different
ways, and it can have a role even in the formulation of what may
initially appear to be a purely aggregative exercise.
Thus, the demands of equality may come up in many distinct
contexts, in quite different ways. But it is also clear that not all the
issues to be settled can be resolved simply by the demands of
equality even in their most diverse forms. For example, while
equality may influence the form of the aggregative objective (e.g.
insisting on an unweighted sum as the proper rnaximand), the de-
mand to maximize that aggregate objective is not—in itself—a de-
mand of equality. We have to recognize the variety and extensive
reach of the demands of equality, without seeking in it a complete-
ness of considerations that cannot possibly be there,
9.6. ALTERNATIVE DEFENCES OF INEQUALITY
When competing with aggregative considerations, the demands of
equality may often be substantially compromised in the endorsed
arrangements. This recognition does not contradict the persistent
relevance and reach of egalitarian values. Indeed, the pursuit of
equality cart be properly evaluated only within a broader context in
which other demands are not arbitrarily ignored.
In the context of reasoned judgement of social arrangements,
inequality in terms of any variable (e.g. incomes, capabilities) may
be defended by using one of, at best, three different types of argu-
ments without generally disputing the relevance of equality for
social arrangements: (I) 'the wrong space' argument, (2) the 'incen-
tive' argument, and (3) 'operational asymmetry' argument.
'The wrong space' argument takes the form of claiming that the
variable in question (i.e. income, or capability, or whatever) is not
the right one in terms of which equality should be sought. It does not
provide, so it is suggested (usually implicitly),the right space for the
demands of equality. Indeed, in this class of arguments, the need for
equality in some other space may be inter alia asserted, such as
29
Meade (1976) provides a welfare-economic analysis in which the claims ofequal
weight in the maximand as veil as equality in the individual levels achieved receive
competing attention, in the space of utilities. See also Mirrlees (1971, 1986), Phelps
(1973), Mueller (1979), Atkinson and Stiglite (1980), Roberts (1980a), Dreze and
Stem (1987), Starrett (1988).

equality of libertarian rights (Nozick 1974), equal rights to enjoy
what one has 'produced' (Bauer 1981), equal and impartial treat-
ments in terms of some procedures (Gauthier 1986), equalweighting
of utilities in the aggregate maximand (Bentham 1789; Harsanyi
1955; Hare 1981), and so on. And because of these demands—
including demands of equality—in the other space, equality cannot
be insisted on in the particular space in dispute (e.g. incomes, or
capabilities, or utility levels, or some other field favoured by the
'egalitarians'),
I have discussed the nature of this argument and its implications
quite extensively earlier on in this monograph (in Chapter 1 and
also earlier on in this concluding chapter), and I shall not elaborate
it further here. Nor arn I discussing here why many of these claims
may be disputable,21
Indeed, the plausibility of valuing equality in
terms of capabilities (with which a lot of this monograph has been
concerned) is itself an argument against insisting, unconditionally,
on equality in other spaces.
The other two arguments are related to each other in bringing in
non-equality considerations in the space in question. They dispute
the demands of the pursuit of equality in that space not by airingthe
superior claims of another space, but by pointing to the conflict
between equality and efficiency (in a general sense) in that space
itself. It may be accepted that inequality in that space may be a bad
thing, but that badness (it is claimed) may be outweighed by its
efficiency advantages. Attempts to eliminate that inequality would
lead to worse consequences, e.g. worsening of the position of all (or
most people).
Efficiency-based critiques of equality can takeat least two distinct
forms, and in the list presented earlier, these were put down as two
different arguments (as they indeed are): the 'incentive' argument
and the 'operational asymmetry' argument. The incentive argument,
which has perhaps been more discussed in the literature, concen-
trates on the need to give people the incentive to do the right thing
for the promotion of the objectives. Inequality may, thus, play a
functionally useful role in encouraging work, enterprise, and in-
vestment.22
Jl
I have presented arguments against the claims of Nozick (1974) and Bauer
(1981), among other claims, in Sen (1982*. 1985e).
22
Unequal outcomes may also be associated with the willingness to take risky but
useful decisions. However, the relation between riskiness and inequalities in income
distribution is more complex than is often presumed; on this see Kanbur (1979).

It is often taken for granted that the objectives must be of the
'aggregative' type (in the sense of promoting individual achieve-
ments irrespective of distribution, e.g. seeking Pareto improvement,
or maximizing the sum-total). But the demands of efficiency can, in
fact, be related to any type of objectives—distributive as well as
aggregative.23
There are issues of efficiency in the promotion of
equality itself along with other objectives.24
The incentive argu-
ment, applied to individuals, deals with the need to provide motiva-
tion and encouragement to individuals so that their choices and
actions are conducive to the promotion of overall objectives,25
These
goals could be purely aggregative, or include distributive goals as
well.
The incentive argument has been invoked, for a very long time, in
disputing the immediate claims of equality. Egalitarian policies
have been criticized on the ground that they hinder the pursuit of
social goals, and this can happen even when these goals include
equality as well. Incentive-based critiques of egalitarianism have
had a fair amount of practical airing in recent years, e.g. in discuss-
ing what went wrong with egalitarian policies in Maoist China, or in
criticizing redistributive features of 'welfare state* policies.26
The third line of argument also relates to the tension between
equality and efficiency, but it concentrates specifically on aggrega-
tive objectives and the need to have inequality in promoting them
because of the necessity of operational asymmetry. The need for
such asymmetry can arise from differences in people's skills and
abilities. For example, it may be argued that givingmore power and
capabilities of particular types (e.g. in running governments, taking
business decisions) to the more able and skilled would help every-
a
On this and related issues, see' Le Grand (1990, 1991).
24
If 'efficiency' is taken to mean wasteless, productive promotion of objectives (no
matter what these objectives are), then the subject-matter of efficiency can encompass
the promotion of equality itself, if that were among the chosen objectives. But the
inclusion of equality among the objectives to be promoted does not rule out the
possibility that a no-nonsense pursuit of equality may lead to less overall
achievement
25
Simple but illuminating examples of the conflict between distributive concerns
and incentive demands can be found in the literature on optimum income taxes; see
e.g. Mirrtees (1971), Stern (1976),Sadka (1977), Seade (1977).
24
There isalsoa basicissueof individual incentivesinthe operation of politicalpro-
cesses (including legislation and administration)which govern the objectives pursued
by statepolicy; on this see Buchanan and Tuttoek (1962) and Buchanan (1975,1985),

one, but this would obviously be accompanied by inequality of
those powers and capabilities. This 'operational asymmetry' argu-
ment applies particularly to inequalities in some specific space, e.g.
powers and capabilities, and may or may not be directly relevant to
other spaces (such as utilities).21
Another variant of the 'operational asymmetry* argument focuses
on the possibility that the need for asymmetric treatment arises from
the social role of asymmetry (e.g. a few people have to take opera-
tional decisions to avoid confusion). In this view, efficiency of opera-
tion would require that some people should have more authority or
power than others. Asymmetric treatment may be necessary, in this
view, even if the people who are in authority are no more talented
than others are. Similarly, indivisibilities in the economic oppor-
tunities, which could not be shared given the nature of technology,
may lead to operational asymmetry on grounds of aggregative
objectives. The operational asymmetry arguments need not focus on
the individual incentive problem as such, but on the social role of
asymmetries, e.g. those related to the use of differences in skills, or
to the need for authority or discipline, or the presence of in-
divisibilities.2
*
9.7, INCENTIVES, DIVERSITY AND EGALITAWANISM
In the economic literature on resource allocation, the 'incentive'
argument in particular has been very extensively explored.29
How
does the emphasis on the diversity of human beings—one of the
27
Indeed, a variant of the 'operational asymmetry* argument may yield the odd
result in the utility space that the more productive should have lower total utilities. If
some people are more productive than others, and can generate more output through
the same amount of work (and the same loss of leisure) than others tan, then a system
of total utility maximization might require that these more productive people should
work harder and end up having less utility than others because of harder work and
smaller leisure. On this, see Mirrlees (1971,1974), and also Roenjer (1985). See also
Atkinson and Stiglitz (1980) and Tuoiaala (1984, 1990).
28
Arguments for asymmetric treatment can arise also in other circumstances and
may inter alia involve incentive problems (without being entirely reducible to it). See
Stiglitz's (1982) discussion of the difficulty of having "horizontal equity' along with
aggregative objectives. The particular case he deals with is that of utilitarianism, and
he considers cases in which 'social welfare (as measured by the sum of utilities) is
higher if individualswho have the same tastes and the same endowments are treated
differently' (p. 2).
2
' See Atkinson and Stiglitz (1980), Auerbach and Feldstein (1987), and Starrett
(1988).

recurrent themes in this monograph—affect the incentive argu-
ment? It is arguable that an explicit recognition, of some types of
deeper human diversities and their roles in inequality evaluation
may have the effect of restraining the force of the incentive prob-
lem. At least, the incentive problem, may have to be stated some-
what differently from the way it is often posed, especially in popu-
lar accounts.
In many economic models (e.g. in welfare economics or in public
finance), the disparity between the achievements of different persons
arises typically from differencesin efforts and in other decision vari-
ables of the individuals, and not from any antecedent diversity in
productive ability (though there may be differences in tastes, e.g. for
responding to risk or to financial rewards). These differences are
clearly related to problems of motivations and opportunities, and
in that context the incentive argument has—rightly—figured fairly
ubiquitously in that literature.30
There is much to be learned from
such analysis.
If, on the other hand, human diversities of particular types—
rather than differences in decisions—are an important factor
behind unequal achievement or freedom, then the incentive argu-
ment, in its straightforward form, may not directly apply. For
example, to the extent that gender or age is responsible for in-
equality of capabilities, the policy response may take the form of
providing special help to members of the more deprived gender or
age categories. Since it is impossible to change one's age rapidly,
and particularly hard to change one's sex, the special treatments
may not generate incentive problems of the standard kind,*' It is, of
course, possible to lie about one's age or gender, but that may not
always be easy nor convincing. In general, the possibility of incen-
tive distortions may be a good deal less with egalitarian policy in
this case than in the standard economic models involving in-
dividuals whose fortunes diverge because of their own chosen
levels of application.
Similarly, if individuals suffering from diseases are offered
special medical or other facilities, the incentive problem may be
relatively light, since people do not typically wish to cultivate
30
This holds not only for the market mechanism, but also for allocation under
other institutional arrangements. For the relevance of such considerations for re-
source allocation in economies with peasants or collectives, see Putterman (1986).
11
See, in the context, Sen (1973a), Akeriof (1978, 1984), and Atkinson (199Ic).

diseases, nor do they usually have much use for the specific medical
and other facilities that may be on offer. To the extent that free or
heavily subsidized medical facilities may make people take less
precaution, there would of course be an incentive effect here too
(since the difference here would be caused by a choice variable, to
wit, being careless), but in most circumstances and with most ill-
nesses people are reluctant to take such risk just because the treat-
ment itself would be free or inexpensive. The different genetic and
environmental risks of illness can be dealt with by providing special
medical facilities without having terrible incentive problems.32
Egalitarian policies to undo the inequalities associated with
human diversity are much less problematic front the point of viewof
incentives than policies to undo inequality arising from differences
in effort and application, on which much of the incentive literature
has tended to focus. Thus, the importance of human diversity in
inequality evaluation, with which we have been much concerned in
much of this monograph, may also have some considerable bearing
on the nature and force of the incentive problem in pursuing
egalitarian policy (particularly in the context of moving towards
less inequalityof elementarycapabilities). This is not a trivial issue,
in so far as antecedent diversities (e.g. in gender, age, class) are
among the central factors behind unequal freedoms that people
have in the world in which we live.
9.8. ON EQUALITY AS A SOCIAL CONCERN
This monograph has been much concerned with exploring capabili-
ties as the basis of judging individual advantage. I have tried to
emphasize that this capability perspective can be used not just for
evaluating equality, but also for assessing efficiency. Efficiency in
the capability space, if defined analogously to the usual definitions
of "economic efficiency' (characterized in terms of the utility space),
would require that no one's capability can be further enhanced
32
For a private insurance company, there will of course be considerable incentive
to find out who arc genetically more prone to illness, since excluding them would
enhance profits. If, however, it is accepted that there are good social arguments for
providing medical coverage for those whoare genetically more prone to disease, then
that policy initiative need not run into a severe barrier of incentive problems. The
"incentive compatibility* in this case contrasts with the problems that arise when the
differences are not due to basic human diversities, but mainly the result of choice of
actions.

while maintaining the capability of everyone else at least at the same
level."
The explicit acceptance of aggregative concerns as an integral part
of social evaluationdoes make a difference to the wayequality itself
may be assessed. The demands of equality cannot be clearly inter-
preted or understood without taking adequate note of efficiency
considerations, The point is not merely that the demands of equality
have to be ultimately weighed against the force of competing de-
mands, when present. It is also that the interpretation of the de-
mands that equality makes has to be assessed in the lightof the other
considerations (e.g. aggregative concerns) that are inter alia recog-
nized, The explicit admittance of other concerns avoids the over-
burdening of equality with unnecessary loads. This general point
may be illustrated with a couple of examples.
First, as wasargued earlier (in Chapter 6, Section 6.2), the ease for
'attainment equality' as opposed to 'shortfall equality' may be
severely weakened if no weight is to be given to aggregative con-
siderations, but that in the presence of aggregative concerns, move-
ments towards attainment equality may have much that is commen-
dable. When person A's potentials permit a maximal achievement
of x, compared with a general maximal achievement of, say, 2x for
all others (the difference may be related, for example, to some
physical disability that person A has), a demand for equality as the
only consideration would tend to have the effect of levelling down
everyone else (withoutdisabilities)to x. Under those circumstances,
shortfall equality would seem to have some comparative merit, and
its programme of equating everyone's achievementsas a proportion
of their respective maximal values would be distinctly less ex-
tremist.
On the other hand, that programme would not yield a patently
just solution either. Person A is disadvantaged in having a lower
maximal achievement, and it is not clear why priority should not be
given to helping A to move as close to x as possible, even at the cost
33
Note that this definition of efficiency takes note of the possibility of incomple-
teness in the ranking of individual capabilities. Given extensive incompleteness, the
requirement that no one's capability be actually reduced would have been more
easily passed. Due to incompatabilities, a change can fail to keep a person at a level
that isjudged to be giving him at least the same overall capability without actually
leading to » decidable reduction of capability. This distinction can be particularly
important in policy judgements, e.g. in assessing the efficiency of competitive market
equilibria in terms of capabilities (on this, see Sen 1992&),

of some people having a lower proportionate achievement vis-a-vis
their own higher maximal 2x, What the situation demands is a
respect for attainment equality, without losing sight of aggregative
considerations, including the demands of efficiency. If the problem
is thus reformulated, person A will have preferential treatment in
being helped to move to his maximal achievement x, without all
others being dragged down to x (from their respective maximal
potential of 2x). The admittance of aggregative and efficiency con-
siderations, thus, removes from the sphere of equality a load that it
cannot comfortably carry, under either interpretation of equality,
No less importantly, the case for giving weight to attainment
equality comes into its own once that iscombined with a concern for
efficiency as well. The ad hoc case for shortfall equality would then
be cut back to size,34
A second example is chosen frorn Rawls's (1990) defence of using
primary goods as the basis of the Difference Principle in his 'justice
as fairness'. He is responding here particularly to a claim (made in
Sen 198Q«) that it is more fair to usecapabilities injudgingindividual
advantage,JS
Rawls argues against equating people's capabilities in
influencing public policy, and reasserts the fairness and justice of a
system in which influential offices are filled through open competi-
tion.3
* While people will have the same opportunity to compete for
these offices which are open to all (and thus, in some sense, enjoy the
same holding of these primary goods in terms of opportunities), they
will end up having unequal capabilities, Rawls points out that when
individuals differ in 'moral and intellectual capacities and skills',
there is nothing unfair or unjust in people with greater skills occupy-
ing influentialpositionsand offices.37
This linkswith the 'operational
asymmetry' argument discussed in the last section.
Let me begin by accepting the substantive conclusion of Rawls
that there may be no injustice in having a selection system for offices
and positions of responsibility that favours the more skilled. The
14
For a fuller discussion of this rather complex issue, see Sect. 6.2 in Ch. 6.
35
See also Ch. 5 above.
** Rawls (1971) demands equality in the distribution of primary goods unless
unequal divisions are advantageous for all. But the specification of primary goods
tends to concentrate on the opportunities (e.g. "holding positions open', p.61), 'All
social values—liberty and opportunity, income and wealth, and the bases of self-
respect—-are to be distributed equally unless an unequal distribution of any, or all,
of these values is to everyone's advantage' (p. 62).
37
Rawls (1990), Lecture 5, Sect. 3: 'Primary Goods and Interpersonal Com-
parisons*.

important question is; why do we agree with thisconclusion? This is
the case, I would argue, not because there is any intrinsic superiority
in more skilful people having more office-related capabilities than
others. The case for merit-based selection of officers and influential
positions relates, ultimately,to the efficiency of such a system.Being
born with lower mental skills is not a fault for which a person
should be penalized (for reasons that John Rawls, more than
anyone else, has taught us to understand). Indeed, if we could have
had equality in the holding of influential offices and positionswith-
out inefficiency and loss of advantages in general, justice would
have (I would argue) demanded that weconsider opting for it.What
makes us accept the inequality involved in this case is precisely the
impossibility of achieving that hypothetical position.
There is a significant inequality in the arrangements under discus-
sion—an inequalitythat can be understood in terms of differences in
capabilities and powers that different people would end up having.
The argument that such an inequality may nevertheless be accept-
able is best understood in terms of its efficiency advantages, and not
through denying that there is any real inequality here, since every-
one had the same opportunity of competing for offices, A significant
inequality has to be acknowledged first, before it is examined as to
whether it isjustified or not. Justifying the inequalityin capabilities
in the case discussed would take the form of arguing thateliminating
it would tend to pull downmany people's capabilities quite substan-
tially and that would be inefficient and unacceptable. The justifica-
tion iscontingent on the aggregativeconsideration workingthisway.
The contingent acceptability of the social arrangement that
Rawts defends does not, therefore, show that focusing on primary
goods is a fairer or better way ofjudging individual advantage than
capabilities. What it really suggests is that demands of equality of
individual advantages have to be supplemented by considerations of
efficiency in generating these advantages.
Rawls himself has outlined the need to consider the demands of
efficiency in his second principle of justice. But the use he can make
of efficiency considerations is somewhat limited by the insistence on
the extremism of giving total priority to the interests of the worst
off.18
This priority makes it harder to justify inequalities through
aggregative considerations, since the focus has to be, lexicographi-
cally, on the worst-off group only. The demands of Rawls'smaximin
3
» On this, see Sen (1970a, 19776).

formula is more easily fulfilled when we have a characterization of
individual advantages (in the form of holding primary goods) that
does not record inequalities that would be readily recorded in the
space of capabilities. But that absence of recording does make the
political scrutiny infonnationally poorer. There are good reasons, in-
stead, for recording significantinequalitiesin capabilities, and exam-
ining whethertheycan or cannot bejustified by efficiency arguments;
Sofar I havenot disputed the substantive claim of Rawls that there
is no unfairness or injustice in allocating offices through open com-
petition, with the more skilled being actually selected (as long as
everyone has the same opportunity to be educated and to compete).
Basing the justification of such inequalities explicitly on their effi-
ciency advantages makes it necessary to scrutinizethe causal connec-
tions related to the claims of efficiency. I do not doubt that inmany
circumstances the procedure that Rawls supports (and whichI tenta-
tively accepted earlier on) wouldprove to bejust right. If, on the other
hand, it turns out that a system by which offices and influential posi-
tions goto peoplewhodo better inopen competition creates a kindof
'meritocracy' that is not so efficient and which leads to people of less
favoured groups being unequally treated (in the exercise of those
offices and positions), then thatjustification would no longer obtain.
In the political debates that raged in India through 1990 on the
proposal of the then Prime Minister V. P, Singh that more than half
the influential jobs in civil service be reserved for members of lower
castes and other disadvantaged groups (the disputes were crucial to
V, P. Singh's eventual loss of parliamentary majority and the fall of
his government), the presumed efficiency advantages of selection
through open competition was severelyquestioned. It is not mypur-
pose heretoargue thatV,P, Singh'sscepticism of the efficiency advan-
tages was correct or incorrect, but only that the justice of the open-
competition arrangementmustbesensitiveto the answerthat wegive
to that question. The issue cannot be settled by a prior fixation of the
distributive pattern of primary goods (e.g. in the form of none of the
offices being reserved for the socially immobile underprivileged).
The justice of the arrangements must be sensitive to the respective
impacts of the different systems on aggregative and distributive
aspects of people's effective freedoms and capabilities. Such an
analysis could, of course, fully support Rawls's recommendation,
but if it did, then that would indeed be a powerful reason for that
recommendation.

9,9, RESPONSIBILITYAND FAIRNESS
John Rawls (1971) and other modern theorists of justice (such as
Ronald Dworkin 1981) have tended to stress the need to see each
person as being peculiarly responsible for matters over which she
has control. In contrast, responsibility is not attributed—nor credit
given—to a person for something she could not have changed (such
as having rich or poor parents, or having or lacking natural gifts),
The lines are sometimes hard to draw, but there ismuch plausibility
in that general differentiation. The analysis presented in this book
has made a good deal of use of that distinction,
Indeed, the criticism of Rawls's theory of'justice as fairness' from
the capability perspective arose partly from our attempt to take
direct note of a person's difficulties—naturally or socially
generated—in converting 'primary goods' into actual freedoms to
achieve (see Chapter 5).A person less able or gifted in using primary
goods to secure freedoms (e.g. because of physical or mental dis-
ability, or varying proneness to illness, or biological or conven-
tional constraints related to gender) is disadvantaged compared
with another more favourably placed in that respect even if both
have the same bundle of primary goods. A theory of justice, I have
argued, must take adequate note of that difference. It is for this
reason that the approach presented here both draws on and criticizes
the Rawlsian theory—it draws on Rawls's illuminating analysis of
fairness and of responsibility to criticize his theory's particular de-
pendence on the holding of primary goods (as opposed to the
freedoms and capabilities the persons respectively enjoy).
The distinction is of significance on another contentious issue, viz,
the choice between achievements and freedoms to judge a person's
relative situation. In dealing with responsible adults, it is more
appropriate to see the claims of individuals on the society (or the
demands of equity or justice) in terms of freedom to achieve rather
than actualachievements. If the social arrangements are such that a
responsible adult is given no less freedom (in terms of set com-
parisons) than others, but he still wastes the opportunities and ends
up worse off than others, it is possible to argue that no unjust in-
equality may be involved. If that view is taken, then the direct
relevance of capability (as opposed to achieved functionings) will
be easy to assert.
It is, however, important to be clear about some qualifications

that must apply to the preceding argument. For one thing, the issue
is quite different in the presence of uncertainties. The predicament of
a person due to adverse happenings over whichhe has no control can
scarcely be dismissed on grounds of personal responsibility,
That responsibility argument is more applicable when the person
himself willingly takes a risk and ends up losing the gamble, since
there will be more scope here for invoking the person's own respon-
sibility. But even here the picture may be made more complicated by
the difficulty of getting adequate information that would permit a
person to make intelligent decisions in risky situations. For
example, the collapse of a well-known insurance company or a
well-regarded bank can hardly be seen as an occasion for hard-
headed dismissal (with no special quarter being given to the unfor-
tunate victims) on the grounds that the victims themselves had
chosen the insurance company or the bank.39
In fact, the case for
concentrating on freedoms to achieve as opposed to actual achieve-
ments depends quite heavily on the knowledge and the ability of the
persons to understand and intelligently choose from the alterna-
tives they really do have.
A closely related issue concerns the way in which capability
accounting must take note of the real freedoms that people in fact
(not just "inprinciple*) enjoy. If social conditioning makes a person
lack the courage to choose (perhaps even to 'desire* what is denied
but what would be valued if chosen), then it would be unfair to
undertake the ethical assessment assuming that she does have that
effective choice. It is a matter of concentrating on the real freedoms
actually enjoyed, taking note of all the barriers—including those
from 'social discipline'.
Indeed, an overdependence on what people 'manage to desire' is
one of the limiting aspects of utilitarian ethics, which isparticularly
neglectful of the claims of those who are too subdued or broken to
have the courage to desire much.40
It would be particularly unfor-
tunate to err in the same way in the capability accounting. There is,
39
Interestingly, the British and the American public responses to such collapses
have been quite different, as illustrated by the enormous protectiveness with which
the 'Savings & Loans' crises have been dealt with in the USA,compared with the
relatively smaller compensations offered to the unlucky depositors of the busted-up
Bank of Credit and Commerce International (BCCI). Even the legal frameworks of
protection of depositors are quite different in the two countries, and involve rather
disparate views of responsibility.
40
This issue is discussed in Sen (198Sa).

150 The Demands of'Equality
however, no need to err in that way, since the capabilities to be
accounted are those that people do actually have (and not those that
they could have hadif they wereless influenced by 'social discipline').
This questionis of particular importance in dealing with entrenched
inequalities that are supported by the victims* conditioned accept-
ance of comparative deprivation (e.g. women's acceptance of sub-
jugated roles in traditional social arrangements).41
9.10. CAPABILITY, FREEDOM AND MOTIVATIONS
The 'capability approach' has something to offer both to the evalu-
ation of well-being and to the assessment of freedom. Considering the
former connection first, the capability approach to well-being
differs from the more traditional concentration on economic opu-
lence (in the form of real income, consumption levels, etc.) in two
distinct respects; (1) it shifts the focus from the space of meansin the
form of commodities and resources to that of functionings which are
seen as constitutiveelements of human well-being; and (2) it makesit
possible—though not obligatory—to take note of the set of alterna-
tive functioning vectors from which the person can choose. The
'capability set* can be seen as the overall freedom a person enjoys to
pursue her well-being.
If the ability to choose between substantivelyimportant alterna-
tives is seen as valuable in leading a worthwhile life, then the
capability set has a further role: it can be directly influential in the
determination of a person's well-being. An alternative combination
x offunctionings can, then, beseen as goingwitha certain freedom to
choose from a set $, to which x belongs. If the well-being that a
person gets from what she does isdependent on howshe came to do it
(in particular, whether she chose that functioningherself), then her
well-being depends not just on x, but on the choice of x from the set
S.42
There are complex problems involved in these issues (discussed in
Chapters 3 and 4). The crucial question here, in the context of well-
being, is whether freedom to choose is valued only instrumentally,
or is also important in itself. The capability approach is broad
41
On thisquestion, see Kynch and Sen (1983), Sen (19«Saf, S990c), Laden (19916).
41
The analysis of welt-being, then, has to be related to the pair (x, S), and not just
to x.

enough to permit both the rival—but interrelated—characteriz-
ations of well-being, and can be used in either way,
It is important to emphasize that even if freedom to choose is
valued only instrumentaHy in the determination of individual well-
being, the extent of freedom enjoyed by each person can, neverthe-
less, be directlyimportant for a good society, Indeed, whether or not
freedom enters individual well-being, individual freedom can be
seen as being constitutive of the goodness of the society which we
have reasons to pursue,43
Our interest in equality of freedoms can, therefore, be related to
different evaluative foundations. In thismonograph I have not parti-
cularly focused on the relative merits of these diverse ethical under-
pinnings,44
but it is useful to keep in view the different foundational
reasons for which we may wish to attach importance to the equality
of freedoms. The perspective provided by the capability approach
can be used to analyse and assess equality of freedoms in relation to
each of these distinct motivations.
Much the same plurality applies also to the approach of seeing
poverty as capability failure. This can be linked to various under-
lying concerns, such as guaranteeing minimal individualwell-being
or providing minima! individual freedoms, and these in turn can be
related to more foundational demands of good—or right—social
arrangements. The case for reorienting poverty analysis from low
incomes to insufficient basic capabilities can be, ultimately, con-
nected with these alternative foundational concerns.
No matter what exact foundational structure we opt for, the re-
orientation from an income-centred to a capability-centred view
gives us a better understanding of what is involved in the challenge
of poverty.45
It provides clearer guidance on the priorities of anti-
poverty policy and also helps us to understand better the genesis of
poverty in apparently unlikely circumstances (e.g. in the rich coun-
43
A related but different line of ethical analysis takes the form of making in-
dividual freedoms the subject-matter of the 'tightness' of socialarrangements, with-
out necessarily invoking any prior notion of the social good. That way of seeing the
problem has distinctly Kantian roots(Kant 1788). On the relevance of the Kantian
foundations for these and related issues, see particularly Rawls (1971, 19880); see
also O'Neill (1989, 1992), and Korsgaard (1992).
44
See,however, Sen (19826, 1985a).
4J
A probing empirical study, done by Schokkaert and VanOotegem (1990), of the
perception of poverty by the Belgian unemployed indicates that the unemployed
themselves may see poverty precisely as lack of substantive freedom.

tries of Europe and America), The wisdom of focusing on poverty
as lack of freedom is consistent with a variety of foundational con-
cerns.
This monograph has been aimed at examining the nature and
reach of the demands of equality. While the analysis has been
mainly conceptual, it has some direct bearing on matters of prac-
tical concern. The analysis has been very substantially motivated by
that connection.

References
ACKERMAN, B. A, (1980). Social Justice in the Liberal State (New Haven,
Conn.; Yale University Press).
—— (1984), "The Storrs Lectures; Discovering the Constitution', Yale Law
Journal, 93,
(1988). 'Neutralities,' mimeographed, Yale Law School.
ADKLMAN, I. (1975). 'Development Economics—A Reassessment of Goals,"
American Economic Review, Papers and Proceedings, 65.
and.M0.Rius, C, T. (1973). Economic Growthand Social Equity inDevel-
oping Countries (Stanford, Calif,: Stanford University Press).
AGARWAL, B. (1986). "Women, Poverty and Agricultural Growth in India*,
Journal of Peasant Studies, 13.
— (1991). 'Social Security and the Family: Coping with Seasonally and
Calamity in Rural India', in Ahmad et al. (1991).
AHLUWALIA, M. S, (1974). 'Income Inequality: Some Dimensions of the
Problem', in Chenery et al, (1974),
(1978). 'Rural Poverty and Agricultural Performance in India', Jour-
nal of Development Studies, 14.
AHMAD, E., and HUSSAIN, A. (1991). 'Social Security in China: A Historical
Perspective', in Ahmad et al. (1991).
Dreze, J., Hills, J,, and SEN, A, (1991) (eds.). Social Security in Devel-
oping Countries (Oxford: Clarendon Press),
AHMED, I. (1983). 'Technology and Rural Women in the Third World',
international Labour Review, 122.
AHOOJA-PATEL, K. (1980) (ed,). Women at Work (Geneva: ILO).
AJQNER, D. J., and HEINS, A, J. (1967). 'A Social Welfare View of the
Measurement of Income Inequality', Review of Income ami Wealth, 13.
AKERUJF, O. (1978). 'The Economics of "Tagging"", American Economic
Review, 68.
(1984). An Economic Theorist's Book of Tales (Cambridge: Cambridge
University Press),
ALAMGIR, M. (1978). Bangladesh: A Case of Below Poverty LevelEquilibrium
Trap (Dhaka: Bangladesh Institute of Development Studies).
—- (1980). Famine in South Asia—The Political Economy of Mass Starva-
tion in Bangladesh (Cambridge, Mass.: Oelgeschlager, Gunn and Main).
ALDWCH, J. (1977). 'The Dilemma of a Paretian Liberal: Some Conse-
quences of Sen's Theory", Public Choice, 30.
ALLARDT, E. (1981), 'Experiences from the Comparative Scandinavian Wel-
fare Study, with a Bibliography of the Project', European Journal of
Political Research, 9.

154 References
ALLARDT, E, (1992). 'Having, Loving,Being:An Alternative to the Swedish
Model of Welfare Research', in Nussbaum and Sen (1992),
ALLEN, T. (1988). The Impossibility of the Paretian Liberal and its Rele-
vance to Welfare Economies', Theory and Decision,24.
——• (1991), 'Economic Development and the Feminization of Poverty", in
Folbre et al. (1991).
AMIEL, Y,, and COWELL, P. (1989), 'Measurement of Income Inequality:
Experimental Test vs. Questionnaire', Discussion Paper TIDI/140,
STICERD, London School of Economics.
AMSDEN, A. H. (1980) (ed.). The Economics of Women and Work (Har-
mondsworth: Penguin).
ANAND, S. (1977). 'Aspects of Poverty in Malaysia*, Review of Income and
Wealth, 23.
(1983), Inequality and Poverty in Malaysia (NewYork: OxfordUniver-
sity Press).
——and HARRIS, C. (1990). 'Food and Standard of Living: An Analysis
Based on Sri Lankan Data', in Dreze and Sen (1990), vol. 1.
—— (1992). 'Issues intheMeasurementof Undernutrition', inOsmani
(1992a).
and KANBUR, R, (1984). 'Inequality and Development: A
Reconsideration*, in Nissen (1984).
———(1986). 'Inequality and Development: A Critique', mimeo-
graphed, St Catherine's College, Oxford, and WarwickUniversity.
— (1990). 'Public Policy and Basic Needs Provision; Intervention
and Achievement in Sri Lanka', in Dreze and Sen (1990), vol. 3.
ANCKAR, D., aad BERNDTSON, E. (1984) (eds.). Essays on Democratic Theory
(Tampere: Finnpublishers),
ANNAS, J. (1992). 'Women and the Quality of Life: Two Norms or One?', in
Nussbaum and Sen (1992).
APFPEI. MAROLIN, F.f and MARGLIN, S. A. (1990) (eds.). Dominating Knowl-
edge; Development, Culture, Resistance(Oxford: Clarendon Press).
ARCHIBALD, G. C., and DONALPSON, D. (1979). 'Notes on Economic In-
equality', Journal of Public Economics, 12,
ARISTOTLE. The Nicomachean Ethics, English trans, in Ross (1980).
ARNESON, R. (1989a). "Equality and Equality of Opportunity for Welfare',
Philosophical Studies, 56.
—— (19896). 'Paternalism, Utility and Fairness', Revue Internationale de
Philosophic, 43.
(1990a). 'Liberalism, Distributive Subjectivism, and Equal Oppor-
tunity for Welfare', Philosophy and Public Affairs, 19.
(19906), 'Primary Goods Reconsidered', Nous, 24.
—— (1990c), 'Neutrality and Utility', Canadian Journal of Philosophy, 20.
(1991). 'A Defence of Equal Opportunity for Welfare*, forthcomingin
Philosophical Studies.

References 155
ARROW, K. J. (1951). Social Choice and Individual Values (New York:
Wiley).
— (1963), Social Choice and Individual Values, 2nd extended edn. (New
York: Wiley).
—— (1977), 'Extended Sympathy and the Possibility of Social Choice*,
American EconomicReview,67.
— (1982a). 'Risk Perception in Psychology and Economies', Economic
Inquiry, 20,
(19S2A). 'Why People Go Hungry', New York Reviewof Books, 29 (15
July).
(1983). "Behaviour under Uncertainty and its Implications for
Policy', in B. P. Stigum and F. Wenstop (eds.). Foundations of Utility and
Risk Theory with Applications (Dordrecht: Reidel).
—- (1991) (ed.). Markets and Welfare (London; Maemillan).
—— and INTRILIGATOR, M. (1986). Handbook of Mathematical Economics,
iii (Amsterdam: North-Holland).
ARTICLE 19 (1990), Starvingin Silence:A Report on Famine and Censorship,
ed. Frances D'Souza (London: Article 19),
ASAHI, J. (19S7). 'On Professor Sen's Capability Theory', mimeographed,
Tokyo.
ASLANBEIQUI, N., and SUMMERHELB, G, (1989). 'Impact of the Responsi-
bility System on Women in Rural China: An Application of Sen's
Theory of Entitlements', World Development, 17.
ASPREMONT, C. D* (1985), 'Axioms for Social Welfare Ordering', in
Hurwicz, Schmeidler, and Sonnenschein (1985).
and GBVERS, L, (1977), "Equity and the Informational Basis of
Collective Choice', Review of Economic Studies, 46.
AHTISAARI, M. (199)). 'Amartya Sen's Capability Approach to the Stan-
dard of Living*, Jonathan Lieberson Memorial Prize Essay, mimeogra-
phed, Columbia University.
ATKINSON, A. B. (1970«). Poverty in Britain and the Reform of Social
Security (Cambridge: Cambridge UniversityPress).
—— (19706). 'On the Measurement of Inequality", Journal of Economic
Theory, 2, (Repr. in Atkinson 1983.)
—— (1972). Unequal Shares (London: Allen Lane, The Penguin Press).
——(1975). The Economicsof Inequality (Oxford: Clarendon Press).
-—- (1983). Social Justice and Public Policy (Brighton: Wheatsheaf and
Cambridge, Mass.: MIT Press),
(1987), 'On the Measurement of Poverty', Econometrica, 5.
(1989), Poverty and Social Security (New York; Harvester
Wheatsheaf).
(I991a). 'The Social Safety Net', mimeographed, STICERD, London
School of Economics.
——(19916). 'What is Happening to the Distribution of Income in the

156 References
UK?', Keynes Lecture, British Academy,
(1991c). 'On Targeting', mimeographed, STICERD, London School
of Economics,
ATKINSON A. B. and BOURGUIONON, F, (1982), 'The Comparison of Multi-
dimensional Distributions of Economic Status*, Review of Economic
Studies, 49,
—, (1987), 'Income Distribution and Differences in Needs', in
Feiwel (1987),
—— and STIOLITZ, J. B, (1980). Lectures on Public Economics (London:
McGraw-Hill),
AUERBACH, A,, and FBLDSTEIN, M. (1987), Handbook of Public Economics
(Amsterdam: North-Holland).
AUSTEN-SMITH, D. (1979). 'Fair Rights', Economic Letters, 4,
(1982). 'Restricted Pareto and Rights', Journal of Economic Theory
26.
BABU, S, C, (1986). 'Identifying the Poor—A Development Approach:
Case Study of a South Indian Village', unpub, manuscript, Iowa State
University.
BALL, S, W. (1987). 'Choosing between Choice Models of Ethics:
Rawlsian Equality, Utilitarianism, and the Concept of Persons', Theory
and Decision, 22: 3,
BANERJEE, N. (1979), 'Women in Urban Labour Market', Labour Capital
and Society, 12.
(1982). Unorganised Women Workers: The Calcutta Experience
(Calcutta: Centre for Studies in Social Sciences).
BANISTER, J. (1987). China's Changing Population (Stanford, Calif.:
Stanford University Press),
BARDHAN, K. (1985). "Women's Work, Welfare and Status', Economic and
Political Weekly, 20 (21-8 Dec.).
BAKDHAN, P. (1974a). "On the Incidence of Poverty in Rural India in the
Sixties', in Srinivasanand Bardhan (1974).
——(19746). "On Life and Death Questions', Economic and Political
Weekly, 9 (Special Number).
(1984). Land Labour and Rural Poverty: Essays in Development Econ-
omics (New York: Columbia University Press),
— (1987). 'On the Economic Geography of Sex Disparity in Child
Survival in India', mimeographed, University of California, Berkeley.
—— (1988). 'Sex Disparity in Child Survivalin Rural India', in Srinivasan
and Bardhan (1988).
BARNES, J. (1980). 'Freedom, Rationality and Paradox', Canadian Journal
of Philosophy, 10.
BARMY, B. (1986). 'Lady Chatterly's Lover and Doctor Fischer's Bomb
Party: Liberalism, Pareto Optimality and the Problem of Objectional
Preferences', in Elster and Hylland (1986).

References 151
BAITEN, A. P, (1964), 'Family Composition, Prices and Expenditure Pat-
terns', in P. Hart et al, (ed$.)> Econometric Analysisfor National Economic
Planning (London: Butterworth),
BASU, A. (1988). Culture, the Status of Women and Demographic Behaviour
(New Delhi: National Council of Applied Economic Research).
et al, (1986), 'Sex Bias in Intrahousehold Food Distribution: Roles of
Ethnic and Soeioeconomic Characteristics', Current Anthropology, 27, .
BASU, K. (1979). Revealed Preference of Governments (Cambridge; Cam-
bridge University Press).
(1984). The Right to Give Up Rights', Economica,51.
—— (I987a), 'Achievements, Capabilities and the Concept of Weil-Being',
Social Choice and Welfare, 4.
—— (1987i). 'Axioms for Fuzzy Measures of Inequality', Mathematical
Social Sciences, 14,
— (1989). *A Theory of Association: Social Status, Prices and Markets',
Oxford Economic Papers, 41.
BATEA, R., and PATTANAIK, P. (1972). 'On Some Suggestions for Having
Non-Binary Social Choice Functions", Theory and Decision,3.1-11.
BAUER, P, T. (1981).Equality, the Third World andEconomicDelusion (Cam-
bridge, Mass.: Harvard University Press).
BAUMOL, W. J, (1986). Superfairness (Cambridge, MA: MIT Press).
BAYER, R.,CAPLAN,A. L,,and DANIEIS, N. (1983)(eds.).In Searchof Equity:
Health Needs and the Health Care System (New York: Plenum
Press),
BICKER, O. S. (1965). *A Theory of the Allocation of Time', Economic Jour~
nal, 75.
(1981).A Treatiseontfief'amily(Cambndge, Mass.:Harvard University
Press).
BECKBRMAN, W. (1979). The Impact of Income Maintenance Programmes on
Poverty in Four Developing Countries (Geneva: ILO).
— and CLARK, S.(1982). Poverty andSocial Security inBritain since 1961
(Oxford: Clarendon Press),
BEPAU, H. A. (1971) (ed.). Justice and Equality (Englewood Cliffs, NJ:
Prentice-Hall).
BEHKMAN, J. R. (1988). 'Intrahousehold Allocation of Nutrients in Rural
India: Are BoysFavoured? Do Parents Exhibit Inequality Aversion?* Ox-
ford Economic Papers, 40.
(1992), 'Intrahousehold Allocation of Nutrientsand Gender Effects; A
Survey of Structural and Reduced-form Estimates', in Osmani (1992a),
and DEOLALIKAR, A, B.(1988), 'Health and Nutrition', in Chenery and
Srinivasan (1988).
—— POLLAK, R,, and TAUBMAN, P, (1989). 'Family Resources, Family Size,
and Access to Financing for College Education", Journal of Political
Economy, 97.

158 References
BEITZ, C. R, (1986), 'Amartya Sen's Resources, Values and Development',
Economics and Philosophy, 2.
BELLMAN, R. E., and ZADEH, L, A. (1970). 'Decision Making in a Fuzzy
Environment', Management Science, 17.
BENERIA, O, (1982) (ed.). Women and Development: The Sexual Division of
Labour in Rural Societies (New York: Praeger).
BBNTHAM, J, (1789). An Introduction to the Principles of Morals and Legis-
lation (London: Payne). (RepubJished Oxford: Clarendon Press, 1907.)
BENTZEL, R, (1970). 'The Social Significance of Income Distribution
Statistics', Review of Income and Wealth, 16.
BEROMAN, B. (1986), The EconomicEmergence of Women (New York: Basic
Books).
BEROSON, A. (1938). 'A Reformulation of Certain Aspects of Welfare
Economies', Quarterly Journal of Economics, 52,
—— (1966). Essays in Normative Economics (Cambridge, Mass.: Harvard
University Press).
BERLIN, I. (1955-6). 'Equality as an Ideal', Proceedings of the Aristotelian
Society, 56,
—— (1969). Four Essays on Liberty, 2nd cdn, (London: Oxford University
Press).
BEKNHOLZ, P. (1974). 'Is a Paretian Liberal Really Impossible?', Public
Choice, 19.
—— (1980), 'A General Social Dilemma: Profitable Exchange and
Intransitive Group Preferences', Zeitschrlft fur Nationalokonomie, 40.
BESLEY, T. (1989). 'Ex Ante Evaluation of Health Status and the Provision
for Ill-Health", Economic Journal, 99,
and KANBUR, R. (1988), 'Food Subsidies and Poverty Alleviation',
Economic Journal, 98,
BtreiLLE, A. (1969) (ed,). Social Inequality (Harmondsworth: Penguin
Books),
-—-(1977), Inequality among Men (Oxford: Blackwell).
-(1981). The Backward Classes and the New Social Order (Delhi:
Oxford University Press),
_ (1983a). The Idea of Natural Inequality and Other Essays (Delhi:
Oxford University Press),
— (19836) (ed.). Equality and Inequality (Delhi: Oxford University
Press).
—— (1987). 'Equality as a Right and as a Policy', LSE Quarterly,1.
(1990). 'Distributive Justice and Institutional Well-Being', text of
V. T. Krishaamachari Lecture, mimeographed, Delhi School of Econ-
omics.
BEDS, J. DE (1989). Markt, Democratic en Vrijhe (Market, Democracy and
Freedom) (Amsterdam: Tjeenk Willink, Zwolle), in Dutch.
BEZEMBINDER, TH.. and VANACKER, P. (1979). 'A Note on Sen's Partial

References 159
Comparability Model*, mimeographed, Department of Psychology,
Katholieke Universiteit, Nijmegcn.
(1986), Tactual versus Representational Utilities and their
Interdimensional Comparisons', mimeographed, Department of Psy-
chology, Katholieke Universiteit, Nijmegen.
BHAROAVA, A. (1991). 'Malnutrition and the Role of Individual Variation
with Evidence from India and the Phillippines*, mimeographed, Depart-
ment of Economics, Universityof Houston.
BHATTACHARYA, N., and CBATTBRJEE, G. S. (1988), 'A Further Note on
Between-States Variation in Level of Living in Rural India*, in Srinivasan
and Bardhan (1988),
— and PAL, P. (1988). 'Variations in Levels of Living across Regions and
Social Groups', in Srinivasan and Bardhan (1988).
BHATTY, I. Z. (1974). 'Inequality and Poverty in Rural India", in Srinivasan
and Bardhan (1974),
BHATTY, Z. (1980). 'Economic Role and the Status of Women: A Case Study
of Women in the Beedi Industry in Allahabad*, ILO working
paper.
BHUIYA, A., et at, (1986). 'Socioeconomic Determinants of Child's Nutri-
tional Status; Boys versus Girls', Food and Nutrition Bulletin, 8.
BIGMAN, D. (1985). 'Aggregate Poverty Measures and the Aggregation of
Individual Poverty: A Reconsideration of Sen's Axiomatic Approach*,
unpub. manuscript, The Hebrew Universityof Jerusalem,
(1986). 'On the Measurement of Poverty and Deprivation*, Working
Paper No. 8602, The Center for Agricultural Economics Research, The
Hebrew University of Jerusalem.
BINMQRE, K., and DASGUPTA, P, (1987) (eds,). The Economics of Bargaining
(Oxford: Blackwell).
BISWAS, T. (1987). 'Distributive Justice and Allocation by the Market: On
the Characterization of a Fair Market Economy', Mathematical Social
Sciences, 14.
BLACKORBY, C. (1975). 'Degrees of Cardinality and Aggregate Partial
Ordering', Econometrica, 43.
——and DONALDSON, D. (1977). 'Utility versus Equity: Some Plausible
Quasi-Orderings', Journal of Public Economics, 7.
——-— (1978). 'Measures of Relative Equality and their Meaning in
Terms of Social Welfare', Journal of Economic Theory, 18.
——•— (1980). 'Ethical Indices for the Measurement of Poverty", Eco-
nometrica, 48,
— (1984). 'Ethically Significant Ordinal Indexes of Relative In-
equality', Advances in Econometrics,3.
— (1988). 'Adult-Equivalence Scales and the Economic Implemen-
tation of Interpersonal Comparisons of Well-Being*, mimeographed,
University of British Columbia.

160 References
——-—- and AUERSPEKG, M, (1981). 'A New Procedure for the Measure-
ment of Inequality within and among Population Subgroups', Canadian
Journal of Economics, 14.
BLACKORBY, C., DONALDSON, D., and WEYMAIK, J, (1984). 'Social Choice
with Interpersonal Utility Comparisons: A Diagrammatic Introduction',
International Economic Review, 25.
—— (1990). 'A Welfarist Proof of Arrow's Theorem', Recherches
Iconomiques de Louvain, 56,
BLAIR, D. H. (1988). 'The Primary-Goods Indexation Problem in Rawls'
Theory of Justice', Theory and Decision, 24,
BLAU, J. H. (1975). 'Liberal Values aod Independence', Review of Economic
Studies, 42,
BLAXTEK, K., and WATERUJW, J. C. (1985) (eds.). Nutritional Adaptation in
Mem (London: John Libbey).
Buss, C. J. (1992). 'Life-Style and the Standard of Living', in Nussbaum
and Sen (1992),
BORDA, J, C, (1781). 'Memoke sur les elections aw scrutin', M&naires des
t'Academie Royals des Sciences,. English trans, by A. de Graaa, Isis, 44
(1953).
Bos, D., ROSE, M., and SEIDL, C. (1986) (eds.). Welfare and Efficiency in
Public Economics (Berlin; Springer-Verlag).
BOSERUP, E. (1970). Women's Role in Economic Development (London:
Allen and Unwin).
—~ (1987). "Inequalitybetween the Sexes', in J. Eatwell, M. Milgate, and
P. Newman (eds.), The New Palgrave; A Dictionary of Economics, ii
(London: Macmillan).
— (1990).'EconomicChangeand theRoles ofWomen'»inTinker(1990fl),
BouiauiQNON, P. (1979), 'Decomposable Income Inequality Measures',
Ecanantetrica, 47.
—— (1990). 'Growth and Inequality in the Dual Model of Development',
Review of Economic Studies, 57.
~— and FIELDS, G. (1990). 'Poverty Measures and Anti-Poverty Policy*,
Recherches tconomiques de Louvain, 56.
BRAITHWAITE, R. B. (1955). Theory of Games as a Toolfor the Moral Philos-
opher (Cambridge: Cambridge University Press).
BEANDT, R. B. (1979). A Theory of the Good and the Right (Oxford: Claren-
don Press).
BRANNIN, J., and WILSON, G. (1987) (eds.). Give and Take in Families
(London: Allen and Unwin).
BRINKERT, C. G, (1980). 'Freedom and Private Property in Marx', in
M. Cohen, Nagel, and Scanlon (1980).
—— (1983). Marx's Ethics of Freedom (London: Routledge),
BREYER, F. (1977). "The Liberal Paradox, Decisiveness over Issues, and
Domain Restrictions', Zeitschrift fur Nationalokonomie, 37.

References 161
— and GARDNER, R. (1980). 'Liberal Paradox, Game Equilibrium and
Gibbard Optimum', Public Choice, 35.
and GIGUOTTI, G, A. (1980). 'Empathy and the Respect for the Right of
Others', Zeitschrift fur Nationalokanomie, 40.
BROCK, D, (1992). 'Quality of Life Measures in Health Care and Medical
Ethics', in Nussbaum and Sen (1992).
BiooER, I, E., and MORRIS, C. T. (1983). 'Socially Weighted Real
Income Comparisons; An Application to India', World Development,
11.
BROOME, J. (1978). 'Choice and Value in Economies', Oxford Economic
Papers, 30.
(1987). 'What's the Good of Equality?' in J, Hey (ed.), Current
Issues in Macroeconomics (London: Macmillan),
(1988), 'What is Wrong with Poverty', London Review of Books, 10
(19 May).
BROWN, M., and CHUANG, C. P. (1980). 'Intra-Household Power and
Demand for Shared Goods', mimeographed, SUNY, Buffalo.
BRYCESQN, D. F, (1985). Women and Technology in Developing Countries
(Santo Domingo: United Nations).
BUCHANAN, A. B. (1982). Marx and Justice (London: Methuen).
BUCHANAN, J. M. (1975). The Limits of Liberty (Chicago: University of
Chicago Press).
— (1984). *On the Ethical Limits of Taxation*, Scandinavian Journal
of Economics, 86.
(1986). Liberty, Market and the State (Brighton: Wheatsheaf Books).
— and TuLiocs, G. (1962). The Calculus of Consent (Ann Arbor,
Mich.: University of Michigan Press).
BUHMANN, B., RAINWATER, L., SCHMAUS, G., and SMEEDING, T. M. (1988).
'Equivalence Scales, Well-Being, Inequality and Poverty: Sensitivity
Estimates across Ten Countries using the Luxembourg Income Study
Data Base', Meview of Income and Wealth, 34.
BURMAN, S. (1979) (ed.). Fit Work for Women (London: Croom Helm).
Buvimc, M., LYCETTE, M., and MCGREEVEY, W. (1983) (eds.). Women
and Poverty in the Third World (Baltimore: Johns Hopkins Press).
CAMPBELL, D. E. (1976). 'Democratic Preference Functions', Journal of
Economic Theory, 12.
(1989). 'Equilibrium and Efficiency with Property Rights and Local
Consumption Externalities', Social Choice and Welfaret 6.
CASE, A., and KATZ, L. (1990). The Company You Keep: Effects on
Family and Neighborhood on Disadvantaged Youth', mimeographed,
Harvard University.
CASSEN, R. (1978). India; Population, Economy, Society (London: Mac-
millan).
CHAKRAVARTY, L, (1986). 'Poverty Studies in the Context of Agricul-

162 References
tural Growth and Demographic Pressure (Case of Post-Independence
India)*, mimeographed, Indraprastha College, Delhi University,
CHAKRAVARTY, S, (1969), Capital and Development Planning (Cambridge,
Mass.: MIT Press),
CHAKRAVARTY, S. R. (1981), 'On Measurement of Income Inequality and
Poverty*, Ph.D. thesis (Indian Statistical Institute, Calcutta).
(I983a). 'Ethically Flexible Measures of Poverty', Canadian Journal
of Economics, 16,
— (1983ft). 'Measures of Poverty Based on Income Gap', Sankhya, 45.
(1984), 'Normative Indices for Measuring Social Mobility', Econ-
omics Letters, 15.
—— (1988), 'Extended Gini Indices of Inequality", InternationalEconomic
Review, 29,
—— (1990). 'Distributional Implications of Minimum Sacrifice Principle*,
mimeographed, Indian Statistical Institute.
—- and CHAKRABORTY, A. B. (1984), 'On. Indices of Relative Depriva-
tion', EconomicsLetters, 14.
— and DUTTA, B. (1987). 'A Note on Measures of Distance between
Income Distributions', Journal of Economic Theory, 41.
-——and ROY, T. (1985). 'Measurement of Fuzziness: A General
Approach', Theory and Decision, 19.
DUTTA, B., and WEYMARK, D. (1985). 'Ethical Indices, of Income
Mobility*, Social Choice and Welfare, 2,
CHAPMAN, B, (1983). 'Rights as Constraints: Nozick vs. Sen', Theory and
Decision, 15.
CHEN, L. C. (1986). 'Primary Health Care in Developing Countries; Over-
coming Operational, Technical and Social Barriers", Lancet, 2.
— HUQ, E,, and D'SouzA, D. (1981). 'Sex Bias in the Family Allocation
of Food and Health Care in Rural Bangladesh', Population and'De-
velopment Review, 7,
CHEN, M. (1986a). A Quiet Revolution; Women in Transition in Rural
Bangladesh (Dhaka: BRAG),
—— (1986&). 'Poverty, Gender and Work in Bangladesh', Economic and
Political Weekly, 21 (1 Feb.).
(1991). 'A Matter of Survival: Women's Right to Work in India and
Bangladesh", paper presented at the WIDER conference on 'Human
Capabilities; Women, Men and Equality*, 14-16 August.
CHENERY, H. B,, and SRINJVASAN, T, N. (1988) (eds.). Handbook of De-
velopment Economics, i (Amsterdam: North-Holland),
AHLUWAUA, M. S., BELL, C. L. G., DUIQY, J. H., and JOLLY, R.
(1974) (eds.). Redistributionwith Growth (Oxford: Clarendon Press),
CmemLNiSKY, O. (1980). 'Basic Needs and Global Models', Alternatives,6,
CHIPMAN, J, (1985). The Theory of Measurementof Income Distribution',
Advances in Econometrics,4.

References 163
CLARK, J, B. (1902). Distributionof Wealth (London; Macmtllan).
CLARK, S., and HEMMING, R, (1981). 'Aspects of Household Poverty in
Britain*, Social Policy and Administration, 15.
__——and ULPH, D, (1981). 'On Indices for the Measurement of
Poverty", Economic Journal, 91.
CLEMHOOT, A. (1979). 'A Life-Cycle Theory of Marriage and Divorce: A
Pareto Optima! Differential Game Model*, in P, Liu and J, F. Sutioen
(eds.), Control Theory in Mathematical Economics(New York: Marcel
Dekker).
CLEMHOUT, S., and WAN, H. Y., JR. (1979). 'Symmetric Marriage, House-
hold Decision Making and Impact on Fertility', Working Paper 152,
Cornell University.
COHEN, G, A. (1978).
— (1986). 'Self-Ownership, World Ownership, and Equality', Social
Philosophy and Policy, 3.
——(1988). History, Labour and Freedom: Themes from Marx (Oxford:
Clarendon Press).
(1989). 'On the Currency of Egalitarian Justice', Ethics, 99,
—— (1990), 'Equality of What? On Welfare, Goods and Capabilities', Re-
cherches economiques de Louvain, 56.
—— (1992). 'Equality of What? On Welfare, Resources and Capabilities',
in Nwssbaum and Sen (1992).
COHEN, M., NAGBL, T. and SCANLON, T. (1980) (eds.). Marx, Justice and
History (Princeton, NJ: Princeton University Press).
COLANDER, D. (1984) (ed.). Neoclassical Political Economy (Boston: Ball-
inger Press).
COOPER, T, C. (1971), 'Poverty*, unpub, manuscript, St Hugh's College,
Oxford.
COUGHUN, P. C. (1986), "Libertarian Concessions of the Private Pareto
Rule*, Econotnica, 53.
COWBLL, F. A. (1977). Measuring Inequality (New York: Wiley),
— (1980), *On the Structure of Additive Inequality Measures', Reviewof
Economic Studies, 47.
(1985). 'Measures of Distributional Change: An Axiomatic
Approach', Review of Economic Studies, 52.
—— (1986), 'Poverty Measures, Inequality and Deeoroposability', in Bos,
Rose, and Seidl (1986).
—— (1988). 'Social Welfare Functions and Income Inequality', London
School of Economics.
and KUOA, K. (1981), Inequality Measurement: An Axiomatic
Approach', European Economic Review, 15.
CROCKER, D. A. (1983). Praxis and Democratic Socialism (Brighton: Har-
vester Press).
(1991a). 'Toward Development Ethics', World Development, 19.

164 References
——-(19916), 'Functioning and Capability; The Foundations of Sen's De-
velopment Ethic", IDEA Montclair Workshop, mimeographed, Color-
ado StateUniversity.
CROCKER, L, (1980). Positive Liberty (London: Martinus Nijhoff).
CULVER, A. J. (1986), 'The Scope and Limits of Health Economies', in
Gerard GSfgen (ed,), ffkonomie des Gesundheitswesens (Berlin: Dunker
and Humbolt).
DAHLBY, B. G, (1987). 'Interpreting Inequality Measures in a Harsanyi
Framework", Theory and Decision, 22.
DAHKENDORF, R, (1988). The Modem Social Conflict: An Essay on the Politics
of Liberty (London: Weidenfeld & Nicolson),
DALTON, H. (1920). 'The Measurement of the Inequality of Incomes', Econ-
omic Journal, 30.
(1925), Inequality of Incomes (London; Routledge).
DANIELS, H. (1983), 'Health Care Needsand Distributive Justice', in Bayer,
Caplan, and Daniels (1983).
DANIELS, N, (1975) (ed,). Reading Rawls (Oxford: Blackwell).
DANZIOEK, S. H., and WEINBBBO, D, H, (1986). Fighting Poverty: What
Works and What Doesn't (Cambridge, Mass.: Harvard University
Press).
DAS, V., and NICHOLAS, R. (1981). '"Welfare" and "Well-Being" in South
Asian Societies', ACLS-SSRC Joint Committee on South Asia (New
York: SSRC).
DAS GUPTA, M. (1987). 'Selective Discrimination against FemaleChildren
in Rural Punjab, India', Population and Development Review, 13.
DASGUPTA, B. (1977). Village Society and Labour Use (Delhi: Oxford
University Press).
DASGOPTA, P, (1980). 'Decentralization and Rights', Economics, 47.
— (1982), The Control of Resources (Oxford;Blackwell).
(1986). 'Positive Freedom, Markets and the Welfare State', Oxford
Review of Economic Policy, 2,
— (1988). 'Lives and Well-Being', Social Choice and Welfare, 5.
(1989), 'Power and Control in the Good Polity', in A. Hamlin and
Pettit (1989),
(1990). 'Weil-Being and the Extent of its Realization in Poor Coun-
tries', Economic Journal, 100.
and HAMMOND, P, (1980). 'Fully ProgressiveTaxation',Journal of Pub-
lic Economics, 13.
—— and HEAL, G, (1979). Economic Theory and Exhaustible Resources
(London: James Nisbet and Cambridge: Cambridge University Press),
—— and RAY, D. (1990). 'Adapting to Undernutrition: Clinical Evidence
and its Implications', in Dreze and Sen (1990), vol. 1.
— SEN, A. K., and STARRETT, D, (1973). 'Notes on the Measurement of
Inequality', Journal of Economic Theory, 6,

References 165
DAVIDSON, D, (1986). 'Judging Interpersonal Interests', in Elster and
Hylland (1986).
DEATON, A, (1980), 'The Measurement of Welfare: Theory and Practical
Guidelines', LSMS Working Paper 7, The World Bank.
(1987), 'The Allocation of Goods within the Household: Adults,
Children and Gender', mimeographed, Princeton University.
(1988). 'Household Behavior in Developing Countries', Occasional
Paper No. 1, Economic Growth Center, Yale University.
and MUELLBAUER, J. (1980). Economics and Consumer Behaviour
(Cambridge: Cambridge University Press).
,__ (1986), 'On Measuring Child Costs with Applications to Poor
Countries', Journal of Political Economy,94.
DEB, R. (1989). "Rightsas Alternative Game Forms: Is There a Difference
of Consequence', mimeographed, Southern Methodist University,
DELBONO, F. (1986). Review article on Commodities and Capabilities, Econ-
omic Notes (Siena), 15.
DE LEONARDO, O., MAURIE, D., and ROTELLI, F. (1986). 'Deinstit-
utionalization, Another Way: The Italian Mental Health Reform*,
Health Promotion, I .
DEN HARTOO, A. P. (1973). 'Unequal Distribution of Food within the
Household", FAQ Newsletter, 10: 4 (Oct.-Dec.),
DBSAI, M. J. (19i4). 'A General Theory of Poverty*, Indian Economic
Review, 19.
(1988). 'The Economics of Famine', in Harrison (1988).
• (1990). 'Poverty and Capability: Towards an Empirically Implemen-
table Measure', mimeographed, London School of Economics.
——and SHAH, A. (1988). *An Econometric Approach to the Measure-
ment of Poverty', Oxford Economic Papers, 40,
BOLTVINNHC, J., and SEN, A. K. (1991). Social Progress Index
(Bogota: UNDP).
DIEWERT, W., and MONTMAEQUETTB, C. (1983) (eds.). Price Level Measure-
ment (Ottawa: Statistics Canada).
DIXON, R. (1982). 'Mobilizing Women for Rural Employment in South
Asia; Issues of Class, Caste and Patronage', Economic Development and
Cultural Change, 30,
—-(1983). 'Land Labour and the Sex Composition of the Agricultural
Labour Force: An International Comparison*, Development and Change,
14.
DQBB, M, H. (1937). Political Economy of Capitalism (London: Routledge).
DONALDSON, D., and ROEMER, J. E. (1987). 'Social Choice in Economic
Environments with Dimensional Variation*, Social Choice and Welfare,
4.
and WEYMARK, J. A. (1980). 'A Single-Parameter Generalization of
the Gini Indices of Inequality', Journal of Economic Theory, 22.

166 References
—___ (1986). 'Properties of Fixed-Population Poverty Indices', Inter'
national Economic Review,27.
DREWNOWSKI, J. (1978). The Affluence Line', SocialIndicators Research, 5.
DnizE, J,, and SEN, A. (1989). Hunger and Public Action (Oxford: Claren-
don Press).
——(1990). The Political Economy of Hunger, 3 vols. (Oxford:
Clarendon Press).
— and STEIN, N. H. (1987). 'The Theory of Cost Benefit Analysis', in
Auerbaeh and Feldstein (1987).
D'SouzA, F. (1988), 'Famine: Social Security and an Analysis of Vul-
nerability', in Harrison (1988).
—-(1990). 'Preface', in Article 19 (1990).
DUTTA, B. (1978), 'On the Measurement of Poverty in Rural India', Indian
Economic Meview, 15,
(1980). 'Intersectoral Disparities and Income Distribution in India;
1960-61 to 1973-74', Indian EconomicReview, 15.
— and RAY, D, (1989), 'A Concept of Egalitarianism under Participation
Constraints', Econometrica, 57.
DWORKIN, 0. (1982). 'Is More Choice Better than Less?' Midwest Studiesin
Philosophy, 7,
DWORKIN, R, (1978). Taking Rights Seriously, 2nd edn, (London;
Duckworth).
(1981), 'What is Equality? Part I: Equality of Welfare', and 'What is
Equality? Part 2: Equality of Resources', Philosophy and Public Affairs, 10.
—— (1985). A Matter of Principle (Cambridge, Mass.: Harvard University
Press),
(1987). 'What is Equality? HI:The Place of Liberty', Iowa Law Review,
73.
DYSON, T. (1987). 'Excess Female Mortalityin India: Uncertain Evidence on
a Narrowing Differential', mimeographed, London School of Economics.
EBBRT, U. (1987). 'Size and Distribution of Incomes as Determinants of
Social Welfare', Journal of Economic Theory, 41.
(1988). 'Measurement of Inequality: An Attempt at Unification and
Generalization*, Social Choice and Welfare, 5.
EDOEWORTH, F, Y. (1881). Mathematical Psychics (London: Kegan Paul),
EDWARDS, R, C., REICH, ML,and WEISSKOPF, T. (1986)(eds.). The Capitalist
System (Englewood Cliffs, NJ: Preatice-Hall),
EICHHORN, W. (1988a) (ed.). Measurement in Economics: Theory and Appli-
cations in Economic Indices (Heidelberg: Physica-Verlag).
—— (19886). 'On a Class of Inequality Measures', Social Choice and Wel-
fare, 5.
and OEHRJO, W, (1982).'Measurement of Inequality in Economies', in
B. Korte (ed.), Modern Applied Mathematics—Optimization and Opera-
tions Research {Amsterdam: North-Holland).

References 167
ELSTER, J, (1979), Ulysses and the Sirens (Cambridge; Cambridge University
Press),
— (1986). Making Sense of Marx (Cambridge; Cambridge University
Press).
——'and HYLIAND, A. (1986) (eds). Foundations of Social Choice Theory
(Cambridge; Cambridge University Press).
EWKSON, R. (1992), 'Descriptions of Inequality; The Swedish Approach to
Welfare Research", in Nussbaum and Sen (1992).
— and ABERG, R. (1987). Welfare in Transition: A Survey of Living Con-
ditions in Sweden (1968-81j (Oxford: Clarendon Press),
— H&NSEN, E, J,, RINOEN, S., and UUSITALQ, H. (1986). The Scandinavian
Model: The Welfare State and Welfare Research (New York: Sharps).
EVANS, M. (1982) (ed.). The Woman Question: Readings on the Subordination
of Women (London: Fontana),
FARMER, A. (1988). The Developing World (Dublin: Development Educa-
tion Support Centre).
FAKEELL, M, J. (1976). 'Liberalism in the Theory of Social Choice", Review
of Economic Studies, 43.
FEIWEL, O, R. (1987) (ed,). Arrow and the Foundations of the Theory of
Economic Policy (London: Macmillan).
FERETOHN, J, A. (1978). 'The Distribution of Rights in Society*, in Gottinger
and Leinfellner (1978),
FIELDS, G. S. (1979). *A Welfare Economic Approach to Growth and
Distribution in the Dual Economy', Quarterly Journal of Economics,
93.
(1980a). Poverty, Inequality and Development (Cambridge: Cambridge
University Press).
—— (19806). 'Reply*, AmericanEconomic Meview, 70.
(1990). 'Do Inequality Measures Measure Inequality?' mimeo-
graphed, Cornell University.
and FBI, J. C, H. (1978). 'On InequalityComparisons', Econometrics,
46.
——and JAKUBSON, G. H. (1990). 'The Inequality-Development Rela-
tionship in Developing Countries*, mimeographed, Cornell University.
FINE, B. J. (1975a). 'A Note on Interpersonal Aggregation and Partial
Comparability', Econometrica, 43.
— (19756), 'Individual Liberalism ina Paretian Society',Journal of Politi-
cal Economy, 83.
— (1985). 'A Note on the Measurement of Inequalityand Interpersonal
Comparability", Social Choice and Welfare, I.
FISHER, F. M. (1956). 'Income Distribution, Value Judgments and Wel-
fare*, Quarterly Journal of Economics, 70.
(1986). 'Household Equivalence Scales and Interpersonal Com-
parisons', mimeographed, MIT.

168 References
(1987). 'Household Equivalence Scales and Interpersonal Com-
parisons', Review of Economic Studies, 54.
-—-and SHELL, K. (1972). The Economic Theory of Price Indices (New
York: Academic Press).
FISHKIN, J, S.(1983). 'Can There Bea Neutral Theory of Justice?* Ethics, 93.
FisHtow, A. (1980). 'Who Benefits from Economic Developtneot? Com-
ment*, AmericanEconomic Meview, 70,
FLOOD, R. (1992). 'Antfaropometrie Measures of Nutritional Status in
Industrial Societies: Europe and North America since 1750', in Os-
mani (19920).
—-and WACHTER, K. W. (1982), 'Poverty and Physical Stature: Evi-
dence on the Standard of Living of London Boys 1770-1870', Social
Science History, 6.
FOQEL, R. W, (1986). 'Nutrition and the Decline in Mortality since
1700: Some Additional Preliminary Findings', Working Paper 182,
National Bureau of Economic Research, Cambridge, Mass.
— (1992). "Second Thoughts on the European Escape from Hunger:
Crop Yields, Price Elasticities, Entitlements and Mortality Rates', in
Osmani (I992a).
—— ENGERMAN, S. L., and TMJSSELL, J. (1982). 'Exploring the Use of
Data on Height: The Analysis of Long Term Trends in Nutrition,
Labour Welfare and Labour Productivity', Social Science History, 6.
Founts, N. (1984). 'Cleaning House: New Perspectives on Household
and Economic Development', mimeographed, New School for Social
Research.
—— BERGMANN, B., AQAEWAL, B., and FLORO, M. (1991) (eds.).
Women's Work in the World Economy (London: Macmillan).
FOLEY, D. (1967). 'Resource Allocation in the Public Sector*, Yale
Economic Essays, 7.
FOSTER, J. (1984). 'On Economic Poverty: A Survey of Aggregate
Measures', Advances in Econometrics, 3.
—— (1985). "Inequality Measurement', in H. P. Young (ed.)> Fair Alto~
cation (Providence, RI: American Mathematical Society).
—- GRBER, J., and THORBECKE, B. (1984). 'A Class of Decomposable
Poverty Measurement', Econometrics, 42.
MAJUMDAR, M, K., and MITRA,T. (1988). 'Inequalityand Welfare in Ex-
changeEconomies', CAEDiscussion Paper No.88-09,CornellUniversity.
— and SHORROCKS, A, F. (1988a). 'Poverty Orderings*, Econametrica, 56.
_ (I988fe), 'Poverty Orderings and Welfare Dominance', Social
Choice and Welfare, 5.
(1991). 'Subgroup Consistent PovertyIndices', Econometric®, 59,
FOUNTAIN, J. (1980). 'Bowley's Analysis of Bilateral Monopoly and Sen's
Liberal Paradox in Collective Choice Theory: A Note', Quarterly Jour-
nal of Economics, 95.

References 169
FOXLEY, A., and RACZYNSKI, D. (1984), 'Vulnerable Groups in Reces-
sionary Situations:The Case of Children and the Youngin Chile', World
Development, 12.
FRANK, R, H, (1985), Choosing the Right Pond: Human Behavior and the
Questfor Status (New York; Oxford University Press).
FRANKFUET, H, (1987). 'Equality as a Moral Ideal*, Ethics,98.
FRIED, C. (1978). Right and Wrong (Cambridge, Mass.: Harvard Univer-
sity Press).
FUCHS, V. R, (1983), How We Live: An Economic Perspective on Americans
from Birth to Death (Cambridge, Mass.: Harvard University Press).
GAERTNEK, W. (1985). 'Justice Constrained Libertarian Claims and
Pareto Efficient Collective Decisions', Erkemtnis, 23.
— (1986). 'Pareto, Independent Rights Exercising and Strategic
Behavior*. Journal of Economics, 46.
— (1988). 'Review of Commodities and Capabilities*, Journal of Econ-
omics, 48.
— (1992), 'Comment on Cohen and Sen*, in Nussbaum and Sen(1992).
-and KR&GER, L. (1981). 'Self-Supporting Preferences and Individual
Rights: The Possibility of Paretian Liberalism', Economka, 48,
•————(1983), 'Alternative Libertarian Claims and Sen's Paradox',
Theory and Decision, 15.
PATTANAIK, P., and SUZUMUEA, K, (1992). 'Individual Rights Revis-
ited", forthcoming in Economiea,
GAJHA, R. (1985). 'Poverty, Technology and Infrastructure in Rural
India*, Cambridge Journal of Economics, 9.
and KAZMI, N. A, (1981). 'Aspects of Poverty in Rural India',
Economics of Planning, 17.
GALBRAITH, J. K. (1952). American Capitalism: The Concept of Counter-
vailing Power (Cambridge, Mass.: Riverside Press).
(1958). The Affluent Society (Boston: Houghton Miffiin).
(1967). The New Industrial State (Boston: Houghton Mifflin).
GAMBETTA, D. (1987) (ed.). Trust and Agency (Oxford: Blackwell).
GARDENFORS, P. (1981). 'Rights, Games and Social Choice', Nous, 15.
—— and PETTIT, P. (1989). 'The Impossibility of a Paretian Loyalist*,
Theory and Decision, 27.
GAE0INBR, J,, HiMMBLWBrr, S., and MACKINTOSH, M. (1975). 'Women's
Domestic Labour', Bulletin of the Conference of Socialist Economists,
GARDNER, R. (1980), "The Strategic Inconsistency of Paretian Liberalism',
Public Choice, 35.
GASTWIRTH, I. L. (1975). 'The Estimation of a Family of Measures of
Economic Inequality', Journal of Econometrics, 3.
GAUTHIER, D. (1986). Morals by Agreement (Oxford: Clarendon Press).
GEVERS, L. (1979). 'On Interpersonal Comparability and Social Welfare
Orderings', Econometrica, 47.

170 References
GIBBARD, A, (1974), *A Pareto-Consistent Libertarian Claim', Journal of
Economic Theory, 7,
(1979), "Disparate Goods and Rawls's Difference Principle: A Social
Choice Theoretic Treatment', Theory and Decision, 11.
— (1986), 'Interpersonal Comparisons: Preference,Good, andtheIntrin-
sic Reward of a Life', in Elster and Hylland (1986),
GJOUOTTI, G. A, (1988). "The Conflict between Naive and Sophisticated
Choice as a Form of the "Liberal Paradox"', Theory and Decision, 24.
GLANTZ, M. H. (1976). The Political Economy of Natural Disaster (New
York: Praeger).
GLOVER, J. (1977). Causing Death and SavingLives (Harmondsworth: Pen-
guin Books),
GOLDIN, C. (1989). Understanding the Gender Gap: An Economic History of
American Women (New York: Oxford University Press),
GQUJSCHMIOT-CLERMONT, L, (1982), Unpaid Work in the Household
(Geneva: ILO),
GOODIN, R, E, (1982), Political Theory and Public Policy (Chicago: Univer-
sity of Chicago Press),
—— (1985). Protecting the Vulnerable (Chicago: University of Chicago
Press).
(1986). 'Laundering Preferences', in Elster and Hylland (1986).
— (1987). 'EgaHtarianism, Fetishistic and Otherwise*, Ethics,98.
— (1988). Reasonsfor Welfare: The Political Theory of the Welfare State
(Princeton, NJ: Princeton University Press).
GOPALAN, C. (1983). 'Measurement of Undernutrition: Biological Con-
siderations', Economic and Political Weekly, 19 (9 Apr.).
GORMAN, W. M. (1956), 'The Demand for Related Goods', Journal Paper
J 3129, Iowa Experimental Station, Ames, la.
—-(1976). 'Tricks with Utility Functions', in M. J. Artis and A. R.
Nobay (eds.), Es$ays in Economic Analysis (Cambridge: Cambridge
University Press).
GOSUNO, J. C. B. (1969). Pleasure and Desire (Oxford: Clarendon Press).
GOTTINGER, H. W., and LEINFELLNER, W, (1978) (eds,). Decision Theory and
Social Ethics (Dordrecht: Reidel).
GOUOUEN, J. A. (1967), 'L-Fuzzy Sets', Journal of Mathematical Analysis
and Applications,
GRAAFF, 1, m v. (1946). 'Fluctuations in Income Concentration', South
African Journal of Economics, 14.
——(1957), Theoretical Welfare Economics (Cambridge: Cambridge
University Press).
——(1977), 'Equity and Efficiency as Components of General Welfare',
South African Journal of Economics, 45,
(1979), Tastes, Valuesand the Foundations ofNormative Economies',
paper read at the 6th Interlaken Seminar on Analysis and Ideology.

References 171
—•—(1985). 'Normative Measurement Theory', unpub. manuscript,
GBANT, J, P, (1978), Disparity ReductionMates in Social Indicators (Wash-
ington, DC; Overseas Development Council).
GRAY, I (1983), John Stuart Mill's Doctrine of Liberty: A Defence
(London: Routledge and Regan Paul),
GREEN, E. T. (1980). 'Libertarian Aggregation of Preferences: What the
"Cease Theorem" Might Have Said', Social Science Working Paper No,
315, California Institute of Technology.
GRIFFIN, J, (1981). 'Equality: On Sen's Weak Equity Axiom*, Mind, 50.
-—-(1982), 'Modern Utilitarianism*, Revue Internationale de philosophic,
36.
(1986), Well-Being (Oxford: Clarendon Press).
GWFRN, K. (1978). International Inequality and National Poverty (London:
Macmillan).
——and KNIGHT, J. (1989) (eds.). Human Development in the 1980s and
Beyond, Journal of Development Planning, 19 (Special Number),
GUTMANN, A. (1980). Liberal Equality (Cambridge: Cambridge University
Press),
——-(1983). 'For and Against Equal Access to Health Care', in Bayer,
Caplan, and Daniels (1983),
GUTTAMACHER, A, F. (1957). Pregnancy and Birth (NewYork: VikingPress).
GWATKIN, D, R., WIUXK, J, R., and WRAY, J. P. (1980). 'The Policy
Implications of Field Experience in Primary Health and Nutrition*,
Social Science and Medicine, 14C.
HAGENAAKS, A. J. M,, and 0E Vos, K. (1989). 'A Comparison between the
Poverty Concepts of Sen and Townsend', mimeographed,
— and VANPEAAG, B, M. S. (1983), 'A Synthesis of Poverty Line Defini-
tions', Report P3.01, Center for Research in Public Economics, Leiden
University.
HAHN, F, H. (1971). 'Equilibrium with Transaction Costs', Econometrics,
39,
——-and HOLLIS, M. (1979) (eds.). Philosophy and Economic Theory
(Oxford: Clarendon Press).
HAKSAR, V. (1979). Equality, Liberty and Perfectionism (Oxford: Clarendon
Press),
HAMADA, K., and TAKAYAMA, N. (1978). 'Censored Income Distributions
and the Measurement of Poverty', Bulletin of International Statistical
Institute, 47.
HAMLIN, A, (1986), Ethics, Economics and the State (Brighton:
Wheatsheaf Books),
-——(1989). 'Rights, Indirect Utilitarianism, and Contractarianism',
Economics and Philosophy, S.
— and PETTIT, P. (1989) (eds.). The Good Polity: Normative Analysis of
the State (Oxford: Blackwell).

172 References
HAMMOND, P. J. (1976a), 'Equity, Arrow's Conditions and Rawls"
Difference Principle', Econometrica, 44.
-—--(19766). 'Why Ethical Measures of Inequality Need Interpersonal
Comparisons', Theory and Decision, 1.
—— (1977). 'Dual Interpersonal Comparisons of Utility and the Welfare
Economics of Income Distribution', Journal of Public Economics, 6,
—— (1978). 'Economic Welfarewith Rank Order Price Weighting*, Review
of Economic Studies, 45.
(1981). 'Liberalism, Independent Rights and the Pareto Principle', in
J. Cohen (ed.), Proceedings of the 6th International Congress of Logic,
Methodology and Philosophy of Science (Dordrecht: Reidel).
— (1982). 'Utilitarianism, Uncertaintyand Information', in Sen andWil-
liams (1982).
• (1985). 'Welfare Economies', in G, Feiwel (ed.), Issues in Contem-
porary Microeconomics and Welfare (Albany, NY: SUNY Press).
• (1986). 'Consequentialirt Social Norms for Public Decisions', in
Heller, Starr, and Stairett (1986).
HAMMOND, R. J. (1951). History of the Second World War:Food (London;
HMSO).
HAMPSHIRE, S, (1982). 'Morality and Convention', in Sen and Williams
(1982),
HANSSON, B. (1977). The Measurement of Social Inequality', in R. Butts
and J. Hintikka (eds,), Fawdational Problems in the Special Sciences
(Dordrecht: Reidel).
HANSSON, S. O. (1988). 'Rights and the Liberal Paradoxes', Social Choke
and Welfare, 5.
HARB, R. M. (1952), The Language of Morals (Oxford: Clarendon Press).
(1963). Freedom and Reason (Oxford: Clarendon Press).
— (1981), Moral Thinking: Its Levels, Methods andPoint {Oxford: Claren-
don Press).
-—(1982). 'Ethical Theory and Utilitarianism', in Sen and Williams
(1982).
HAREL, A., and NITZAN, S. (1987). The Libertarian Resolution of the Pare-
tian Liberal Paradox', Journal of Economics, 47.
HAXMAN, G. (1977). The Nature of Morality (New York: OxfordUniversity
Press).
HAERISON, G. A. (1988). Famines (Oxford: Clarendon Press).
HARWSS, B. (1977). 'Paddy Milling: Problems in Policy and Choice of Tech-
nology', in B. F. Farmer (ed.), The Green Revolution (London: Billing).
— (1990). 'The Intrafamily Distribution of Hunger in South Asia', in
Dreze and Sen (1990), vol. 1.
and WATSON, E. (1987). "The Sex-Ratio in South Asia', in J. H. Mom-
sen and J. Townsend (eds.), Geography of Gender in the Third World
(London: HutcMnson).

References 173
HAESANYI, J. C. (1955), 'Cardinal Welfare, Individualistic Ethics and Inter-
personal Comparisons of Utility", Journal of Political Economy, 63,
—— (1982). 'Morality and the Theory of Rational Behaviour', in Sen and
Williams (1982),
HART, H. L, A. (1961). The Concept of Law (Oxford; Clarendon Press).
— (1973), 'Rawls on Liberty and its Priority', University of Chicago Law
Review, 40. (Repr. in Daniels 1975.)
HART, K. (1987). 'CornmoditiEation and the Standard of Living", in Sen et al,
(1987).
HAUSMAN, D. M., and MCPHEKSGN, M. S. (1991). 'Taking Morality
Seriously; Economics and Contemporary Moral Philosophy', mimeo-
graphed, University of Wisconsin and WilliamsCollege.
HAWTHOEN, G. (1987). 'Introduction*, in Sen et al. (1987).
HAYEK, p. A. (1960). The Constitution of Liberty (London: Routtedge and
Kegaa Paul),
—— (1967), Studies inPhilosophy, Politics,andEconomics(Chicago;Univer-
sity of Chicago Press).
HELLER, W, P., STARS, R. M., and STAMUBTT, D. A. (1986) (eds.). Social
Choice and Public Decision Making: Essays in Honor of Kenneth J. Arrow
HELM, D. (1986). The Assessment: The Economic Borders of the State',
Oxford Review of Economic Policy, 2.
HELPMAN, E,, and SADKA, E. (1978). 'Optimal Taxation of Full Income*,
International Economic Review, 19.
HEMMELL, V., and SINDBJERO, P. (1984). Women in Mural China: Policy
towards Women before and after the Cultural Mevolution (Copenhagen:
Scandinavian Institute of Social Studies and Humanities Press).
HEMMING, R. (1984). Poverty and Incentives: The Economics of Social
Security (Oxford: Clarendon Press).
HICKS, J, R. (1939). Value and Capital (Oxford: Clarendon Press).
(1940). 'The Valuation of the Social Income*, Economica, 7,
HILPINEN, R. (1971) (ed.). Deontic Logic: Introductory and Systematic Read-
ings (Dordrecht: Reidel),
HIMMELFARB, O, (1984). The Idea of Poverty (London: Faber & Faber).
HIRSCH, F. (1977), Social Limits to Growth (London: Routledge).
HIBSCHMAN, A. O. (1958). The Strategy of Economic Development (New
Haven, Conn.: Yale University Press).
— (1982). Shifting Involvement (Princeton, MI; Princeton University
Press),
HOBSBAWM, E. (1964). Labouring Men (London: Weidenfeld and Nieol-
son).
-— (1989), Politicsfor a Rational Left (London; Verso).
HONDEWCH, T. (1985) (ed.). Morality and Objectivity: A Tribute to J, L,
Mackie (London: Routledge aad Kegan Paul).

174 References
HOSSAIN, I, (1990). Poverty as Capability Failure (Helsinki:Swedish School
of Economics),
HOUTHAKKER, H, S. (1950). 'Revealed Preferenceand the Utility Function',
Economical, 17.
HULL, C. H, (1899) (ed.). The Economic Writings of Sir William Petty
HURLEY, S. (1985), 'Objectivity and Disagreement*, in Honderich
(1985).
— (1989). Natural Reasons (Oxford: Clarendon Press),
HURWICZ, L., SCHMEIDLER, D,, and SONNENSCHEIN, H, (1985) (ed$.).
Social Goals and Social Organisation: Essays in Memory of Elisha
Pazner (Cambridge: Cambridge University Press),
HUSSAIN, A., Liu, H,, and Liu, X. (1989). 'Compendium of Literature
on the Chinese Social Security System*, mimeographed, London
School of Economics,
HUTCHENS, R, M, (1986). 'Segregation Curves, Lorenz Curves and
Inequality in the Distribution of People across Groups', New York
School of Industrial and Labor Relations, CornellUniversity.
HYULAND, A. (1986). "The Purpose and Significance of Social Choice
Theory: Some General Remarks and Application to the "Lady
Chatterley Problem"', in Elster and Hylland (1986).
ISENMAN, P. (1980). 'Basic Needs: The Case of Sri Lanka', World
Development, 1.
IVENGAR, N, S. (1989). 'Recent Studies of Poverty in India', Journal of
Quantitative Economics, 5,
JAIN, D. (1985). "The Household Trap: Report on a Field Survey of
Female Activity Patterns', in Jain and Banerjee (1985).
-—-and BANERJEE, N. (1985). Tyranny of the Household: Investigative
Essays on Women's Work (New Delhi: Vikas).
IASSO, G. (1981). 'Who Gains and Who Loses under Alternative In-
come Distribution Regimes that have Identical Magnitudes of the
Gini Coefficient', Proceedings of the Social Statistics Section of the
American Statistical Association, 1981.
JAYAWARDENA, K. (1986). Feminism, and Nationalism in the Third World
(London: Zed Books),
JAYAWARBENA, L. (1974), 'Sri Lanka', in Chenery et al (1974).
JENCKS, C. (1972), Inequality (New York: Basic Books).
JOHANSSON, S. R. (1991), 'Mortality, Welfare and Gender: Continuity
and Change in Explanations for Male/Female Mortality Differences
over Three Centuries', Continuity and Change,
JORGENSON, D. W., and SLBSNJCK, D. T. (1983). 'Individual and Social
Cost of Living Indexes', in Diewert and Montmarquette (1983).
_ __(j9g4a), 'inequality in the Distribution of Individual Wel-
fare', Advances in Econometrics,3.

References 175
(19846). "Aggregate Consumer Behaviour and the Measure-
ment of Inequality', Review of Economic Studies, 51,
_ (1986). 'Redistribution Policy and the Elimination of Poverty*,
discussion paper. Harvard Institute of Economic Research.
______ (1987). 'Aggregate Consumer Behaviour and Household
Equivalence Scales', Journal of Business and EconomicStatistics, 5,
LAV, L. J,, and STOKER, T. M. (1980). 'Welfare Comparison under
Exact Aggregation', American EconomicReview, 70.
KAKWANI, N. (19800). Income, Inequality and Poverty (New York: Oxford
University Press).
(19806). 'On a Class of Poverty Measures', Econometrica, 48. .
—— (1981). 'Welfare Measures: An International Comparison', Journal of
Development Economics,8,
-—-(1986). Analysing Redistribution Policies (Cambridge; Cambridge
University Press).
-—-(1988), 'Income Inequality, Welfare and Poverty in a Developing
Economy with Applications to Sri Lanka', Social Choice and Welfare,
5.
(1992). 'Measuring Undernutrition with Variable Calorie Require-
ments', in Osmani (19920).
and SUBBARAO, K. (1990). 'Rural Poverty in India, 1973-86', Working
Paper WPS 526, The World Bank.
KALAI, E., and SMOMMNSKY, M. (1975). 'Other Solutions to Nash's
Bargaining Problem', Econometrica, 43.
KANBUR, R, [S. M. R.J (1979). 'Of Risk Taking and Personal Distribution of
Income', Journal of Political Economy, 87.
— (19820). 'Entrepreneurial Risk Taking, Inequality and Public Policy',
Journal of PoliticalEconomy, 87.
(1982ft). 'The Measurement and Decomposition of Inequality and
Poverty: A Selective Survey', in F. van der Ploeg (ed.)» Handbook of
Applicable Mathematics: Economics(Chichester: John Wiley and Sons).
(1987). 'The Standard of Living: Uncertainty, Inequalityand Oppor-
tunity', in Sen el al. (1987).
—— and HADDAD, L. (1990), 'How Serious is the Neglect of Intrahousehold
Inequality', Economic Journal, 100.
and STROMBERa, J. O. (1986). 'Income Transitions and Income Dis-
tribution Dominance", Journal of Economic Theory, 96,
KANEKO, M., and NAKAMURA, M. (1979). The Nash Social Welfare Func-
tion", Econometrica,47.
KANGER, S. (1957). New Foundations far Ethical Theory (Stockholm). (Repr.
in Hilpinen 1971.)
(1972). 'Law and Logic', Theoria, 32.
(1985). 'On Realization of Human Rights', Acta Philosophka Fennica,
38.

176 References
KANT, I, (1785), Fundamental Principles of Metaphysics of Ethics,English
trans, by T, K, Abbott (London: Longmans, 1907).
-—(1788). Critique of Practical Reason, English trans, by L. W. Beck
(New York: Liberal Arts Press, 1956).
KAPTEYN, A., and VANPRAAQ, B. M. S. (1916). 'A New Approach to the Con-
struction of Family Equivalent Scales', European Economic Review, 7.
KARNI, E, (1978). 'Collective Rationality, Unanimity and Liberal Ethics*,
KELLY, J. S. (1976«). 'The Impossibility of a Just Liberal', Economica, 43.
(1976£). 'Rights-Exercising and a Pareto-Consistent Libertarian
Claim", Journal of Economic Theory, 13,
—— (1978). Arrow Impassibility Theorems (New York: Academic Press).
KELSEV, D. (1985). 'The Liberal Paradox: A Generalization*, Social Choice
and Welfare, 1.
KBLSEY, D. (1988), 'What is Responsible for the "Paretian Epidemic"*,
Social Choice and Welfare, 5,
KERN, L. (1978). 'Comparative Distribution Ethics: An Extension of Sen's
Examination of the Pure Distribution Problem', in Gottinger and Lein-
fellner (1978).
KHAN, Q. M. (1985). 'A Model of Endowment Constrained Demand for
Food in an Agricultural Economy with Empirical Application to
Bangladesh*, World Development, 13.
KOLAKOWSKI, L. (1978). Main Currents of Marxism; Its Origin, Growth and
Dissolution (Oxford: Clarendon Press).
KOLM, S. CH. (1969). 'The Optimal Production of Social Justice', in Mar-
golis and Guitlon (1969).
——(1976). 'Unequal Inequalities', Journal of Economic Theory, 12.
._ (1977), 'Multidimensional Egalitarianisoi*, Quarterly Journal of Econ-
omies, 91.
KOOPMANS, T. C. (1964). 'On the Flexibility of Future Preferences', in M. W.
Shelley and Jf. L, Bryan (eds.), Human Judgments and Optimattty (New
York: Wiley).
KOKNAI, J. (1988). 'Individual Freedom and the Reform of Socialist
Economy', European Economic Review, 32.
KORSOAARD, C. (1992), 'Comments on Cohen and Sen', in Nussbaum and
Sen (1992).
KREIXE, W., and SHORKOCKS, A. F. (1978) (eds,). Personal Income Distribu-
tion (Amsterdam: North-Holland).
KREPS, D. (1979). 'A Representation Theorem for "Preference for
Flexibility"', Econometrica, 47.
KRISHNAJI, N. (1987). 'Poverty and the Sex Ratio: Some Data and Specula-
tions', Economic and Political Weekly, 22.
KOMAH, B, G. (1987). 'Poverty and Public Policy; GovernmentIntervention
and Levels of Living in Kerala, India', D.Phil, thesis (Oxford).

References 177
—— (1989), 'Gender, Differential Mortality and Development;The Experi-
ence of Kerala1
, Cambridge Journal of Economies, 13.
KUMAR, A. K. S, (1992), 'Maternal Capabilities and Child Survival in Low
Income Regions: Economic Analysis of Infant Mortality Differentials in
India*, Ph.D. dissertation, Harvard University.
KUNDU, A,, and SMITH, T. E. (1983), 'An Impossibility Theorem on Poverty
Indices*, International Economic Review,24,
KUZNETS, S. (1961), Six Lectures on Economic Growth (New York: Free
Press of Gleneoe).
-—- (1966). Modern Economic Growth (New Haven, Conn,: Yale Univer-
sity Press).
(1973). Population, Capital and Growth: Selected Essays (London:
Heinemann).
KyNCH, J. (1985). 'How Many Women are Enough? Sex Ratios and the
Right to Life*, Third World Affairs 1985 (London: Third World Found-
ation).
— and SEN,A. K. (1983). 'Indian Women: Well-Being and Survival',
Cambridge Journal of Economics, 7.
LADEN, T. (I991a). 'Games, Fairness and Rawls's A Theory of Justice,
Philosophy and Public Affairs, 20.
——(19916). 'Freedom, Preference and Objectivity: Women and the
Capability Approach", mimeographed, Department of Philosophy, Har-
vard University,
LAMBEBT, P. J. (1985). 'Social Welfare and the Gini Coefficient Revisited',
Mathematical Social Sciences,9.
• (1989). The Distribution and Redistribution of Income (Oxford: Basil
Blackwell).
and WEALE, A. (1981). 'Equality, Risk-Aversion and Contractarian
Social Choice", Theory and Decision, 13,
LANCASTER, K. J. (1966). 'A New Approach to Consumer Theory', Journal
of Political Economy, 74,
(1971). Consumer Demand; A New Approach (New York: Columbia
University Press),
LABMQHE, C. (1987). Patternsof Moral Complexity (Cambridge: Cambridge
University Press).
LASLETF, P. (1991). A Fresh Map of Life: The Emergence of the Third Age
(Cambridge, Mass.: Harvard University Press).
LE BRETON, K., and TRANNOV, A. (1987). 'Measures of Inequality as an
Aggregation of Individual Preferences about Income Distribution: The
Arrowian Case', Journal of Economic Theory, 41.
and UWARTB, J. R. (1985). Topologies! Aggregation of la-
equality Preorders', Social Choice and Welfare, 2.
LE GRAND, J. (1982). The Strategy of Equality: Redistribution and the Social
Services (London: Allen and Unwin).

178 References
—— (1984). 'Equity as an Economic Objective', Journal of Applied Philos-
ophy, 1.
(1990). 'Equity versus Efficiency: The Elusive Trade-off*, Ethics, 10,
(1991), Equity and Choice (London: Harper Collins),
LEMMING, P. B. (J989). 'Liberalism and Capabilities: Some Remarks on the
Neutrality Debate*, mimeographed, Rotterdam University.
LETWIN, W. (1983) (ed.). Against Equality (London: Macmillan).
LBVI, I. (1982), 'Liberty and Welfare', in Sen and Williams (1982).
——(1986). Hard Choices (Cambridge: Cambridge University Press),
LEWIS, O, W., and UJJPH, D. T. (1987). 'Poverty, Inequalityand Welfare',
Discussion Paper 87/188, University of Bristol.
LWDAHL, L. (1977). Position and Change: A Study of Law and Logic (Dor-
drecht: Reidel).
LINDBECK, A. (1988). 'Individual Freedom and Welfare State Policy', Euro-
pean Economic Review, 32,
LIFTON, M, (1983). 'Poverty, Undernutrition and Hunger', World Bank
Staff Working Paper (Washington, DC: The World Bank).
——(1985). 'A Problem in Poverty Measurement', Mathematical Social
Sciences, 10.
LITTLE, I. M. D. (1950). A Critique of Welfare Economics (Oxford: Claren-
don Press).
LOPEZ, A. D., and RUZICKA, L, T, (1983) (eds.). Sex Differences in Mortality
(Canberra: Department of Demography, Australian National
University).
LOURY, O. (1987). 'Why Should We Care about Group Inequality?' Social
Philosophy and Policy, 5.
LOUTFI, M, F. (1980). Rural Women: Unequal Partners in Development
(Geneva: ILO).
LUCAS, O. R. (1990). 'African Famine: New Economic and Ethical Perspec-
tives', Journal of Philosophy, 87.
LUCAS, J. R. (1965). 'Against Equality', Philosophy, 40.
(1980). On Justice (Oxford: Clarendon Press).
LUCE, R. D., and RAIFFA, H. (1957). Gamesand Decisions (New York:Wiley).
LUKEK, W. (1986). 'Welfare Economics, Positivist Idealism and Quasi-
Experimental Methodology", mimeographed, University of Texas at
Austin.
LUKES, S. (1985). Marxism and Morality (Oxford: Clarendon Press).
—— (1990). 'Equality and Liberty: Must They Conflict?', mimeographed,
European University Institute.
LYDALL, H, F. (1966). The Structureof Earnings (Oxford: Clarendon Press).
MAASOUMI, E. (1986). 'The Measurement and Decomposition of Multi-
dimensional Inequality", Econometrica, 54.
—- (1989). 'Continuously Distributed Attributes and Measures of Multi-
variate Inequality', Journal of Econometrics, 42,

References 179
McCoro, C, and FREEMAN, H. P. (1990), 'Excess Mortality in Harlem',
New England Journal of Medicine, 322 (18 Jan.),
MCDOWELL, J. (1981), 'Non-Cognitivistn and Rule Following*, in S. H,
HoltZBian and C, M. Leigh (eds.)> Wittgenstein; To Follow a Rule
(London: Routledge and Kegan Paul).
— (1985). 'Values and Secondary Qualities', in Honderich (1985),
MCELROY, M. B., and HORNBY, M. J, (1981). 'Nash-Bargained Household
Decisions: Toward a Generalization of the Theory of Demand', Inter-
national Economic Review, 22.
MAclNTYKB, I, D, A. (1987), '"The Liberal Paradox; A Generalisation" by
D. Kelsey', Social Choice and Welfare, 4,
-—-(1988), "Justice, Liberty, Unanimity and the Axioms of Identity',
Theory and Decision,24.
MACK, J., and LANSLEV, S, (1985), Poor Britain (London: Allen and Unwin).
MACKIE, J. L. (1978a). Ethics; Inventing Right and Wrong (Harmonds-
worth: Penguin),
—-(19786). 'Can There be a Rights-Based Moral Theory?' Midwest
Studies in Philosophy, 3,
— (1986), 'The Combination of Partially-Ordered Preferences', in J. L,
Mackie, Persons and Values (Oxford: Clarendon Press).
MCLEAN, I. (1980). 'Liberty, Equalityand the Pareto Principle', Analysis,
40,
MACLEOP, A, M, (1984). 'Distributive Justice, Contract and Equality',
Journal of Philosophy, 81.
MeMuRMN, S. M, (1980) (ed.). Tanner Lectures on Human Values, i (Salt
Lake City: Universityof Utah Press and Cambridge: Cambridge Univer-
sity Press).
MCPHBBSON, M. S, (1982), 'Mill's Moral Theory and the Problem of Prefer-
ence Change", Ethics, 92,
—— (1984), 'Limits of Self-Seeking; The Role of Morality in Economic
Life', in Colander (1984).
MAJUMDAR, T. (1969). 'Revealed Preference and the Demand Theorem in a
Not Necessarily Competitive Market*, Quarterly Journal of Economics,
S3.
MANSER, M., and BROWN, M. (1980), 'Marriage and Household Decision
Making; A Bargaining Analysis', International Economic Review, 21.
MARCUS, R, B. (1980). 'Moral Dilemmas and Consistency', Journal of
Philosophy, 1977.
MAROUN, S. A. (1984). Growth, Distribution and Prices (Cambridge, Mass.:
Harvard University Press).
MARGOUS, }., and GUITTON, H. (1969) (eds.). Public Economics (London:
Macmillan).
MARX, K. (1844), The Economic and Philosophic Manuscript of 1844, English
trans. (London: Lawrence and Wishart),

180 References
MARX, K. (with F. ENOELS) (1845-46), The German Ideology, English trans,
(New York: International Publishers, 1947),
—— (1857-58). Gnrndrisse; Foundations of the Critique of Political
Economy, English trans, by M. Nicolaus (Hanmondsworth: Penguin
Books, 1973).
(1867). Capital, vol. i, English trans. (London: Allen and Unwin,
1938).
——(1875). Critique of the Gotha Program, English trans. (New York:
International Publishers, 1938).
MASKIN, E. (1978), *A Theorem on Utilitarianism*, Review of Economic
Studies, 45.
(1979). 'Decision-Making under Ignorance with Implications for
Social Choice', Theory and Decision, 11,
MAZUMDAR, V. (1985). Emergence of Women's Questions in India and the
Role of Women's Studies (New Delhi: Centre for Women's Development
Studies).
MEADE, J. E, (1955), Trade and Welfare (Oxford: Oxford University Press).
—— (1965). Efficiency, Equity and the Ownership of Property (Cambridge,
Mass.: Harvard University Press).
— (1976). The Just Economy (London: Allen and Unwin).
WEEKS, 0, (1991) (ed.). Thoughtful Economic Man (Cambridge: Cambridge
University Press).
MEHKAN, F. (1976). 'Linear Measures of Economic Equality', Econome-
trica, 44.
MEYER, M. (1987). 'Multidimensional Correlation and the Measurement
of Ex-Post Inequality under Uncertainty', unpub,manuscript, St John's
College, Oxford,
MEZZETTI, C. (1987). 'Paretian Efficiency, Rawlsian Justice and the
Nozick Theory of Right*, Social Choice and Welfare, 4.
MIES, M. (1982). The Lace Makers ofNasrapur: Indian Housewives Produce
for the World Market (London: Zed Press).
MILIBANB, R. (1977), Marxism and Politico (Oxford: Clarendon Press).
MILL, J. S. (1859). On Liberty (London). (Republished Harmondsworth:
Penguin, 1974.)
-(1861). Utilitarianism (London).
MILLER, B. (1981). The Endangered Sex: Neglect of Female Children in
Rural North India (Ithaca, NY: Cornell University Press).
— (1982). 'Female Labor Participation and Female Seclusion in Rural
India: A Regional View', Economic Development and Cultural Change,30.
——(1984). 'Child Survival and Sex Differential in the Treatment of
Children', Medical Anthropology, 8,
MINCER, J. (1962). 'Labor Force Participation of Married Women', in
H. G. Lewis (ed.), Aspects of Labour Economics (Princeton,NJ: Princeton
University Press). (Repr. in Amsden 1980.)

References 181
— and POLOCHBK, S, (1974), 'Family Investments in Human Capital
Earnings of Women*, Journal of Political Economy, 82. (Repr, in
Amsden 1980.)
MIRRLEES, J. (1971). 'An Exploration in the Theory of Optimum Income
Taxation', Review of Economic Studies, 38,
—— (1974). 'Notes on Welfare Economics, Information and Uncertainty*,
in M, Baleh, D. McFadden, and S. Wu (eds.), Essays on Economic
Behaviour under Uncertainty (Amsterdam: North-Holland).
—— (1982). The Economic Uses of Utilitarianism', in Sen and Williams
(1982).
(1986). 'The Theory of Optimal Taxation*, in Arrow and Intriligator
(1986).
MJTRA, A. (1980). Implications of Declining Sex Ratio in India's Population
(Bombay: Allied Publishers).
MOOKHEBJEE, D., and SHORROCKS, A. (1982). 'A Decomposition Analysis of
the Trends in UK Income Inequality*, Economic Journal, 92.
MORRIS, M. D. (1979). Measuring the Conditions of the World's Poor: The
Physical Quality of Life Index (Oxford: Pergamon Press).
MOULIN, H. (1989). 'Welfare Bounds and Fair Allocation of Private
Goods', mimeographed, Virginia Polytechnic Institute.
— (1990). 'Interpreting Common Ownership", Recherches icanomltpiesde
Louvain, 56.
MUELLBAUER, J. (1974a). 'Household Composition, Bngel Curves and
Welfare Comparisons between Households: A Duality Approach", Euro-
pean Economic Review,5.
-—(1974ft). 'Inequality Measures, Prices and Household Composition",
Review of Economic Studies,41.
— (1978). 'Distributional Aspects ofPrice Comparisons', inR. Stoneand
W. Peterson (eds.), Economic Contributions to Public Policy (London:
Macmillan).
—— (1987), 'Professor Sen on the Standard of Living', in Sen et at. (1987).
MueitBR, D. C. (1979), Public Choice (Cambridge: Cambridge University
Press).
MURRAY, C. and CHEN, L, (1990). 'Health Transitions: Patterns and Dyna-
mics',mimeographed, Center for Population Studies, Harvard University,
MUSGRAVE, R. (1959). The Theory of Public Finance (New York: McGraw-
Hill).
NAGEL, T. (1979). Mortal Questions (Cambridge: Cambridge University
Press).
— (1980). The Limits of Objectivity", in McMurrin (1980).
— (1986). The View from Nowhere (New York: Oxford University Press).
NASH, J. F. (1950). The Bargaining Problem', Econometrtca, 18.
NBWBBKY, D. M. G. (1970). 'A Theorem on the Measurementof Inequality*,
Journal of Economic Theory,, 1.

182 References
No, Y.-K, (1971), 'The Possibility of a Paretian Liberal: Impossibility
Theorems and Cardinal Utility', Journal of Political Economy,79,
(1979), Welfare Economics(London: Macmillan),
NISSEN, H. P. (1984) (ed.). Towards Income DistributionPolicies(Tilburg:
European Association of Development Research and TrainingInstitute),
NORDHAUS, W., and TOWN, J. (1972). 'Is Growth Obsolete?', in National
Bureau of Economic Research, Economic Growth: Fiftieth Anniversary
Colloquium (New York: NBER),
NOZICK, R, (1973). 'Distributive Justice', Philosophy and Public Affairs, 3,
— (1974), Anarchy, Stale and Utopia (Oxford: Blackwell).
(1989), The Examined Life (New York: Simon & Schuster).
NURMI, H, (1984). 'On Taking Preferences Seriously*, in Anckar and
Berndtson (1984).
— (1987), Comparing Voting Systems (Dordrecht: Reidel).
NUSSBAUM, M, C. (1985), Fragility of Goodness: Luck and Ethics in Greek
Tragedy and Philosophy (Cambridge; Cambridge University Press),
(I988a). 'Nature, Function, and Capability: Aristotleon Political Dis-
tribution*, Oxford Studies in Ancient Philosophy (supplementary volume).
(19886). 'Non-Relative Virtues: An Aristotelian Approach*, Midwest
Studies in Philosophy, 13; revised version in Nussbaum and Sen (1991).
(1991«). 'Human Functioning and Social Justice: In Defence of
Aristotelian Essentialism*, paper presented at the WIDER conference
on 'Human Capabilities: Women, Men and Equality", 14—16 Aug.
— (19916). 'Emotions and Women's Capabilities', paper presented at
the WIDER conference on 'Human Capabilities: Women, Men and
Equality', 14-16 Aug.
and SEN, A, K. (1992) (eds.). The Quality of Life (Oxford: Clarendon
Press).
NYGRAD, F., and SANDSTROM, A. (1981), Measuring Income Inequality
(Stockholm: Almqvist and Wiksell International),
OKIN, S, M. (1987), 'Justice and Gender', Philosophy and Public Affairs, 16
(Winter).
(1989), Justice, Gender and Family (New York: Basic Books).
OKUN, A. (1975). Equality and'Efficiency: The Big Tradeoff (Washington:
Brookings).
O'NEILL, O. (1986), Faces of Hunger (London: Allen and Unwin),
(1989). Constructions of Reason; Explorations of Kant '$ Practical Phil-
osophy (Cambridge: Cambridge University Press).
(1992), "Justice, Gender and International Boundaries', in Nussbaum
and Sen (1992).
OSMANI, S, R. (1978). 'On the Normative Measurement of Inequality*,
Bangladesh Development Studies, 6.
— (1982), Economic Inequality and Group Welfare (Oxford: Clarendon
Press),

References 183
(1990a). 'Nutrition and the Economics of Food: Implications of
Some Recent Controversies', in Dreze and Sen (1990), vol. 1.
— (19906). 'Freedom and the Choice of Space', mimeographed,
WIDER, Helsinki.
— (I992<z) (ed.). Nutrition and Poverty (Oxford: Clarendon Press),
forthcoming,
(19926). 'Undernutrition: Measurement and Implications', in
Osmani (1992a).
—— (1992c). 'Comments', in Nussbaum and Sen (1992).
OTTEN, M. W., TEUTSCH, S. M., WILLIAMSON, D. P., and MARKS, J, S.
(1990). 'The Effect of Known Risk Factors in the Excess Mortality
of Black Adults in the United States', Journal of the American
Medical Association, 263 (9 Feb.).
PADMANABHA, P. (1982). 'Trends in Morality*, Economic and Political
Weekly, 17 (7 Aug.).
PALMER, J., SMEEDING, T., and TOEREY, B. (1988). The Vulnerable:
America's Young and Old in the Industrial World (Washington, DC:
Urban Institute Press),
PANIKAR, P. O, K., and SOMAN, C. R. (1984). Health Status of Kerala
(Trivandrym: Centre for Development Studies).
PANT, P., et al. (1962). Perspective of Development 1961-1976, Impli-
cations of Planning for a Minimum Level of Living (New Delhi: Plan-
ning Commission of India).
PANTULU, Y. V. (1980). 'On Sen's Measure of Poverty', mimeographed,
Sardar Patel Institute of Economic and Social Research.
PAPANEK, H. (1990). 'To each less than she needs, from each more
than she can do: Allocations, Entitlements and Value', in Tinker
(1990a).
PARFJT, D. (1984). Reasonsand Persons (Oxford: Clarendon Press).
PATTANAIK, P. K. (1988). 'On the Consistency of Libertarian Values',
Economica, 55.
and SENGUFfA, M, (1991). 'A Note on Sen's Normalization
Axiom for a Poverty Measure*, mimeographed.
and Xu, Y. (1990). 'On Ranking Opportunity Sets in Terms of
Freedom of Choice', Recherches economiques de Louvain, 56.
PAYNE, P. R. (1985). 'Nutritional Adoption in Man: Social Adjustments
and their Nutritional Implications' in Blaxter and Waterlow (1985).
-— and LIPTON, M., with LONGHURST, R., NORTH, J., and TREAGUST, S.
(1988). 'How Third World Rural Households Adapt to Dietary Energy
Stress', mimeographed, International Food Policy Research Institute,
Washington, DC,
PAZNER, E. A., and SCHMEIDLER, D. (1974). 'A Difficulty in the Concept of
Fairness*, Review of Economic Studies, 41.
PEACOCK, A. T., and ROWLEY, C. K. (1972), 'Welfare Economics and the

184 References
Public Regulation of Natural Monopoly', Journal of PoliticalEconomy,
W,
PEN, J. (1971). Income Distribution: Facts, Theories, Policies (New York:
Praeger).
PERBLU-MINETTI, C. R. (1977). 'Nozick on Sen: A Misunderstanding',
Theory and Decision, 8,
PETTY, W. (1676). Political Arithmetick. (Republisfaed in Hull 1899.)
PHBUPS, E. &.(1973) (ed.). Economic Justice (Harrnondsworth: Penguin).
—- (1977). 'Recent Developments in Welfare Economics; Justice et
equite", in M. D. Intriligator, Frontiers of Quantitative Economics, Hi
(Amsterdam: North-Holland). (Repr. in Phelps 1980.)
(1980). Studies in Macroeconomic Theory, ii. Redistribution and Growth
(New York: Academic Press).
Ptoou, A. C, (1952). The Economics of Welfare, 4th edn. with eight new
appendices (London: Macmillan).
PtGTT, C. (1978). 'Rawls' Theory of Justice: An Impossibility Result*, in
Oottinger and Leinfellner (1978).
POCOB, T. W. (1989). Realizing Rawls (Ithaca, NY: Cornell University
Press).
PQLLAK, R. A., and WALES, T. J. (1979). 'Welfare Comparisons and Equiva-
lent Scales', American Economic Review,69.
—_____ (1981), 'Demographic Variables in Demand Analysis", Econome-
trica, 49.
PRESTON, S., KEYWTZ, N., and SCHQEN, R. (1972). Causes of Death; Life
Tables of National Populations (New York: Seminar Press).
PUTNAM, H. (1987). The Many Faces of Realism (La Salle, III.: Open Court).
—— (1992), 'Objectivity and the Science/Ethics Distinction*, in Nussbaum
and Sen (1992).
PUTNAM, R. A. (1991). 'Why Not a Feminist Theory of Justice?' paper
presented at the WIDER conference on 'Human Capabilities: Women,
Men and Equality", 14-16 Aug.
POTTERMAN, L. (1986). Peasants, Collectives,and Choice (Greenwich, Conn.:
JAI Press).
PYATT, O, (1976). 'On the Interpretation and Disaggregation of Gini Coeffi-
cients', Economic Journal, 86.
(1987). 'Measuring Welfare,Poverty and Inequality*, Economic Jour-
nal, 97.
RAB, D. (1981). Equalities (Cambridge, Mass.: Harvard UniversityPress).
RAM, N. (1990). 'An Independent Press and Anti-Hunger Strategies: The
Indian Experience*, in Dreze and Sen (1990), vol. 1.
RAMACHANDRAN, V. K. (1990). Wage Labour and Unfreedom in Agriculture
RAMSEY, F. P. (1928). 'A Mathematical Theory of Saving', Economic Jour-
nal, 38.

References 185
(1978). Foundations: Essays in Philosophy, Logic, Mathematics and
Economics (London: Routtedge).
RAVALIJON, M. (1987), Markets and Famines (Oxford; Clarendon Press),
and VANDEWALLE, D. (1988). 'Poverty Orderings of Food,Pricing
Reforms', Discussion Paper 86, Development Economics Research
Centre, University of Warwick.
RAWLS, J. (1958), "Justice as Fairness", Philosophical Review, 67,
(1971). A Theory of Justice (Cambridge, Mass.; Harvard University
Press).
(1982). 'Social Unity and Primary Goods', in Sen and Williams
(1982).
(1985). "Justice as Fairness; Political not Metaphysical*, Philosophy
and Public Affairs, 14.
— (1.987). The Idea of an Overlapping Consensus*, Oxford Journal of
Legal Studies, 1.
(1988<a), 'Priority of Right and Ideas of the Good*, Philosophy and
Public Affairs, 17.
—— (19886). "Reply to Sen', mimeographed, Harvard University.
(1988e). 'The Domain of the Political and Overlapping Consensus',
mimeographed, Harvard University.
(1990). 'Political Liberalism', to be published by Columbia Univer-
sity Press.
• FWE0, C.t SEN, A., and SCHELUNG, T. (1987). Liberty, Equality, and
Law, ed. S. McMurrin (Cambridge; Cambridge University Press; and
Salt Lake City: University of Utah Press).
RAY, R. (1984a). 'A Class of Decomposable Poverty Measures: A Correc-
tion and a ModifiedPoverty Measure', unpuh. manuscript, University of
Manchester.
(19846).-'On Measuring Poverty in India: A Synthesis of Alternative
Measures', wnpub. manuscript, University of Manchester.
RAZ, J, (1986). The Morality of Freedom (Oxford: Clarendon Press).
RBDBY, S. (1988). "An Independent Press Working against Famine; The
Nigerian Experience", Journal of Modern African Studies, 26,
REOAN, D. H. (1983). 'Against Evaluator Relativity: A Response to Sen',
Philosophy and Public Affairs, 12.
RILEY, J, (1986). 'Generalized Social Welfare Functional: Welfarism,
Morality and Liberty*, Social Choice and Welfare, 3.
(1987). Liberal Utilitarianism: Social Choice Theory and J, S, Mill's
Philosophy (Cambridge; Cambridge University Press).
(1989a). 'Rights to Liberty in Purely Private Matters: Part F, Econ-
omics and Philosophy, 5,
(1989*). 'Rights to Liberty in Purely Private Matters: Part IF, Econ-
omics and Philosophy, 6.
RINGEN, S. (1984). Towards a Third Stage in the Measurement of Poverty',

186 References
unpub. manuscript, The Swedish Institute for Social Research, Univer-
sity of Stockholm.
(1987), The Possibility of Politics: A Study of the Economyof the Wel-
fare State (Oxford; Clarendon Press),
RKKIN, C, (1987). China's PoliticalEconomy (Oxford; Clarendon Press).
RISKJN, D, (1988). 'Reform: Where is China Going?' mimeographed,
Queens University, New York, and East Asia Center, Columbia
University,
ROBBINS, L. (1932). An Essay on the Nature and Significance of Economic
Science (London: Allen & Unwin).
—— (1938). 'Interpersonal Comparisons of Utility", Economic Journal, 48.
ROBERTS, K. W. S. (1980a). 'Interpersonal Comparability and Social Choice
Theory', Review of Economic Studies, 47.
(19806). 'Price Independent Welfare Prescriptions', Journal of Public
Economics, 13.
ROCHFORD, S. C. (1981). 'Nash-Bargained Household Decision Making in a
Peasant Economy', mimeographed.
—— (1982). 'General Results of Stable Pairwise-Bargained Allocations ia a
Marriage Market', mimeographed.
ROEMER, J. (1982). A General Theory of Exploitation and Class (Cambridge,
-—(1985). 'Equality of Talent', Economicsand Philosophy, 1.
—— (1986a). 'An Historical Materialist Alternative to Welfarism', in Elster
and Hylland (1986).
— (19866). 'Equality of Resources Implies Equality of Welfare", Quar-
terly Journal of Economics, 101.
—— (1988). Free to Lose: An Introduction to Marxist Economic Philosophy
(Cambridge, Mass.: Harvard University Press).
- (1990), 'Welfarism and Axiomatic BargainingTheory*, Rechercheseco~
nomiques de Louvain, 56,
• (1992). 'Distributing Health: The Allocation of Resources by an Inter-
national Agency*, in Nussbaum and Sen (1992).
ROGERS, B. (1980). The Domesticationof Women (London: Tavistock).
ROSENZWEIG, M. R., and SCHULTZ, 1, P. (1982). 'Market Opportunities,
Genetic Endowments, and Intra-Family Resource Distribution: Child
Survival in Rural India*, American Economic Review, 11.
Ross, D.(1980) (ed.). Aristotle: TheNicomachean Ethics (Oxford: Clarendon
Press).
ROTBERG, R. L, and RABB, T. K. (1985). Hunger and History (Cambridge:
Cambridge University Press).
ROTH, A. E. (1979). Axiomatic Models of Bargaining (Berlin: Springer-
Verlag).
ROTHSCHILD, M., and STIGUTZ, J. (1973). 'Some Further Results in the
Measurement of Inequality', Journal of Economic Theory, 6,

References 18?
ROWNTREE.B. S.(1901). Poverty: A Study of Town Life (London; Longmans),
— (1941). Poverty and Progress(London: Longmans).
RYAN, A. (1979)(ed,). The Idea of Freedom; Essays in Honour of Isaiah Berlin
SADKA, E. (1977). 'On Progressive Income Taxation*, American Economic
Review, 67,
SAMUELSON, P. A, (1938). 'A Note on the Pure Theory of Consumer
Behaviour', Economica, 5.
——(1947), Foundations of Economic Analysis (Cambridge, Mass.: Har-
vard University Press).
(1956). 'Social Indifference Curves', Quarterly Journal of Economies,!®,
SASTRY, S. A. R. (1977). 'Poverty, Inequality and Development: A Study
of Rural Andhra Pradesh', Anvesak, 7.
-—— (I980a). "ASurvey of Literature on Poverty Income Distribution and
Development*, Artha Vijnana, 22.
-(1980ft). 'Poverty: Concepts and Measurement', Indian Journal of
Economics, 61,
SAWHILL, I. V. (1988). 'Poverty in the U.S.: Why is it so Persistent?',
Journal of Economic Literature, 26,
SCANMJN, T, M, (1975). 'Preference and Urgency', Journal of Philosophy, 72.
(1982). 'Contractualisrn and Utilitarianism', in Sea and Williams
(1982).
— (1988o). 'The Significance of Choice', in Tanner Lectures on Human
Values, viii (Salt Lake City; University of Utah Press).
(19884), 'Notes on Equality*, mimeographed, Harvard University.
— (1990). 'The Moral Basis of Interpersonal Comparisons', mimeo-
graphed, forthcoming in J. Elster and J, Roemer (eds,), Interpersonal
Comparability,
•(1992), 'Value, Desire and Quality of Life', in Nussbaum and Sen
(1992).
SCHEFFLER, S. (1982). The Rejection of Consequentialism (Oxford: Claren-
don Press),
— (1988) (ed,), Consequentialism and Its Critics(Oxford: Oxford Univer-
sity Press).
SCUELUNG, T, C, (1960). The Strategy of Conflict (Cambridge, Mass.: Har-
vard University Press).
SCHBLJUNG, T. (1984). Choice and Consequences (Cambridge, MA: Harvard
University Press).
SCHOKKAERT, E., and VAN OOTEOEM, L, (1990), 'Sen's Concept of the Liv-
ing Standard Applied to the Belgian Unemployed*, Recherches economi-
ques (if Louvain, 56.
School of Public Health, Harvard University (1985). Hunger in America:
The Growing Epidemic (Cambridge, Mass.: School of Public Health,
Harvard University).

188 References
SCHOTTER, A, (1985). Free Market Economics; A Critical Appraisal (New
York: St Martin's Press).
SCHWAETZ,T.(1981).'The UniversalInstabilityTheorem*, Public Choke,37,
— (1986). TheLogicof Collective Choice (NewYork: Columbia Univer-
sity Press).
SCMMSHAW, N. (1987), 'Biological Adaptation in the Maintenance of
Nutrition and Health', mimeographed, MIT.
— TAYLOR, C. E.,and GQPALAN, J. E,(1968). Interactions of Nutritionand
Infection (Geneva: World Health Organization).
SEABJBGHT, P. (1989). 'Social Choice and Social Theories', Philosophy and
Public Affairs, 18.
-— (1992). "Pluralism and the Standard of Living', in Nussbaum and Sen
(1992).
SHADE, J. (1977). 'On the Shape of Optimal Tax Schedules', Journal of
Public Economics!, 1.
SEASTEAND, E, and DIWAN, R. (1975). 'Measurement and Comparison of
Poverty and Inequality in the United States', presented at the Third
World Econometric Congress, Toronto,
SEIDL, C. (1975). 'On Liberal Values', Zeitschrift fur Nationaldkonomie, 35.
—— (1986a). 'Poverty Measures: A Survey', in Bos, Rose, and Seidl (1986).
-— (19866). 'The Impossibility of Nondictatorial Tolerance*, Journal of
Economics: Zeitschrift fur Nationaldkonomie, 46.
- (1990). 'On the Impossibility of a Generalization of the Libertarian
Resolution of the Liberal Paradox", Journal of Economics, 51.
SEN, A. K. (1970a). Collective Choice and Social Welfare (San Francisco:
Holden-Day), (Republished Amsterdam: North-Holland, 1979.)
(19706). Interpersonal Aggregation and Partial Comparability*, Eeo-
nametrica, 38. (Repr. in Sen 1982«.) (*A Correction*, Econornetrica, 40,
1972.)
——(1970c). 'The Impossibility of a Paretian Liberal" Journal of Political
Economy, 78. (Repr, in Hahn and Hollis 1979 and Sen 1982a.)
——(1973a). On Economic Inequality (Oxford; Clarendon Press and New
York: Norton). (Also referred to in the text as OEI.)
(19736). 'Behaviour and the Concept of Preference', Economics, 40.
(Repr. in Sen 1982o.)
-—-(1973c). "Poverty, Inequality and Unemployment; Some Conceptual
Issues in Measurement*, Economic and Political Weekly, 8.
(1913d). 'On the Development of Basic Economic Indicators to Supple-
meat GNP Measures', United NationsEconomic Bulletinfor Asia and the
Far East, 24.
(1974). 'Informational Bases of Alternative Welfare Approaches:
Aggregation and Income Distribution', Journal of Public Economics, 4,
(1976a), 'Poverty; An Ordinal Approach to Measurements', Econorne-
trica, 44. (Repr. in Sen 1982a.)

References 189
(19766). 'Real National Income', Review of Economic Studies, 43.
(Repr. in Sen 1982a.)
(1976c). 'Liberty, Unanimity and Rights', Economica,43.(Repr. inSen
1982a.)
—- (1977«). 'Social Choice Theory: A Re-Examination', Econometnca,
45. {Repr. in Sen 1982a.)
— (1977ft). 'On Weights and Measures: Informational Constraints in
Social Welfare Analysis", Econometrics, 45, (Repr. in Sen 1982a.)
— (1977c). 'Rational Fools: A Critique of the Behavioural Foundations
of Economic Theory', Philosophy and Public Affairs, 6, (Repr. in Sen
1982«.)
-— (1978a), 'On the Labour Theory of Value: Some Methodological
Issues*, Cambridge Journal of Economics, 2.
(1978ft). 'Ethical Measurement of Inequality; Some Difficulties", in
Krellc and Shorrocks (1978). (Repr. in Sen 1982«.)
—— (1979a). 'Personal Utilities and Public Judgements:or What's Wrong
with Welfare Economics?" Economic Journal, 89. (Repr. in Sen I982a.)
(1979ft), 'Utilitarianism and Welfarism*, Journal of Philosophy, 76.
— (1979c). 'The Welfare Basis of Real Income Comparisons', Journal of
Economic Literature, I, (Repr, in Sen 1984.)
(1979d), 'Informational Analysis of Moral Principles', in R. Harrison
(ed), RationalAction (Cambridge: Cambridge University Press).
—— (1979e). 'Issues in the Measurement of Poverty', Scandinavian Journal
of Economics, 81.
—— (1980a). 'Equality of What?*in McMurrin(1980). (Repr. in Sen 1982o;
and in Rawls et al 1987.)
— (19806). 'Description as Choice", Oxford Economic Papers, 32.(Repr.
in Sen 1982o.)
(1980-81). 'Plural Utility', Proceedings of the Aristotelian Society,
80.
(1981a). Poverty and Famines; An Essay on Entitlement and Deprivation
—— (1981ft),'Public Action and the Quality of Life in Developing Coun-
tries', Oxford Bulletin of Economics and Statistics, 43.
(1982a). Choice, Welfare and Measurement (Oxford: Blackwell and
Cambridge, Mass.: MIT Press).
— (19826). 'Rights and Agency', Philosophy andPublic Affairs, 11. (Repr.
in Scheffler 1988.)
(1982c). "Libertyas Control: An Appraisal", Midwest Studies in Philos-
ophy, 7.
(1983a). 'Liberty and Social Choice", Journal of Philosophy, 80.
(19836). 'Evaluator Relativityand Consequential Evaluation', Philos-
ophy and Public Affairs, 12.
— (1983c). 'Development: Which WayNow', Economic Journal, 93.

190 References
— (I983<0- 'Poor, Relatively Speaking", Oxford Economic Papers, 35,
(1984). Resources, Values and Development (Oxford; BJackwell and
Cambridge, Mass,: Harvard University Press),
(1985a). 'Well-being, Agency and Freedom; The Dewey Lectures
1984', Journal of Philosophy, 82,
— (1985ft). Commodities andCapabilities(Amsterdam; North-Holland).
—— (1985c). *AReply to Professor Townsend', Oxford Economic Papers,
37.
(9$5J), 'Women, Technology and Sexual Divisions', Trade and De-
velopment, 6.
(1985e). The Moral Standing of the Market', Social Philosophy and
Policy, 2.
—— (1986a). 'Social Choice Theory*, in Arrow and Intriligator (1986).
(19866). "Information and Invariance in Normative Choice', in
Heller, Starr, and Starrett (1986),
(1987). On Ethics and Economics (Oxford: Blackweli),
•—— (1988a). 'Freedom of Choice: Concept and Content', EuropeanEcon-
omic Review, 32.
—— (1988ft). The Concept of Development', in Chenery and Srinivasao
(1988), vol. I.
—— (1988c). 'Africa and India; What do we have to Learn from Each
Other?' C, N. Vakil Memorial Lecture, 8th World Congress of the
International Economic Association; published in K.. J. Arrow (ed.),
The Balance between Industry and Agriculture in Economic Development
(London: Maemillan).
—— (1989«). 'Women's Survival as a Development Problem', Bulletin of
the American Academy of Arts and Sciences, 43 (Nov.), (A revised
version published in New York Review of Books, Christmas Number,
1990.)
—— (19896). 'Economic Methodology; Heterogeneity and Relevance',
Social Research, 56.
•—— (1990o). 'Welfare, Freedom and Social Choice: A Reply', Recherches
iconontiques de Louvain, 56,
——(19906). 'Justice: Means versus Freedoms', Philosophy and Public
Affairs, 19.
(1990c). 'Gender and Cooperative Conflicts', in Tinker (1990a).
— (1991a). 'Welfare, Preference and Freedom', forthcoming inJournal of
Econometrics,
(19911?). 'Well-Being and Capability', forthcoming in Nussbaum and
Sen (1992).
—-(199k). The Nature of Inequality', in Arrow (1991).
—~(l992a). 'Minimal Liberty', forthcoming in Economica,
— (19926), 'Markets and Freedoms', text of John Hicks Lecture, forth-
coming in Oxford Economic Papers,

References 191
— (1992c). 'On Indexing Primary Goods and Capabilities', mimeo-
graphed, Harvard University.
— and SENGUPTA, S, (1983), 'Malnutrition of Rural Indian Childrenand
the Sex Bias*, Economic and Political Weekly, 19 (Annual Number).
and WILLIAMS, B. (1982) (eds.). Utilitarianism and Beyond (Cambridge:
____ MUBLLBAUER, J., KANBUR, R., HART, K., and WILLIAMS, B, (1987). The
Standard of Living (Cambridge; Cambridge University Press).
SEN, P. K. (1986). 'The Gini Coefficient and Poverty Indexes: Some Recon-
ciliations', Journal of the American Statistical Association, 81.
SERAGELBIN, I. (1989). Poverty, Adjustment and Growth in Africa (Wash-
ington, DC: The World Bank).
SHARMA, U. (1980). Women, Work and Property in North-West India
(London: Tavistock).
SHESHINSKI, E. (1972), 'Relation between a Social Welfare Function and
the Gini Index of Inequality', Journal of Economic Theory, 4.
SHKLAR, J. (1990). TheFaces of Injustice (NewHaven: Yale UniversityPress).
SHORROOCS, A. F. (1980), 'The Class of Additively Decomposable In-
equality Measures', Eeonometrica, 48.
(1982). 'Inequality Decomposition by Factor Components', Eco~
nometrica, 50.
(1983). 'Ranking Income Distributions*, Economica, 50.
(1984). 'Inequality Decomposition by Population Subgroups', Eeo-
nometrica, 52.
——(1988). 'Aggregation Issues in Inequality Measurement*, in Eichhorn
(1988a).
and FOSTER, J. B. (1987). 'Transfer Sensitive Inequality Measures',
SIDOWICK, H. (1874). The Method of Ethics (London: Macmillan).
SILBEK, J. (1983). 'ELL (Equivalent Length of Life) or Another Attempt at
Measuring Development', World Development, 11.
SLOTTJE, D. J. (1984). "An Analysis of the Impact of Relative Price
Changes on Inequality in Size Distribution of Various Components of
Income: A Multidimensional Approach*, Southern Methodist Univer-
sity, Dallas.
SMART, J. J. C., and WILLIAMS, B. (1973). Utilitarianism; For and Against
(Cambridge: Cambridge University Press),
SMEBDINO, T., RAINWATER, L., and O'HiooiNS, M. (1988). Poverty, In-
equality, and the Distribution of Income in an International Context
(Brighton: Wheatsheaf).
SMITH, ADAM, (1776). An Inquiry into the Nature and Causes of the Wealth
of Nations, (Republished London: Home University, 1910.)
——(1790), The Theory of Moral Sentiments, revised edn. (Republished
Oxford: Clarendon Press, 1975.)

192 References
SOLOW, R, M. (1984). 'Relative Deprivation?' Partisan Review,51.
SONSTEGAARD, M. (1987). 'A Reply to Perelli-Minettf, Theory and De-
cision, 22; 3.
SPARROW, J. (1977), Too Much of a Good Thing (Chicago: University of
Chicago Press).
SRINIVAS, M. S, (1962). Caste in Modern India and Other Essays (Bombay:
Asia Publishing House).
SRINIVASAN, T. N. (1981). 'Malnutrition: Some Measurement and Policy
Issues', Journal of Development Economics, 8.
(1992). 'Undernutrition; Concepts, Measurements, and Policy Impli-
cations', in Osmani (1992«).
and BARDHAN, P. (1974) (eds.). Poverty and Income Distribution in
India (Calcutta: Statistical Publishing Society).
____(1988) (eds.). Rural Poverty in South Asia (New York:
Columbia University Press).
STARRETT, D. A. (1988). Foundations of Public Economics (Cambridge:
STBINBR, H. (1982). 'Individual Liberty', Proceedings of the Aristotelian
Society, 82.
(1990). 'Putting Rights in their Place', Recherches econamiqws de
Louvain, 56.
STERN, N. H. (1976). 'On the Specification of Models of Optimum
Income Taxation*, Journal of Public Economics, 6,
STEVENS, D., and FOSTER, J. (1978). "The Possibility of Democratic
Pluralism', Economica, 45.
STEWART, F, (1985). Planning to Meet Basic Needs (London: Macmillan).
——(1988). 'Basic Needs Strategies, Human Rights and the Right to
Development', Development Studies Working Paper No. 2, Queen
Elizabeth House.
STIGLITZ, J. E. (1982), 'Utilitarianism and Horizontal Equity: The Case
for Random Taxation', Journal of Public Economics, 18.
STOCKER, M. (1990), Plural and Conflicting Fater (Oxford: Clarendon Press).
STRASNICK, S. (1976). 'Social Choice Theory and the Derivation of
Rawls* Difference Principle*, Journal of Philosophy, 73.
STKEBTEN, P. (1981). Development Perspectives (London: Macmillan).
—— (1984). 'Basic Needs: Some Unsettled Questions', World Develop-
ment, 12.
—-with BUKKI, S. J., HAQ, MAHBUBQUL, HICKS, N., and STEWART, F.
(1981). Firxl Things First; Meeting Basic Needs in Developing Countries
(London: Oxford University Press).
SUIRAMANIAN, S. (1987). 'The Liberal Paradox with Fuzzy Preferences',
Social Choice and Welfare, 4.
SUODBN, R. (1981). The Political Economy of Public Choice (Oxford:
Martin Robertson).

References 193
(1985). 'Liberty, Preference and Choke', Economics and Philosophy, I.
——- (l9S6),'K.evlew of CotnmoditiesandCapabilities',Economic Journal,96.
—— (1989). 'Maximizing Social Welfare, Is it the Government's Busi-
ness?* in A. Hamlin and Pettit (1989),
SUKHATME, P. V. (1977). Nutrition and Poverty (New Delhi: Indian
Agricultural Research Institute).
— (1982), 'Autoregulatory Homcostatic Nature of Energy Balance',
American Journal of Clinical Nutrition, 35.
SUNOAEAM, K,,and TENDULKAR, S. D. (1981).'Poverty Reductionin the Sixth
Plan', Working Paper 233, Delhi School of Economics.
SUPHES, P.(1966). 'Some Formal Models of Grading Principles', Synthese, 6.
(1977). The Distributive Justice of Income Inequality',Erkenntnis, 11.
(1987). 'Maximizing Freedom of Decision: An AxiomaticAnalysis', in
Feiwel (1987).
(1988). 'Lorenz Curves for VariousProcesses: A Pluralistic Approach
to Equity', Social Choice and Welfare, 5.
SUZWMURA, K. (1978). "Onthe Consistencyof Libertarian Claims', Review of
Economic Studies, 45, fA Correction', 46.)
—— (1980). 'Liberal Paradox and the Voluntary Exchange of Rights-
Exercising', Journal of Economic Theory, 22,
(1983). Rational Choice, Collective Decisionsand Social Welfare (dan-
bridge: Cambridge University Press).
— (1988). 'Introduction' to the Japanese trans, ofCommodities and Capa-
bilities (Tokyo: Iwanami).
(1991).'AlternativeApproaches to LibertarianRights',inArrow(1991).
SVEDBERG, P. (1988). 'Undenratrition in Sub-Saharan Africa: Is there a Sex
Bias?' Working Paper 46, WIDER, Helsinki.
(1990). 'Undernutrition in Sub-Saharan Africa: A Critical Assessment
of the Evidence", in Dreze and Sea (1990), vol. 3.
SWAMINATHAN, M. (1988). 'Inequality and Economic Mobility', D.Phil, dis-
sertation, Oxford University.
SZAL, R, J. (1977). 'Poverty Measurement and Analysis', ILO Working
Paper WEP 2-23/WP60,
Szrao, G. G. (1987). "Hirschmsn versus Herfindahl: Some Topological
Properties for the Use of Concentration Indexes', Mathematical Social
Sciences, 14.
SZKETBS, S.(1986). 'TheImportanceof Social Intervention in Britain's Mor-
tality Decline C.18SO-1914: A Reinterpretation', Discussion Paper 121,
Centre for Economic Policy Research, London.
TAKAYAMA, N. (1979). "Poverty,Income Inequalityand their Measures: Pro-
fessor Sen's Axiomatic Approach Reconsidered', Econometrica, 47.
TAWNEY, R. H. (1931). Equality (London: Allen & Unwin).
TAYLOR, C. (1979). 'What's Wrong with Negative Liberty?', in Ryan (1979).
— (1982), 'The Diversity of Goods', in Sen and Williams (1982).

194 References
TAYLOR, L. (1977). 'Research Directions in Income Distribution, Nutrition
and the Economics of Food', Food ResearchInstitute Studies, 15.
TEMKIN, L. S. (1986), 'Inequality', Philosophy and Public Affairs, 15.
—— (1989), 'Inequality', D.Phil, thesis (Oxford);revised version to be pub-
lished by Clarendon Press, Oxford.
THEIL, H. (1967), Economicsand Information Theory (Amsterdam: North-
Holland).
THOMSON, W., and VAWAN, H, (1985). Theories of Justice Based on Sym-
metry1
, in Hurwicz, Schmeidler, and Sonnenschein (1985).
THON, D. (1979). *On Measuring Poverty', Review of Income and Wealth^
25.
THON, D. (1982), 'An AxioniatizatioB of the Gini Coefficient', Mathematical
Social Sciences, 2,
THUSOW, L. D. (1975). Generating Inequality (New York: Basic Books).
——(1987). 'A Surge in Inequality', Scientific American,256.
TILLY, L. A. (1983). 'Food Entitlement, Famine and Conflict', in Rotberg
and Rabb (1985).
(1985), 'Sex and Occupation in Comparative Perspectives', mimeo-
graphed, New School for Social Research, New York.
TINBBRGBN, J, (1970). 'A Positive and Normative Theory of Income Dis-
tribution*, Review of Income and Wealth, 16,
TINKER, I. (1990a) (cd.). PersistentInequalities (New York: Oxford Univer-
sity Press).
— (19906). 'A Context for the Field and for the Book",inTinker (19900).
and BRAMSEN, M. B. (1976). Women and World Development (Wash-
ington, DC: Overseas Development Council).
TOWNSEND, P, (1979). Poverty in the United Kingdom (Harmondsworth:
Penguin).
(1985). 'A Sociological Approach to the Measurement of Poverty: A
Rejoinder to Prof. Amartya Sen1
, Oxford Economic Papers, 37.
TSAKLQGLQU, P. (1988). *A Family of Decomposable Poverty Indices',
unpub. manuscript, University of Bristol
TUQMALA, M. (1984). 'On the Optimal Income Taxation: Some Further
Numerical Results', Journal of Public Economics,17,
— (1990). Optimal Income Tax and Redistribution (Oxford: Clarendon
Press).
UNDP (1990). The Human Development Report 1990 (New York: United
Nations Development Programme).
(1991). The Human Development Report 1991 (New York: United
Nations Development Programme).
UNICEF (1987). The State of the World's Children 1987(Oxford: Oxford
University Press).
UNICEF (1992), The State of the World's Children 1992 (Oxford:
Oxford University Press).

. References 195
USHER, D, (1968), The PriceMechanism and the Meaning of National Income
Statistics (Oxford; Clarendon Press).
VAIDYANATHAM, A. (1985), 'Food Consumption and the Size of People:
Some Indian Evidence", Economicand Political Weekly, 20,
—-—(1987). 'Poverty and Economy: The Regional Dimension', mimeo-
graphed, paper presented at the Workshop on Poverty in India, Queen
Elizabeth House, Oxford,
VAIUENTYNE, P. (1989). 'How to Combine Pareto Optimally with Liberty
Considerations', Theory and Decision,27,
VAN GINNEKEN, W. (1980), 'Some Methods of Poverty Analysis: An Appli-
cation to Iranian Data 1975-76', World Development, 8.
VAN PARIJS, P, (1990a). 'Equal Endowments as Undominated Diversity*,
Recherches economiques de Louvain, 56,
(19906). The Second Marriage of Justice and Efficiency', Journal of
Social Policy, 19,
— (1991). "Why Surfersshould beFed: The Liberal Case for an Uncondi-
tional Basic Income", Philosophy and Public Affairs, 20,
VAN PRAAG, B, M. S. (1968). Individual Welfare Functions and Consumer
Behaviour (Amsterdam: North-Holland),
— (1978), 'The Perception of Welfare Inequality', European Economic
Review, 10.
(1992). 'The Relativity of the Welfare Concept*, in Nussbaum and Sen
(1992),
• HAGENAARS, A, J,M.,andVANWEEWEN,H,(1982).'Poverty in Europe*,
Journal of Income and Wealth, 28.
VAWAN, H. (1974). "Equity, Envy and Efficiency', Journal of Economic
Theory, 9,
(1975). 'Distributive Justice, Welfare Economics and the Theory of
Fairness', Philosophy and Public Affairs, 4.
VAUGHAN, M. (1985), 'Famine Analysis and Family Relations", Past and
Present, 108.
-— (1987). TheStory of an African Famine: Gender andFamine in Twentieth
Century Malawi (Cambridge; Cambridge University Press).
VAUGHAN, R. N. (1987). 'Welfare Approaches to the Measurement of
Poverty", EconomicJournal, 97.
VEEBA, S., et al. (1987). Elites and the Idea of Equality (Cambridge, Mass.:
Harvard University Press),
VISARIA, P. (1961). The Sex Ratio of the Population of India, Monograph 10,
Census of India 1961 (New Delhi: Office of the Registrar General).
WAAL, A. DE (1989). Famine that Kills (Oxford: Clarendon Press).
WALDRON, I, (1976). 'Why do Women Live Longer than Men?' Social
Science and Medicine, 10.
(1983). 'The Role of Genetic and Biological Factors in Sex Differences
in Mortality", in Lopez and Ruzicka (1983).

196 References
WAUJRON, J, (1984) (ed.). Theories of Rights (Oxford: Oxford University
Press).
WALSH, V, (1964), 'Discussion:The Status of Welfare Comparisons', Philos'
ophy of Science, 31.
—— (1987). 'Philosophy and Economies', in J, Ealwell, M. Milgate, and
P. Newman (eds,), The New Palgrave: A Dictionary of Economics, iii
(London: Macmillan),
(1991), 'Rationality, Allocation and Reproduction: Some Key Con-
cepts of Microtheory', mimeographed.
WALZER, M. (1983). Spheres of Justice: A Defence of Pluralism and Equality
(Oxford: BlackweU).
— (1992), 'Objectivity and Social Meaning', inNussbaum and Sen (1992).
WEALE, A, (1980). 'The Impossibility of Liberal Egalitarianism', Analysis,
40.
WEBSTER, M. (1986). 'Liberals and Information', Theory and Decisions, 20.
WBDDBWJURN, D, (1961). The Aged in the Welfare State (London: Bell),
WESTBN, P. (1982). "The Empty Idea of Equality*, Harvard Law Review, 95.
WEYMARK, J, (1981). 'Generalized Gini Inequality Indices', Mathematical
Social Sciences, L
WHJTEHEAD, A. (1985). 'Gender and Famine in West Africa", Review of
African PoliticalEconomy.
-—- (1990). 'Rural Womenand Food Production in Sub-Saharan Africa*,
in Dreze and Sen (1990), vol. 1.
WIGGINS, D. (1985), 'Claims of Need', in Honderich (1985).
(1987). Needs, Values, Truth (Oxford: Blackwell),
WILLIAMS, A. (1985). 'Economics of Coronary Bypass Grafting*, British
Medical Journal, 291 (3 Aug.),
— (1991). 'What is Wealth and Who Creates It?', in J. Mutton, S.
Hutton, T, Pinch, and A, Shicll (eds.), Dependency to Enterprise
(London: Routledge).
WILLIAMS, B, (1962), The Idea of Equality', in P. Lastett and W. G,
Runisman (eds.), Philosophy, Politics and Society, Second Series
(Oxford: BlackweU).
. (1972). Morality: An Introduction to Ethics (New York: Harper &
Row).
(1973a), Problems of the Self (Cambridge: Cambridge University
Press).
(1973d). 'A Critique of Utilitarianism', in Smart and Williams(1973).
(1981). Moral Luck (Cambridge: Cambridge University Press).
(1985). Ethics and the Limits of Philosophy (London: Fontana and
Cambridge, Mass.: Harvard UniversityPress).
— (1987). 'The Standard of Living: Interests and Capabilities', in Senet
al. (1987).
WILSON, G. (1987). Money in the Family (Aldershot:Avebury).

References 197
WOLF, M. (1987). Revolution Postponed: Women in Contemporary China
(Stanford, Calif.: Stanford University Press).
WOLFSON, M. C. (1974), Strength of Transfers, Stochastic Dominance, and
the Measurement of Economic Inequality (Ottawa; Statistics Canada).
WOLLHEIM, R, (1955-6), 'Equality and Equal Rights', Proceedings of the
Aristotelian Society, 56,
WOULD BANK (1984), China: The Health Sector (Washington, DC: The
World Bank),
-— (1990). The World Development Report 1990 (Oxford: Oxford Univer-
sity Press).
— (1991), The World Development Report 1991 (Oxford: Oxford Univer-
sity Press).
WRICLESWORTH, J, (1982). 'The Possibility of Democratic Pluralism: A
Comment', Economica, 49,
(1985). Libertarian Conflicts in Social Choice (Cambridge: Cambridge
University Press).
WYON, J, B., and GORDON, I. E. (1971). The Khanna Study (Cambridge,
Xu, Y, (1990). 'The Liberal Paradox: Some Further Observations', Social
(1991). 'Urgency and Freedom*, mimeographed, Murphy Institute,
Tulane University,
YAARI, M, E., and BAR-HOLEL, M. (1984). 'On Dividing Justly", Social
YITZHAKI, S. (1979). 'Relative Deprivation: A New Approach to the Social
Welfare Function', Quarterly Journal of Economics, 93.
YOONG, H, P, (1986) (ed.). Fair Allocation,American Mathematical Society,
Symposia in Applied Mathematics; forthcoming.
YOUNG, K,, WOLKOWITZ, C., and MeCuiAAOH, R. (1981) (eds.). On Mar-
riage and the Market: Women's Subordination in International Perspective
(London: CSE Books),
YSANDEI, B. (1992). 'Comment on Erikson', in Nussbaum and Sen (1992),
ZADEH, L. A. (1965), 'Fuzzy Sets', Information and Control, 8.
ZAMAONI, S, (1986). 'Introduzione', in A. K, Sen, Scelta, Benessere, Equita
(Bologna: II Mulinao),
ZEUTHEN, F. (1930). Problems of Monopoly and Economic Welfare (London:
Routledge).

Name Index
Aberg, R, 39,42, 113,JJ6
Ackerman, B, A. 18, 77
Adeiman, I, 109
AgarwaJ, B, 124, 125
Ahluwalia, M. S, 105
Ahmad, E. 125
Ahmed, I. 124
Ahtisaari, M. 42, 51
Aigncr, D, J. 93
Akertof, G. 60, 142
Alamgir, M. 105
AWrich, 3, 32
Aliardt, E. 39, 113, 116
Allen, T. 33, 124
Amid, Y. 95
Amsden, A. H, 122
Anand, S. mi, 38, 93, 103, 105,106
Annas, J. 122
Apffd Marglin, F. 59
Archibald, G. C, 93
Aristotle S, 39, 52, 91
Arneson, R. 33, 42, 45
Arrow, K. J. xiii, 32, 74, 104
Article 19 72
Asahi, J. 42
Aslanbeigui, N. 125
Atkinson, A, B, xiii, 24, 29-30, 89, 93,
94,95-101, 105, 106, 113, 118,S21,
138, 141,142
Auerbach, A. 141
Auersperg, M. 93, 121
Austen-Smith, D. 32
Babu, S, C, 105
Banerjee, N. 123, 124, 125
Banerji, C, xiii
Banister, J. 123, 124
Bardhan, K, 124
Bardhan, P. xiii, 109, 123, 124
Bar-Hillel, M, 95
Barnes, J. 32
Barry, B, 32
Barten, A. P. 121
Basu, A. 123, 124'
Basu, K, xiii, 32, 42, 47, 93
Batra, R, 32
Bauer, P. T. xiii, ! 5, 119, 139
Baumol, W. J. 74, 93
Bayer, R. 118
Beckerman, W. xiii, 103, 107
Behrman, J, R. 123, 124
Beitz, C. R. 42, 45
Bellman, R, E. 47
Beneria, O. 124
Bentham, J, 14, 139
Bentzel, R. 93
Bergmans?, B, 122
Bcrgson, A. 32
Berlin, I. 22, 41, 67
Bernholz, P, 32
Besley, T, 106
Biteille, A, xiii, 14, 44, 117, 118,122
Beus, J. de xiii, 42
Bezembinder, T. 47
Bhargava, A. 38
Ihattacharaya, N, 93-4, 127
Bhatty, I. Z. 105
Bigtnan, D, 106
Blackorby, C. xiii, 29,47, 74,93,94 97,
99, 105, 106, 121,122
Blair, D. H. 84
Blau, J, H. 32
Blaxter, K, 37
Bliss, C. L xiii, 42
Boltvinnik, J. 90, 109
Borda, J. C, 104
Boserup, E. 123,124
Bourguigrwn, F, 93, 94, 106, 121
Bramsen, M. B. 123
Brandt, R. B. 15
Branncn, J. 42, 125
Brenkert, C. G. 41
Breyer, F. 32
Brison, S, xiii, 57
Brock, D. 42, 44
Brodcr, I. E. 93, 94
Broome, J. xiii, 56, 88
Brown, M. 124
Bryceson, D. F. 125
Buchanan, A. E. 41, 118
Buchanan, J, M. xiii, 13, 22, 41, 136,
140

200 Name Index
Buhmann, B, 106, 121 Dworkin, G. 59
Buvinic, M. 123, 124 Dworkin, R. xiii, 8, 12-13,14, 19, 33-4,
37-8, 42, 74, 77, 80, 89, 148
Campbell, D. E. 32, 67
Caplan, A. L, 118 Ebert) jj, 93
Case, A. 113 Edgeworth, F. Y. 6, 14
Chakravarty, L. 105 Edwards, R C 118
Chakravarty, S, 89 Eichhorn, W. 93-4
Chakravarty, S. R, 47, 93, 105, 106 Eister> j. J2> 41, 59, $7, jig
Chapman, B. 32 Erikson, R, 39, 42, 113, 116
Qwtterjee, G, S. 93-4, 127
Chen, L, C. 123, 127
Chen, M, 123, 124, 125,
CWcWlnisky, G, 109
Chipman, J, 93 Feldstein, M. 141
Clark, J, B. 119 Ferejohn, J. A. 32
Clark, S. 103, 105, 106
Cohen, G.A. xiii, 33,41,42,45, 51,65,
Ug Fisher, F. M, 29,94
Cohen, M. 118 Fishkin, J. S. 77
Coughlin, P. C. 32 Fwblow, A. 105
Cowell, F. A, 93, 95, 105, 106, 132 Folbrc, N. 122,124
Crocker, D, A. 42, 44 Fole
y* D
- 7
f
Culyer, A. J. 39, 42 Foster, J. xiii, 29,32,93-4,97,104,106,
Dacy D xiii Fountain, J. 32
DahTMdorf, R, xiii, 71, 89, 118 tmMnit'^'JS;^
Dalttm, H. 21, 24, 89, 95-100, 118 freeman, H, 1. 114
Daniels, N, 75, 118 Fuchs, V. R. 39
Danziger, S, H. 113
Das Gupta, M, 123, 124 Gaertner, W. xiii, 32, 33, 42, 67
Dasgupta, P, xiii, 24, 38, 42, 93, 94, 96, Gaiha, R. 105
105, 109 Galbraith, J, K. 71
d'Aspremont, C. xiii, 74, 96 Gandhi, Mahatma 1J2
Davidson, D, 55 GSrdenfors, P, 32, 33
Deaton, A. xiii, 93, 121, 122, 125 Gardiner, J. 124
Deb, R, 33 Gardner, R. 32
Delbono, F. 42 Gastwirth, J, L. 93
De Leonardo, G. 42 Gauthier, D. 18, 139
Deolalikar, A. B. 123 Gchrig, W. 93
Desai, M, J. xiii, 42, 51, 90, 109, 116 Gevers, L. xiii, 74, 96
Diwan, R, 105 Gibbard, A. 32, 55, 84
Dixon, R, 124 Gigliotti, G, A. 32, 33
Dobb, M, H. 118, 119 Gtentz. M. H. 77
Donald, A. xii Glover, J. xiii, 56
Donaldson, D. 29, 74, 93, 96, 97, 99, Goldin, C. 122
105, 106, 121, 122 Goodin, R, E, 15, 22, 23, 42, 110, 118
Drewnowski, J. 106 Gopalan, C. 37
Drfee, J, xiii, 29, 42, 65, 72, 109, 115, Gosling, J, C. B, 43
124, 125, 126, 128, 138 Gouguen, J. A. 47
D'Souza, D, 123 Graaff, J, de V. 35, 93, 94, 105
D'Souza, F, 72 Grant, J. P. 109
Dutta, B. xiii, 93-4, 105 Green, E, T. 67
132
Farell, M. J. 32, 67
Fei, J. C. H.
Fei, J. C. H. 93
Fields , G. S. 93, 95, 105, 106
Fine, B. J. 32,

Name Index 201
Greer, J. 93, 106
Griffin, J. xiii, 15, 44, 54
Griffin, K. xiii, 42, 109
Gutmaan, A. 22
Haddad, L. 123
Hagenaars, A. J. M, 114
Hahn, F, H. 94
Haksar, V. 22, 45
Halbertal, M. xiii
Hamada, K. 106
Hamlin, A. 33, 42
Hammond, P, I. xiii, 32, 67, 74, 93,
94, 105
Hampshire, S, 17
Hansson, B, 93, 98
Hansson, S, O. 33
Haq, Mahbubul xiii
Hare, R, M. xiii, 3, 14, 17, 18, 56,
108, 137, 139
Hard, A, 33
Barman, G, 18
Harris, C, 38
Harriss, B. 123, 124
Harsanyi, J, C. 6, 14, 18, 74, 137, 139
Hart, H. L. A. 75, 87
Hart, K. 31, 42
Hausrnan, D, M, 60
Hawes, S, xiii
Hawthorn, G, 31, 42
Hayek, F. A. 22
Heal, G. 94
Heins, A, J. 93
Helm, D. 42
Hemming, R. 105, 106
Hicks, I. R. 6, 35
Bimmelfarb, G. Ill
Himmelwdt, S. 124
Hirschman. A, O. xiii, 56, 57, 60
Hobsbawro, E, xiii, 118
Homey, M. L 124
Hossain, I. 42, 109, 134
Houthakker, H. S, 35
Huq, E. 123
Hurley, S. 18, 33, 56, 132
Hussain, A. xiii
Hutchens, R, M. 93
Hylland, A. 32, 67
Jain, D, 123, 124, 125
Jasso, G, 93
Jayawardena, K. xiii
Jayawardena, L, xiii
Jencks, C. 118
Jennings, J, xiii
Johansson, S. R. 123
Jorgenson, D, W, 93, 106, 121
Kakwani, N. xiii, 42, 93, 105, 106
Kanbur, R. [S. M. R,] xiii, 31, 42, 93,
106, 123, 139
Ranger, S. 132
Kant, I. 17, 151
Kapteyn, A. 114
Kami, E, 32
Katz.ll. 113
Kazmi, N. A. 105
Kelly, J, S. 32, 67
Kelsey, D. 32
Keohane, R, xiii
Kern, L. 93
Klasen, S. xiii
Klein, F. x
Knight, J, xiii, 42, 109
Kolakowski, L. 41. 118
Koto, S. Ch, 24, 29, 30, 74, 93, 94,
95, 97, 105
Kooptnans, T. C. 51, 63
Korsgaard, C. 151
Kornai, J. 31
Krcps, D, 51, 63
Kruger, L. 32, 67
Kuga, K. 93
Kumar, B. G. 42, 123, 127
Kumar, A. K. S, 42
Kundti, A. 106
Kuznets, S. xii, 118
Kynch, J, 86, 123, 124, 125, 150
Laden, T, 24, ISO
Lambert, P. J, 93, 132
Lanjouw, P. xiii
Lansley, S. 39, 114
Larmore, C, 77
Laslett, P, 113
Le Breton, K, 93
Le Grand, J. 93, 118, 140
Lehning, P. fl. 38
Lctwin, W. 14-15
Levi, I. xiii, 32, 70, 133
Lewis, G. W. 106
Lindahl, L. 132
Lindbeck, A. 31
Liptou, M, 38, 106
Little, I. M. D. 35
Louiy, G, 117

202 Name Index
Loutfi, M. F. 124
Lucas, J. R, 15
Luker, W. 42
Lukes, S, 22, 4!
Lyeette, M. 123, 124
Lydall, H. F. 118
Maasoumi, E, 93, 94
McCord, C, 114
McCulkgh, R. 124
McDowell, L 18
McElroy, M. B. 124
MeGreevey, W. 123, 124
Maclntyre, I. D, A. 33
Mack, J, 39, 114
Mackie, L L. xiii, 18, 19
Mackintosh, M. 124
McLean, L 32
McPherson, M. S. 60
Majumdar, M, K. 94
Majumdar, T, 35
Manser, M. 124
Marcus, R. B, 70
Marglin, S. A. xiii, 59, 118
Mara, K, 41,52, 118-21
Maskin, E, 74
Maurie, D. 42
Mazutndar, V. 124
Meade, J, E. 21, 138
Meeks, G. 60
Mehran, P. 93
Meyer, M, 93
Mezzetti, C, 33
Miiibaitd, R. 118
Mill, J. S, 14, 132
Miller, B, 123, 124
Mirrlees, J, xiii, 14, 29, 74, 138, 140,
141
Mitre. A, 123
Mitra, T. 94
Mookherjee, D. 93
Morris, C, T, 93, 94, 109
Morris, M. D, 109
Moulin, H, 74
Mttellbauer, J. xiii, 31, 42, 93, 121,
122
Mueller, D. C, 32, 138
Mutnford, L. xiii
Murray, C. 127
Musgrave, R. g9
Nagel, T. 12-13, 18, 19, 56, 61, 118,
132
Wanda, A, R. xiii
Newbery, D. M. G. 93, 105
Ng.Y.-K, 32
Nitzan, S, 33
Nozick, R. xiii, 3, 13, 20-1, 27, 32,
41, 74, 86, 131, 139
Nurmi, H. 33, 67
Nussbaum, M. C. xiii, 39, 42, 45, 61,
70, 91, 109, 122
Nygrad, F, 93
O'Higgins, M. 106
Okin, S. M, 122, 124
O'Neill, O. 18, 42, 122, 151
Osmani, S, R. xiii, 38, 93, 94, 105,
106
Otten, M. W, 114
Pal, P. 93, 127
Palmer, J. 113
Panikar, P. G, K, 127
Pant, P. 109
Pantoto, Y. V, 105
Papanek, H. 124, 125
Parfit, D. xiii, 18, 19, 56
Pattanaik, P. xiii, 32, 33, 63, 68, 106
Payne, P. R. 38
Pazner, E, A. 74, 93
Peacock,A. T. 32
Pen, J. 93
Perelli-Minetti, C. R, 32
Pettit, P. 33
Phclps, E. S. 74, 75, 138
Pigott, A. C, 14
Plott, C, 84
Pogge, T. W. 38, 42, 77
Pollak, R. A. 121
Proudhon, P. J. 118
Putnam, H. xiii, 18
Putnam, R, A. xiii, 85, 122
Putterman, L. 142
Pyatt, G, 93, 105, 106
Quine, W, V, O. x, xiii
Rae, D, xiii, 14, 23
Rainwater, L, 106
Ram, N, 72
Ramachandran, V, K, xiii, 41
Ramsey, F. P. 89
Ranis, G. xiii
Rivailion, M, xiii, 106

Name Index 203
Rawls, J. xi, mi, 8, 10, 12-13, 14, 18,
19, 22, 26-7, 33-4, 37-8, 40, 41-2,
56, 74, 75-87, 89, 91-2, 95, 145-8,
151
Ray, D, 38, 94
Ray, R, 105, 106
Raz, L 77, 132
Reddy, S, ml, 72
Regan, D, H. 19
Reich, M, 118
Riley, J. 32, 33, 42, 67, 74
Ringen, J. 42
Risfcin, C. 126
Roberts, K. W, S. 74, 93, 94, 96, 105,
138
Rochford, S. C. 124
Roemcr, J. 42, 10, 118, 141
Roosevelt, F, 68-9
Rosenzweig, M, R, 123, 124
Ross, D, 52
Rotelli, F. 42
Rothschild, E. xiii
Rothschild, M. 24, 93, 96
Rowley, C, K. 32
Rowntrce, B. S, 111
Roy, T, 47, 93
Rudra, A. xiii
Sadka, E, 140
Samuelson, P. A. 6, 32, 35
Sandstrom, A. 93
Sastry, S. A. R. 105, 106
Scanion, T. M, xiii, 12-13, 17-18,
55, 118
Scheffler, S. 19
Schelling, T, xiii, 56, 59
Schmeidler, D. 74, 93
Schokkaert, E. 151
School of Public Health, Harvard
University 114
Sehotter, A. 32
Schulfer, A. xii
Schultz, T, P. xii, xiii, 123, 124
Schwartz, T, 32
Scrimshaw, N, 38
Seabright, P. 42
Seade, J. 140
Seastrand, F. 105
SeidI, C, 32, 105, 106, 116
Selassie, Emperor Haile 77-8
Sen, P. K, 105
Sengupta, M. 106
Sengupta, S. 123, 124, 125
Shah, H. 116
Shell, K. 29
Sheshinski, E, 93, 105
Shkiar, J. 78
Shorrocks, A. F, xiii, 93-4, 106
Singh, V. P. 147
Slesnick, D, T. 93, 106, 121
Slottje, D, L 93
Smeedtog, T. 106, 113
Smith, Adam 109, 115, 116
Smith, T. E, 106
Soroan, C, R, 127
Sonstegaard, M, 33
Srinivas, M, S, 122
Srinivasan, T. N. xiii, 37
Starrett, D, A. xiii, 24, 93, 96, 105,
138, 141
Steiner, H, xiii, 42
Stern, N, H. xiii, 29, 138, 140
Stevens, D. 32
Stewart, F, 42, 109
Stiglitz, J, 24, 93, 96, 138, 141
Strasnick, S, 74
Streeten, P. 42, 109
Stromberg, J, O, 93
Subramanian, S. 33
Sugden, R, 32, 40
Sukhatme, P. V. 37
SMmmerfield, G. 125
Sundaram, K, 105
Suppes, P. 14, 22, 63, 74
Suzumura, K. xiii, 32, 33, 42, 67, 74,
104, 133
Svedberg, P, 125
Svedrofsky, A. xiii
Szal, R. J, 105
Takayama, N, 106
Tawney, R. H, 21
Taylor, C, 17, 41
Taylor, L. 10S
Temkin, L, S, xiii, 93, 95
Tendulkar, S. D, 105
Theil, H, 93
Thomson, W. 74
Than, D, 93, 106
Thorbecke, E. 93, 106
Thurow, L. D. 118
Tilly, L. A. 125
Tinbergen, J. 93
Tinker, L 123, 124, 125
Torrey, B. 113
Townsend, P. 113, 115, 116

204 Name Index
Trannoy, A, 93
Tuitock, G, 13, 136, 140
Tuomaia, M, 29, 141
Ulph, D, 106
UNDP 90, 109, 125
UNICEF 125
Urtarte, J, R, 93
Usher, D, 115
Vaidyanathan, A. 37, 127
Vallentyne, P. 33
van Acker, P. 47
van de Walle, D. 106
van Ginneken, W, 105
Van Ootegem, L. 151
Van Parijs, P. xiii, 42, 74, 77, I IS
van Praag, B, M. S. 114
van Weeren, H. 114
Varian, H, 74
Vaughan, M. 123,125
Vaughan, R, N. 106
Verba, S, 44, 118
Waldron, I. 123
Wales, T. J. 121
Walsh, V. 60, 70
Wanner, E, xii
Waterlow, 3. C, 37
Watson, E. 123, 124
Weale, A. 32
Webster, M. 32
Wedderburn, D. 113
Weinberg, D. H. 113
Weisskopf, T, 118
Westeo, P, 23
Weymark, J. 74, 93, 96, 106
Whitehead, A. 123
Wiggins, D, 18, 56
Williams, A, 39,42
Williams, B. xiii, 14, 17, 19, 31, 42,
43, 44, 56, 61, 70, 94, 132
Williams, P. xiii
Wilson, O. 42, 125
Wiseberg, L, 77
Wolfson, M. C. 93
Woliheim, R. 22
Woifcowitz, C, 124
World Bank 125
Wriglesworth, I, 32, 67, 74
Xu, Y. 33, 63, 68
Yaari, M. E, 95
Yeati, W. B. 70
Yitzhaki.S, 105
Young, K, 124
Ysander, B. xiii, 39, 113, 116
Zadch, L, A. 47
Zamagni, S, xiii, 42
Zamagni, V. xiii

Subject Index
agency achievement and freedom
5-6, 40, 36-42, 69-72, 132
agency contrasted with well-being
5-6, 56-9, 59-62, 70-2
agency success, instrumental versus
realized 57-8, 64-6
aggregative considerations 7-8,
90-3, 98-101, 136-41, 141-3
ambiguity as a part of conceptual
accuracy 46-8, 48-9, 82-4, 133-5,
144
basal equality 12-13, 19-21, 26, 74,
131, 132-3, 134 137
basic needs 2, 109
basic capabilities 45, 69, 108-9,
109-16, 127-8
Borda procedure of weighting 104-5
capabilities, discrimination and
weighting xi, 4-5, 39, 41-4, 44-6,
46-9, 70, 108-9
capability sets 4-5, 7-10, 39-55,
68-9, 74, 79-87, 88, 91-2, 94, 100,
107-16, 121-8, 129-30, 135,
143-52
caste 7, 121-2, 147
class 7, 9-10, 54-5, 109-12, 117-22
choice, freedom of 4-6, 7-8, 35-6,
37-8, 41-2, 49-53, 54-5, 58, 59-62,
62-4, 64-9, 70, 79-87, 100-1,
111-12, 132, 148-9, 151-2
counterfactuals and the evaluation
of freedom 64-6, 66-9, 135,
149-50
data limitation and its effects 47-9,
52-3, 105-6, 112, 122, 135
deprivation and reduced aspirations
6-7, 54-5, 127, 149-50
desert and productive contribution
14-45, 118-19, 138-9
disability and disadvantage 5, 20,
27, 28-9, 37-8, 81-2, 83-4, 85, 91
diversity of ends 27, 75-9, 82-4, 85-7
diversity of human beings x-xi, 1, 2,
7, 10, 19-21, 27-8, 33, 81-4, 85-7,
88-9, 90-2, 100-1, 109-12, 120-2,
129, 131, 141-3
effective freedom 5-4, 27, 31-4,
37-8, 41, 64-6, 66-9, 81-2, 82-4,
85-7, 100-1, 122-4, 135, 136-8,
141, 143-50, 150-2
efficiency considerations 6-8, 10,
25-4, 90-3, 98-9, 100-1, 136-41,
141-3, 143-4, 144-7
equality of what? ix, x-xi, 1-4,
12-16, 19-21, 22-3, 130-1
equality, why? 3-4, 14-16, 16-19,
130-1
equality as an allegedlyempty
demand 23-6
fairness 18, 74, 75, 91, 148-51
famines 66, 72, 77-8
female-male ratio: biological and
social factors 92, 123-5, 127
focal variables 2, 3,12-15, 19-21,
23-6, 28, 50, 72, 74
freedom and its alleged
disadvantages 5-4, 59-64
freedom and well-being, possible
conflicts between 59-62
freedom as control 58, 64-6, 135
freedom, negative vi$-&-vi$ overall
41,64-6,87
freedom to achieve w, 2-3, 4-8, 27,
31-8, 39-42, 49-53, 54-5, 56-72,
81-4, 85-7, 100-1, 121, 129, 131,
132, 146-7, 148-9, 150-2
freedom nr-a-vw actual achievement
5, 32-3, 50-4, 62-4, 64-9, 135
functioning* 4-5, 7, 20, 39-42,44,
49-53, 54, 76-84, 87, 88, 90-3, 94,
100, 108-11, 121-2, 132, 135
fuzzy sets 47-9, 133
group inequality 117-18
illness and disease-proneness xii, 5,
27, 33, 37-8, 40, 65-6, 66-9, 81-2,
83-4, 100, 109-10, 113, 122,127,
142-3, 148
impartiality 16-18, 21, 23, 24, 131

206 Subject Index
incentives 10, 72, 105-6, 138, 139,
140, 141-3
income distribution x, 2, 9, 24,
28-30, 81, 88-9, 93-4, 95-101,
104-7, 109-12, 118, 121, 128, 132,
15!~2
incompleteness of judgements 5, 7,
46-9, 131-5, 143-4
incompleteness, fundamental
vis-a-vis pragmatic 133-4
informational basis of judgements 2,
23-4, 42-4, 73-5, 80-1, 129-31,
131-2, 135, 138-9
interpersonal comparisons of
welfare 6-7, 32, 82-6, 88, 93-5,
97-101, 136
interpersonal disagreements on
values 5, 46-9, 81-2, 83-4, 85-7,
108-9
interregional and intercountry
comparisons 125-8
intersectional valuation 46-8, 83-5,
132-3
intrahousehoid divisions 113-14,
121, 122-5
justice, theories of 2-4, 8-10, 73-87,
93-5, 143-7, 148-52
Kerala 126-8
liberty and libertarianism x, xi, 2, 3,
20-3, 25, 26, 32-3, 87, 122, 132,
138-9, 151
means of freedom, distinguished
from the extent of freedom xi,
26-8, 33-4, 36-8, 79-80, 81-4,
85-7, 143-7
measurement problems 5,46-9,
52-3, 64-9, 82-4, 105-4, 122,
131-5, 144
methodological problems ix-xii,
10-11, 17-19, 46-7, 48-9, 129-30
needs I, 2, 14-15, 107-9, 109-12,
120-1, 122
neutrality and the 'political'
conception of justice 75-9
old and the young xi, 82, 85, 91-2,
100,113-14,141
opportunity, equality of 2, 7-8,
26-8, 33, 34-6, 75, 79-85, 86, 118,
145-7
Pareto efficiency ('Pareto
optimality*) 136-7
partial orderings 45, 46-9, 70, 82-4,
85-4, 131-5, 144
plurality, internal 2, 131-5, 137
political incentives 72
poverty, the idea of 107-9, 110-12,
112-14, 116
poverty, distribution-sensitive
measures 103-7, 115-16
poverty, head-count and income-gap
measures 102-7
poverty as capability failure 108-9,
109-12, 112-14, 114-16, 134,
151-2
poverty evaluation 8-9, 44-5, 71-2,
81-3, 102-16, 134, 150-2
poverty in rich countries 9, 109-12,
114-16, 121-2, 123, 125, 151-2
primary goods 8-9, 19-20, 26-8, 33,
74, 78, 79-84, 85-7, 88-9, 91-2,
110, 121, 146-7, 148-9
race 114-15, 121-2, 145-7
Rawlsian theory of justice xi, 8-10,
12, 18, 26-8, 33-4, 38, 75-87, 91-2,
95, 145-9
real income comparisons 34-6
resources, equality of 8, 12, 19-20,
26-8, 35-8, 74, 80-1, 87, 121
responsibility 19, 145-7, 148-50
rhetoric of equality and uniformity
xi, 1, 30
rights x, 20-1, 25, 88, 119, 131, 132
risk and insurance 34, 140, 142, 143,
148-9
shortfalls and attainments 89-92,
123-4, 125, 144-5
social choice theory 32, 46-9, 67, 74,
77-8, 84, 89-93, 97-100, 103-7, 133
social welfare 29, 304, 32, 89, 94-5,
95-101
space, selection for evaluative
frameworks x-xi, 2-3, 19-21,
22-6, 27-8, 30, 42-4, 82-4, 88-9,
129, 130-2, 138-40
undernourishment 5, 7, 33-4, 37-8,
40, 81-2, 109-10, 111-12, 115-16,
122, 124-5

Subject Index 207
utilitarianism x, xi, 6-7, 13-14, 31-2, weighting and evaluation 42-3,
43, 73-5. 94-7, 99, 132, 137-8, 44-6, 46-8, 69-70, 80, 532-5, 144-5
149-50 welfarism 6-7, 43-4, 53-5, 73-5,
utilities 6-7, 25, 29, 35, 43-4, 53-5, 95-101, 149-50
73-4, 88, 94, 95-101, 131, 132, well-being ix, 28, 39-42, 43-4, 48-9,
136-7 49-53, 56-7, 59-~62, 62-9, 69-72,
80,88-9, 110, 129,132, 150-1
value objects 25, 42^», 58, 68-9, well-being freedom 5, 6-7, 49-53,
85-8, 88-9 59-62, 63-4, 65-9, 72, 111-12, 129,
138
weak equity axiom 15 women and gender inequality xiii, 7,
wealth xi, 1, 2, 19-20, 76-7, 81, 88, 9, 10, 85-6, 89-92, 100, 112-14,
118 122-5, 127-8, 142, 148, 149-50

Amartya Sen - Inequality Reexamined (Russell Sage Foundation Books) (1992).pdf

More Related Content

Similar to Amartya Sen - Inequality Reexamined (Russell Sage Foundation Books) (1992).pdf

Recently uploaded

Amartya Sen - Inequality Reexamined (Russell Sage Foundation Books) (1992).pdf