Jurisprudence

An International Journal of Legal and Political Thought

ISSN: 2040-3313 (Print) 2040-3321 (Online) Journal homepage: http://www.tandfonline.com/loi/rjpn20

The Conceptual Priority of Injustice

Eric Beerbohm

To cite this article: Eric Beerbohm (2014) The Conceptual Priority of Injustice, Jurisprudence,
5:2, 329-343, DOI: 10.5235/20403313.5.2.329

To link to this article: http://dx.doi.org/10.5235/20403313.5.2.329

Published online: 07 May 2015.

Submit your article to this journal

Article views: 22

View related articles

View Crossmark data

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=rjpn20

Download by: [University of Exeter] Date: 01 April 2016, At: 05:44

 DOI: http://dx.doi.org/10.5235/20403313.5.2.329 (2014) 5(2) Jurisprudence 329–343

REVIEW SYMPOSIUM

Amartya Sen, The Idea of Justice (2009)

Downloaded by [University of Exeter] at 05:44 01 April 2016

The Conceptual Priority of Injustice

Eric Beerbohm*

In ordinary political practice, deliberation about justice is workmanlike. You sense

that some institutional relation between us is off. Say our state has begun requir-
ing me to work, without ensuring that jobs are available. You may not even know
precisely what is objectionable about this relationship. When I ask you for work and
remind you that employment is a condition of my subsisting, what is wrong with my
position as supplicant? You form the conviction that this workfare policy is unjust
without the ability to offer a theory of exploitation. There is little need for you to
imagine a full package of welfare programs—let alone a complete set of political
institutions—that would satisfy the principles of a working theory of justice. Your
confidence in your conviction is not embarrassed by the fact that this conviction is
‘pre-theoretical’.
After all, you are engaging with justice as a decision problem. Should you work
to overturn this policy? What are the acceptable levers you might pull—whether
direct or through your place in the market—to challenge this law? From your point
of view, the task that political philosophers have assigned themselves can seem wil-
fully detached. Here’s a familiar way of framing it in distributive terms. We have
a set of resources that need to be parcelled out. Our constraints include a world
of scarcity and some minimal degree of individual self-interestedness. Among the
total universe of choices, what distributional profile would be most fair to each
member of society? This point applies with equal force to political philosophers
who resist conceiving the problem as essentially distributive. Libertarians have
posed the problem of justice as identifying a set of individual transactions—each
of which can be defended as a free, non-coerced site of choice. Notice again how
* Harvard University, USA.
 330 Jurisprudence

other-worldly such a program can sound. The task of describing a hypothetical his-
tory of giving and taking goods without duress appears no less removed than the
egalitarian’s version of the problem.
The gap between justice as a decision problem and as a theoretical target has
only grown in recent years. GA Cohen urged that we understand normative prin-
ciples as higher-order properties that float free from facts of material scarcity and
constraints of human nature. When we are comparing rules of regulation, we will
need to draw upon general facts about human psychology, economics and politi-
cal institutions. Even your conviction that a work requirement is morally dubious
appeared to rely upon some loose social-scientific assumptions. Neo-Platonism
rejects the thought that theorists of justice face an essentially interdisciplinary
task. For Cohen, the subject of justice looks nothing less than transcendental—it
is ‘not what we should do but what we should think, even when what we should
Downloaded by [University of Exeter] at 05:44 01 April 2016

think makes no practical difference’.1 What follows for the division of labour for
reflection on justice is significant. We set up a firewall between the problems of
implementation and pure questions of principle. Philosophers chip away at the
unadorned principles—recognising the requisite detachment from empirics—
while democratic practitioners take up strategic and tactical questions of how to
make institutions and social practices less unjust.
Amartya Sen has powerfully challenged this way of framing the theory and prac-
tice of justice. In The Idea of Justice he argues that leading theorists of justice have
been ‘transcendental’ in character, univocally devoting themselves to answering
the question: what would a completely just society look like?2 On his charge, this
approach is both unnecessary and insufficient for deliberating in a meaningful way
about how we are to overcome the most pressing injustices of our time. Here I will
take up these two claims in turn. The space-clearing aim of my paper rejects the
necessity view. I argue that Sen drives an implausibly stark wedge between transcen-
dental and comparative approaches. The task of selecting the lesser injustice does
not typically require that we have worked out the single, all-encompassing account
of justice. But it does make essential reference to some transcendental theory or,
more likely, a set of underformulated theories.
That our grasp of an ‘all-out’ theory can be seriously incomplete doesn’t
undermine the ambition to find covering principles. We need some minimal moral-
theoretic knowledge of justice—in its idealised form—to make pronouncements
about relative injustices. This objection plays a constructive role. For, once we clar-
ify the mutually supportive relationship between transcendental and comparative
theory, we are in a better position to see how subject-changing Sen’s challenge
is. I then draw out the implications of this methodological shift for individuals
and groups attempting to reform unjust structures. Deliberating about injustice
implicates particular epistemic and deliberative powers, and it is irreducibly posi-
tion-sensitive. We have reason to give explanatory priority to injustice because

1 GA Cohen, ‘Facts and Principles’ (2003) 31(3) Philosophy & Public Affairs 211.
2 Amartya Sen, ‘What Do We Want from a Theory of Justice?’ (2006) CIII (5) Journal of Philosophy 215;
Amartya Sen, The Idea of Justice (Harvard University Press, 2011). Page references to this work are
given in the text in parentheses.
 Review Symposium: Amartya Sen, ‘The Idea of Justice’ 331

decision problems call for direction absent in theories oriented around solving the
traditional theoretic problem of justice.

1. Field Guides vs User Manuals

I want to use this distinction to tease out two ways to treat a theory of justice. The
first sees it as a reference work. A theory of justice takes on a primarily evaluative
character, providing us with a way to benchmark a given institution against a set
of principles. It can guide our decision making, but it does so indirectly, by devel-
oping a reference point against which to compare actual institutions. Put in this
language, Sen wants to say that transcendental theories are seriously incomplete
field guides. Their single reference point—all-out moral acceptability—makes
Downloaded by [University of Exeter] at 05:44 01 April 2016

them ill-equipped to say something useful about the comparative moral odiousness
of two institutional schemes.
Contrast this with a second approach, which treats a theory of justice as a kind
of user’s manual. Here the stress is on the decision-guidance that the theory issues.
If you treat these instructions as reason-giving in your deliberation, you will likely
reduce the injustice of a practice. Sen introduces the metaphor of a handbook.
His worry is that all-out theories provide us with user instructions of a sort, but
their guidance is insensitive to the position of the prospective reformer: as a ‘grand
revolutionary’s one-shot handbook’ (100). Imagine a citizen who searches for guid-
ance in reforming a health care system. The principles found in theories of priority
or equality may prove unhelpful. Egalitarian theories offer principles that are so
demanding that nothing close to them can be feasibly adopted. The call to restruc-
ture one’s economic system as a whole may not be usable. A candidate reformer will
demand a handbook whose instructions are sensitive to the incremental nature of
her reforming aims.
It is tempting to collapse these two approaches. The theoretical project of test-
ing the significance of relative injustices cannot be divorced from the practical
project of selecting among unjust feasible option sets. My claim is not that these are
independent enterprises, but that they are usefully—though perhaps not neatly—
separable. Our health care reformer may have in hand a theory that tells her which
of two possible health care systems is less unjust. As a reference work, then, the the-
ory is successful. But that same theory may not be institutionally sensitive enough
to suggest to her what actions she needs to take to approach the more morally
acceptable option. This case suggests that these two approaches are in principle
separable.
Before I develop this distinction further, let me guard against a natural worry.
Talk of user’s manuals of justice invites the thought that one is operating with an
implausible conception of the aims and purposes of a normative theory. It is clearly
false to think of a theory of justice in anthropomorphic terms, issuing judgments
that its end-users struggle to follow. This kind of language takes philosophical lib-
erties. But even the milder language of ‘instructions’ paints a misleading picture
of practical reasoning with moral principles. Take the experience of following the
 332 Jurisprudence

instructions of a cookbook. An individual benighted about the gustatory properties

of various ingredients can succeed if he meticulously follows the serially ordered
directions in The Joy of Cooking. We don’t think of the ‘usability’ of principles of a
theory of justice in similar terms. There is nothing similarly automatic about the
advice that its principles offer. We should recognise that language like this risks
treating theories of justice as far more mechanically usable than desirable—or even
possible. So long as we recognise this hazard, I think this distinction can be illumi-
nating, and now I will attempt to show how.
There is a crucial difference between looking to a theory of justice as a reference
work and treating it as an instructional aid. The person reasoning about Hittite and
Greek slavery is consulting a theory of justice for a theoretical insight. In contrast,
the reformer stands in an importantly different relation to the theory of justice. She
expects it to help her act in the world. Sometimes we relate to theories of the physi-
Downloaded by [University of Exeter] at 05:44 01 April 2016

cal world in analogous ways. I might leaf through a textbook on theoretical physics,
searching for some way to explain an observed natural phenomenon that puzzles
me. If, however, I find myself trapped in a lunar orbiter, subject to dangerous tem-
peratures, that same textbook may be consulted. But my attempt to extract from
it usable information in the face of an emergency is a very different kind of enter-
prise. Now I am only concerned with acquiring the bare knowledge to successfully
perform some task. While these seem to me to be two distinct ways of making use
of a theory of justice, they are not independent approaches. The reformer can also
use her adopted theory of justice as a reference work. Suppose, after she attempts
to refigure an objectionable practice, she looks back at her series of reforming
actions. Now she can use the theory as a purely evaluative tool—to ‘bear her own
survey’, in Hume’s phrase.

2. Full-Coverage Theories

We might characterise the normative family of transcendental theories by their

aims or their deficits. Start with the positive condition. Transcendental theories are
idealising—they set out to identify principles that, if satisfied by a social order, would
license a fully just arrangement. Their promise is of full coverage. They owe us a
verdict on any problem within their space—real or imagined. If Sen is right, they
are deficient both as reference and as instructional tools. For theoretical reason-
ers, pondering the moral wrongness of competing injustices, they are insufficiently
responsive. For practical reasoners, seeking principles that issue guidance under
unjust background conditions, they are, in Sen’s words, ‘silent’—or at the very
least, they do not offer the kind of guidance that a practitioner of justice may rea-
sonably demand. They aim to deliver principles that comprehensively issue verdicts
under any situation of choice.
Full-coverage principles must be capable of evaluating any problem of justice
that we can imagine—however divorced that problem is from the world we live in.
We are in need of a theoretically satisfying answer to Louis, an imagined charac-
ter of egalitarians, who has deliberatively cultivated expensive tastes in claret and
 Review Symposium: Amartya Sen, ‘The Idea of Justice’ 333

plover’s eggs and now demands compensation from the state.3 Transcendentalists
will not be moved by the rejoinder that this thought experiment is unreal or its
resolution is lacking moral urgency. By their own lights, they need an answer to any
distributive problem, regardless of its distance from actual political practice. The
positive aim of transcendental theories, then, is to avoid the spectre of intuitionism
by providing a theory that yields a set of highly deterministic verdicts.4
The other way to demarcate transcendental theories is by noticing their built-in
limitations. Sen often characterises theories with a ‘totalist’ ambition by reference
to their deficiencies. Their encompassing ambition leaves them subject to two
kinds of indeterminacies in flawed environments. We can call the first disability:

Evaluative Indeterminacy. Transcendental theories fail to structure our theoreti-

cal deliberation about the gravity of relative injustices.
Downloaded by [University of Exeter] at 05:44 01 April 2016

To see this first shortcoming, suppose I engage in a purely evaluative exercise. I

want to know the relative moral wrongness of slaveholding by the Hittites and the
Greeks. Which of these two social arrangements was more offensive to principles of
justice? Unless I accept a strong thesis of ‘the relativism of distance’5 —of the kind
that Bernard Williams defended—I can pursue this line of inquiry knowing that
its results are not going to bear upon my own deliberative situation. It is, I think,
a perfectly intelligible question. The deficiency is that transcendental theories can
fail to issue any verdict when presented with two rival unjust practices. Notice that
this kind of failure comes in the assessment stage. I am in the market for a piece
of moral knowledge, not practical guidance, and yet I may find no answer when I
deliberate within the system of my favourite transcendental theory of justice. We
can distinguish this deficiency from:

Decision-Guiding Indeterminacy. Transcendental theories fail to structure our

practical deliberation in acting to reform unjust practices and institutions.

Transcendentalists can find themselves deadlocked when they attempt to figure out
where their moral priorities lie under background injustices. Consider the project
of trading-off injustices. A Rawlsian may find herself torn between countenancing a
procedural or substantive infraction. For instance, you may have to choose between
fostering a democratic procedure that ensures greater power for the least well-
off representative person or material gains to that same person. A theory whose
principles simply warn us to avoid both kinds of infractions may not be satisfying.
I find the case of trading-off injustice compelling. But the Rawlsian could well find
that the principles suffer from what Sen calls a ‘relational silence’—they routinely
propose a single institutional jump that ignores the difficult choices involved in
the thorny transition to more a more just state of affairs. Counterexamples like this
are evidence that Sen has identified a deficiency in accounts of distributive justice.

3 Ronald Dworkin, Sovereign Virtue: The Theory and Practice of Equality (Harvard University Press, 2000).
4 Henry Sidgwick, The Methods of Ethics (Hackett, 7th edn 1981 [1907]).
5 Bernard Arthur Owen Williams, Ethics and the Limits of Philosophy (Harvard University Press, 1985).
 334 Jurisprudence

These two deficits are useful in setting up Sen’s alternative approach. Compara-
tive theories of justice aim to avoid these two kinds of indeterminacies. For Sen,
comparativists do not share the idealising project, since they do not seek the ‘iden-
tification of a perfectly just society’.6 But this way of distinguishing the approaches
has the potential to mislead. For one thing, it is not clear that theorists from the
transcendentalist tradition would accept this formulation. To see its peculiarity, con-
sider the analogue in moral philosophy. Suppose you set out to identify the morally
perfect person. Unless you happen to believe in a robustly determinate, monistic
theory of value, such a search would be quixotic. For on most plausible-sounding
approaches to morality, there is no such thing as the morally perfect person. To be
sure, there is a set of principles that, if flawless adhered to, would be constitutive
of a morally exemplary life. Unless the principles are highly restrictive, there are
presumably many ways of living that can succeed in satisfying these principles.
Downloaded by [University of Exeter] at 05:44 01 April 2016

This line of argument is amendable to thinking about theories of justice as

well. It isn’t obvious that transcendental approaches have devoted themselves to
identifying a perfectly just society—at least in the plain sense that this formulation
implies. I take it that such an imaginative exercise would involve describing the
institutional details of an idealised political system, writ large and small. An enter-
prise like this may be philosophically interesting, akin to building a model train-set
or a dollhouse. Plato’s efforts at sketching the panoramas that we know as Kallipolis
and Magnesia may come close to approximating a model-theoretic approach. In
the Laws we learn about the Nocturnal Council’s unusual meeting hours and the
differential tax rates paid by citizens.7 Contemporary theories of justice, on both
the left and right, don’t understand themselves as providing an intricate descrip-
tion of what an ideal just society would look like. If anything, they are marked by
an excessive modesty in what verdicts principles of justice can provide in complex
institutional settings.8 What they identify is a set of principles that may be satisfied
through a range of institutional designs. They are regularly agnostic about how this
range is to be drawn. Robert Nozick, conscious that he hasn’t provided a libertar-
ian theory of rectification, famously refuses to rule out a system of redistribution.
John Rawls stresses his agnosticism about economic arrangements. Neither claims
to have spelled out what the perfectly just society would look like.
Second, the formulation of a ‘perfectly just society’ has the potential to mislead.
The phrase is used to pick out the site of transcendental theories of justice. The
worry is that it conflates a just set of institutions from a broader moral evaluation
that we reserve for societies as a whole. Granted, some Rawlsians have acknowledged
that possessing a just basic structure does not exhaust the range of objects that are
susceptible to judgements of justice. Susan Moller Okin pioneered this expansion
of the scope of distributive justice, and GA Cohen powerfully developed this cri-

6 Ibid, 225.
7 Plato, Compete Works, John M Cooper (ed) (Hackett, 1997).
8 ‘Thus on many questions of social and economic policy we must fall back upon a notion of quasi-
pure procedural justice: laws and policies are just provided that they lie within the allowed range,
and the legislature, in ways authorized by a just constitution, has in fact enacted them’: John Rawls,
A Theory of Justice (Harvard University Press, rev edn 1999) 362.
 Review Symposium: Amartya Sen, ‘The Idea of Justice’ 335

tique.9 But those who wish to distinguish between a just basic structure and what
we might call cosmic justice will find Sen’s expansive marker problematic. They are
in the position to deny that they are formulating principles that, if satisfied, would
indicate a just society. For them, it is conceivable that we could have a fully just set
of ground rules alongside a series of injustices in what TM Scanlon calls ‘informal
politics’, where inequalities of power can be sustained against background fairness
of formal institutions.10 I should acknowledge that this second point is, as far as I
can tell, relatively minor. It requires tidying up the stated aim of transcendentalists.

3. The Necessity Claim

What is the relationship between transcendental and comparative theory? Is the

Downloaded by [University of Exeter] at 05:44 01 April 2016

latter necessary for the former? If not, in a world as seriously imperfect as our
own, why not postpone transcendental theorising for the indefinite future? How
we answer this question is important for theorists of justice. While Sen comes close
to concluding that transcendental approaches are pointless, my focus will remain
on his official position. The necessity claim holds that transcendental views do not
stand in any essential relation to comparative approaches. My strategy is to cast
doubt on the thesis in two ways. First, I will claim that the principal argument for it
is unsound. Second, I will suggest the difficulty of resisting a certain kind of neces-
sary relation between transcendental and comparative approaches. If this view is
correct, then the argument is also invalid. Sen arrives at this position through this
line of reasoning:

The Practitioner’s Argument:

Premise 1. We live in a world so unjust that our most urgent problems involve
comparative choices among injustices.
Premise 2. Selecting among comparative injustices doesn’t (typically) require
us to be in possession of the single, correct theory of justice.

So, it is unnecessary for us to direct our theoretical energy at identifying the princi-
ples that a perfectly just society would satisfy.
I began this paper by acknowledging the plausibility of Premise 1. We don’t find
ourselves gathering together, de novo, aiming to construct institutions that are fully
just. So let us grant Premise 1 and focus our attention on Premise 2. In its defence,
Sen offers two arguments by analogy. The first begins with this premise:

Aesthetic Analogy. One can compare the relative beauty of the works of Picasso
over Dalí without identifying the most beautiful picture in the world.

9 Susan Moller Okin, Justice, Gender, and the Family (Basic Books, 1991); GA Cohen, ‘Where the Action
Is: On the Site of Distributive Justice’ (1997) 26(1) Philosophy & Public Affairs 3.
10 Thomas Scanlon, The Difficulty of Tolerance (Cambridge University Press, 2003).
 336 Jurisprudence

Folk aesthetics accepts this claim as correct, and arguendo we can follow the lead of
common sense. To start with a thought experiment, imagine a renowned scholar of
art. Mary has devoted her life to the study of oil-based canvas painting. But there’s
a twist. At every stage her instructors and colleagues have kept her from being
exposed to the Mona Lisa.11 Now even if we suppose that da Vinci’s painting is
the aesthetically best, and Mary lacks all contact with that piece, it doesn’t neces-
sarily disqualify her from making comparative judgements among lesser works. It
is worth noting that examples of this kind don’t demonstrate the independence of
judgements about the best and judgements about the better. We could imagine
that, later in her career, Mary accidentally comes across a textbook that contains
the Mona Lisa. Identifying this painting as the best—in light of the principles of aes-
thetics that she has developed over the years—forces a certain amount of revision
in her theory of the beautiful. It could lead her to reverse her previous judgement
Downloaded by [University of Exeter] at 05:44 01 April 2016

between Picasso and Dalí. So even if we accept the Aesthetic Analogy and its rel-
evance to theorising about justice, this argument doesn’t yield a conclusion strong
enough to support the necessity thesis.
Well aware of the difficulties of this analogy, Sen repairs to a second argument
of the same basic form:

Distance Analogy. One can compare the heights of Mount Kanchenjunga and
mont Blanc without being about to pick out the tallest mountain in the world.

As an epistemic claim, this is surely right. Knowing about the superlative height
of Everest—even knowing that such a mountain of that height exists—would be
unnecessary for making this comparative judgement. The general principle that
Sen cites follows from reflection on examples like this one: we can make relative
judgements without knowing the object in the world that most closely approxi-
mates the relevant metric, whether that be height, weight or beauty.
Yet this analogy fails to lend support to the necessity thesis. How might an anal-
ogy with theories of justice be advanced? Suppose I am attempting to determine
which of two regimes, say North Korea and Iran, is more unjust. It is plausible to
think that I can form this judgement without canvassing all of the world’s nations
and picking out the regimes that most and least comply with the principles of jus-
tice that I take be to correct. I don’t need to know the ‘supreme alternative’, in
Sen’s parlance, to render a judgement about this more local judgement. Once we
state the analogy in this way, however, its limitations become clear. Now Sen wants to
argue that one need not engage in transcendental theorising about justice in order
to make comparative judgements. The upshot of his analogy, however, suggests a
different conclusion. There are two possible kinds of ignorance at work here: (1)
I may not know what actual regime comes closest to approximating justice; (2) I

11 This thought experiment bears a relation to Mary’s Room. In that case, Frank Jackson asks us to
imagine Mary, an accomplished neurophysiologist who specialises in colour. Mary has been raised in
a black and white environment. Does she know everything there is to know about the colour red, or
when exposed to a colour television does Mary learn something new? See Frank Jackson, ‘What Mary
Didn’t Know’ (1986) 83(5) Journal of Philosophy 291.
 Review Symposium: Amartya Sen, ‘The Idea of Justice’ 337

may not know what moral principles are constitutive of a perfectly just regime. The
problem is that Sen’s analogy doesn’t speak to the latter kind of ignorance. But the
necessity thesis is a claim about the relationship between two approaches to justice,
not to knowledge of which worldly practices most embody a theory of justice.
Even if the necessity claim lacks a sound argument in its favour, it is valuable
to pause and investigate its surface appeal. What is intuitively attractive about the
thought that we can engage in comparative judgements of injustices without any
appeal to transcendental principles? Let me suggest two sources of appeal. The first
is that transcendentalists operate at a nosebleed level of ideality. This worry is as
old as Glaucon’s needling of Socrates. And since Rawls introduced the distinction
between ideal and nonideal theory, critics have raised doubts about the practical
value of investigating principles for a world of universally compliant individuals.12
We can reply directly to this concern in a moment—I think it is best bundled with
Downloaded by [University of Exeter] at 05:44 01 April 2016

the second worry and handled together.

Let me sound a concessionary note. No plausible normative inquiry can afford
to be indifferent to the charge of over-idealisation. I suspect that every theorist of
justice will accept that there is some point at which excessive idealisation makes
one’s theory troublingly distant from our world. To see this, imagine a political
theorist who has devoted her life’s work to constructing a moral theory for an imag-
ined people who were physically and psychologically incapable of suffering. Even if,
as her career wanes, she arrives at principles that survive the scrutiny of soundness
and validity, there is a remaining question that all moral inquiry is susceptible to:
what is the point of principles designed for individuals who so fail to resemble the
vulnerable creatures that we understand ourselves to be? If the necessity thesis is
correct, the task of arriving at principles that a fully just citizenry would select may
appear to be a project nearly as valueless as our imagined theorist, whose work is
designed for creatures wholly different than us. Perhaps, then, Sen’s critique draws
its plausibility from concerns about ideality.
The second source of the thesis’s appeal derives from a positive impulse—what
we might call a need-to-know practicality. An incompletely worked-out theory may
deliver plausible results in the set of worlds it is likely to be applied in. Even without
a more fine-grained, exhaustive formulation, a rougher version may provide all the
guidance we will need. You know well that Newtonian physics is incomplete. But so
long as the velocity of the objects around us remains below a certain point, it per-
forms admirably in our daily lives. It is more than adequate to predict the path of a
golf ball and to launch a lunar explorer to Mars. Perhaps we have reached a similar
stage in our collected theories of justice. Available accounts are incomplete—even
demonstratively incorrect at the margins—but they are usable, given the environ-
ment they will be deployed in.13 To make vivid this thought, consider this pair of
cases:

12 John Rawls, A Theory of Justice (Belknap Press, 1971).

13 I argue for this in In Our Name: The Ethics of Democracy (Princeton University Press, 2012).
 338 Jurisprudence

The Partial Capability Approach: It’s 1979, and Amartya Sen has just introduced
basic capability equality in his Tanner Lecture.14 We have reason to pay attention
to what a person can do rather than what a person does.

The Full Capability Approach: It’s 1985, and Amartya Sen formalises his capability
framework. Our capabilities are a ‘set of functioning vectors within his or her
reach’.15

Here we have two accounts of capabilities, not all-out distributive justice. But my
point about incompleteness carries over. An individual criticising an unjust struc-
ture for discriminating against individuals with disabilities may not need the fuller
approach. She will do fine in 1979. Telescoping from this case, we can imagine a
range of decision problems that would only require the less formal characterisation
Downloaded by [University of Exeter] at 05:44 01 April 2016

of capabilities. Unjust environments have a way of seriously constraining a reform-

er’s political options. So it may be precisely in those settings that less guidance is
due. Here the choices may be between the level of strategies and tactics of reform.
If our task is to sketch precisely what institutions a completely society may justifiably
contain, we will need a currency of justice that has greater content. We would do
better to wait until the 1985 formalisation of capabilities. You can be sympathetic
to this line of thought without embracing the Necessity Thesis. Our political theo-
ries, like our theories of the universe, are always going to be works-in-progress. We
shouldn’t postpone applying them until we have arrived at what we take to be a
completed theory. I believe that Premise 2 draws its plausibility from the thought
that incompleteness is not necessarily a barrier to deliberating about the compara-
tive injustice of actual institutions.
Even if we accept Premises 1 and 2, the conclusion of the Practitioner’s Argu-
ment does not yet follow. There remains, I think, a necessary relation between
transcendental and comparative approaches. To determine that Iran is a more
unjust regime than North Korea, surely I need to know something about the princi-
ples that would signify a perfectly just social arrangement. Perhaps I’m torn between
a subset of egalitarian or libertarian views, but the transcendental ambition of these
views remains. The Necessity Thesis is committed to denying this relation. I find it
difficult to see how a comparative theorist of justice could make no use of principles
that state the necessary conditions that must be satisfied for an institution or prac-
tice to be morally acceptable. In this vein, Michael Sandel explains the inevitably of
theory in this way: ‘For all we may resist such ultimate questions as the meaning of
justice and the nature of the good life, what we cannot escape is that we live some
answer to these questions — we live some theory — all the time.’16 It is clear that
ultimate questions of this ilk would qualify as transcendental in their ambition.

14 Amartya Sen, ‘Equality of What?’ in The Tanner Lectures in Human Values, vol 1 (Cambridge University
Press, 1980) 217.
15 Amartya Sen, ‘Well-Being, Agency and Freedom’ (1985) 82(4) Journal of Philosophy 201.
16 Michael Sandel, Democracy’s Discontent: America in Search of a Public Philosophy (Harvard University
Press, 1996).
 Review Symposium: Amartya Sen, ‘The Idea of Justice’ 339

What Sandel is claiming is that there is an inevitable appeal to such principles.

What needs clarification is his remark that ‘we live some theory’. The temptation is
to infer that you and I are implicitly living a highly worked-out account of justice.
This is unlikely to be the case. What is more plausible is that you and I are operat-
ing under a second-order theory that underdetermines a set of first-order theories.
We saw this above, when the informal capabilities approach offered us an incom-
plete theory. Comparative judgements parasitically depend on principles that have
a transcendental aim—even when they are underformulated. In the next section I
will suggest that these two approaches synergistically relate.

4. Serviceable Principles
Downloaded by [University of Exeter] at 05:44 01 April 2016

Even if the necessity claim is false, the sufficiency claim is field-disrupting. It is a

call for us to redirect our philosophical energies to developing serviceable princi-
ples for responding to injustice.17 In formulating his project in this way, I draw an
analogy from Robert Nozick’s picture of serviceable truths. These are beliefs ‘true
enough’ that they achieve our aims when acted upon, given our specified environ-
ment.18 Sen’s home truth is that we lack serviceable principles. We need to assign
a kind of priority to injustice that has not been the focus of debates in distributive
justice—whether domestic or globalised. This home truth, I believe, has been over-
shadowed by Sen’s criticisms of over-idealised theory. There is an irony here. All-out
theories, in their abstraction and idealisation, engage in ‘a search for the supreme
alternative among all possible alternatives’ (95). But they fall short on their own
terms, failing to provide coverage in perfectly familiar conditions of injustice. Even
if we arrived at fully worked-out principles in a clean-room environment, we would
not know everything there is to know about justice. Sen puts it well in his article
precursor to his book:

If such spotless justice were the only focus of attention in a theory of justice, then the
institutional preconditions would form a kind of ‘entry barrier’, leading to an absti-
nence from applying justice theory to situations in which those exacting institutional
demands are not only currently met but cannot be met in the foreseeable future.19

What is insufficient about theories of justice that work from a background of ideal-
ised conditions is that they are incapable of fulfilling—at least on their own—both
roles that we assign to theories of justice. Despite their success as reference works,
they fall short when consulted as instructional materials. Our division of labour
should reflect the different purposes of a theory of justice. Earlier we imagined a
range of dilemmatic situations where full-compliance principles of justice weren’t

17 Sen puts the change in subject this way: ‘A transcendental approach to justice gives—or can give—
quite distinct and distant from the type of concerns that engage people in discussions on justice and
injustice in the world’: The Idea of Justice, 96.
18 Robert Nozick, The Nature of Rationality (Princeton University Press, 1993) 68.
19 Sen, ‘What Do We Want’ (n 2) 226.
 340 Jurisprudence

happily ‘ported’ to real-world institutional choices. A political theory that single-

mindedly devotes itself to producing transcendental principles will be susceptible
to the charge that it neglects a principal motivation of theoretical reflection about
acceptable social institutions: to serve not as a revolutionary’s handbook, but as a
more quotidian reformer’s manual.
Rather than assess Sen’s social choice approach, in my limited remaining space I
want to imagine two broad agendas that give priority—investigative or conceptual—
to injustice in our assessment of political institutions. The more modest version is
what I am calling the investigative priority of injustice. We should significantly redirect
our resources to developing serviceable principles tailored to decision problems
that we face today. What we produce will not hang together in the traditional sense
of a theory. Our serviceable principles, considered together, may not be especially
simple to state. There may be many more of them than their transcendental rela-
Downloaded by [University of Exeter] at 05:44 01 April 2016

tives. And they will be seriously limited in their ability to provide verdicts across
environments. In any case, the range of theories of distributive justice can seem
deadlocked. David Lewis liked to say that a deadlock is probative evidence that sub-
ject matter is difficult, not intractable. Nothing follows about whether there is a fact
of the matter about which rival principles are correct. The disagreements between
equality, priority and sufficiency theorists are deep and non-trivial. Whether these
theorists are climbing the same mountain, they are not aiming for the same peak—
to take Mill’s reinvigorated phrase.
If we assign investigative priority to injustice, we are shifting the subject way
from first-order distributive theory to the second-order question of the political
activist: what should individuals do with principles of justice in projects of reform?
The motivation can be put simply. The ethics of identifying and reforming unjust
institutions is not exhausted by a substantive theory of justice. And the mode of
ethical reasoning of the reformer is radically different from the theory-designer.
Political philosophies sometimes talk of ‘applying’ principles to actual institutions,
but political actors don’t see themselves as appliers. The function of principles of
justice in their day-in-day-out decision making is hardly clear. The test of service-
able principles is in their use. Are they effective at helping us identify and reform
injustices on the ground that may be occluded by our position within complex
institutional and social networks?
Notice the insistent focus on how things are for flesh-and-blood agents, who
have to figure out how to act within a system of diversely motivated non-compli-
ers. The perspective of the political reformer is not analogous to the institutional
designer. The way this ‘outsider’ deliberates about the moral acceptability of
institutions—whether the Boy Scouts of America or the United Nations—will be
institutionally one step removed. The prospective reformer wonders how to act in
response to the moral and rational failings of others precisely when one lacks direct
access to legislative action. This focus should not come at the expense of questions
of institutional design. How individuals should understand their reforming obliga-
tions is a hybrid of interpersonal and institutional concerns.
To see the significance of this subject-change, compare two pictures of the
reformer. The first is self-consciously realising an ideal of justice in her actions.
 Review Symposium: Amartya Sen, ‘The Idea of Justice’ 341

‘Even most nonphilosophers who are active in the cause of justice’, writes A John
Simmons, ‘do in fact have in mind, however vaguely, an ideal of justice toward
which they take their campaigns to be ultimately directed.’20 We are all transcen-
dentalists, the thought goes. Some of us face greater time pressure than others,
and some have more theoretical materials at our grasp, but there is an essential
continuity between theory and practice. If we accept the sufficiency claim, a very
different picture of the reformer emerges. She needn’t be chagrined by her lack of
an ideal of justice. She retains confidence in her convictions about the presence of
practices that subordinate and subject some for the good of others. Her decision
making is not a result of calculation or the subtle weighing of positive principles of
justice, however unformed.
For the task of identifying an unjust institution is radically different from con-
structing a theory of justice. For one thing, the repertoire of skills that allow us to
Downloaded by [University of Exeter] at 05:44 01 April 2016

pick out an injustice seems considerably broader. It will not do to have a familiarity
with the reflective equilibrium and bare training in first-order logic. The demanded
skills of practical reasoning will be a mix of cognitive and affective. We will need
to read and respond to the basic elements of the moral world—to draw from Bar-
bara Herman’s work on what moral reasoning is really like for the deliberator.21
Consider the imaginative act of engaging vividly with the particulars of poverty. If
we have little first-hand acquaintance with deprivation, it may be difficult to vividly
enter into particulars—the hectoring calls from creditors and the experience of
having one’s food stamp allowance cut. Increasing our vividness of a person’s or
group’s position can improve our deliberative position, for status quo institutions
can deprive us of appreciating moral claims as seriously as we might.
These examples point to the gap between constructing justice and identifying
injustice. The theoretical enterprise of theory design is a fairly routinised activity.
We test an abstract principle—say that the moral value of achieving a benefit for
an individual depends on their level of well-being—by bringing it to bear on a
high-level description of a distributive pattern. If the result flies in the face of an
intuitive judgement, we modify the initial principle. Notice how little resemblance
this mode of reasoning bears to the task of identifying an injustice. This mode
of deliberation involves trying to assess how weighty various considerations are. It
seems especially amenable to heuristics like vividness and affective engagement
with the interests one is assessing. Individuals who have first-hand experience with
disadvantage may have a certain epistemic edge in this mode of reflection. For oth-
ers, a certain amount of meta-reasoning about one’s current deliberative position
may be an essential anecdote. Is my place within this institution obscuring features
that are salient for the identification of an injustice?
Now let’s turn to a second way to think of the priority of injustice, one that is
philosophically more ambitious. Suppose we assign conceptual priority of injustice over
justice. This move can sound mysterious until we look for analogous efforts across
philosophy. Some leading epistemologists have argued that knowledge has concep-

20 A John Simmons, ‘Ideal and Nonideal Theory’ (2010) 38(1) Philosophy & Public Affairs 36.
21 Babara Herman, Moral Literacy (Harvard University Press, 2008).
 342 Jurisprudence

tual priority over belief. We get the concept of belief as ‘something derived from
the concept of knowledge’.22 Timothy Williamson has forcefully challenged the
tempting thought that belief is conceptually prior to knowledge. So, too, we may
conclude that Sen’s subject-change is as much theoretical as it is practical. Taken
seriously, it urges us to treat the very concept of injustice as more basic than justice,
so that the latter is derived from the former. I take this methodological shift to be
inspired by Sen’s approach rather than a statement of his official position. You
might worry that taking his thought in this direction makes him more of a grand
revolutionary than he would accept.
The basic idea—which I will describe and not defend—is that we have reason
to ask what properties make an institutional arrangement unjust independently
from and prior to attempts to characterise a just arrangement. We have seen that
theories of distributive justice typically proceed in two stages. They first investigate
Downloaded by [University of Exeter] at 05:44 01 April 2016

the moral properties of a perfectly just system—made under a series of simplify-

ing assumptions. They then describe how their principles operate when idealising
assumptions are not satisfied. If we endorse the conceptual priority of injustice, this
is not a fruitful way to proceed. We should change our order of inquiry, investigat-
ing the moral unjustifiability of structures where we stand in unjustifiable relations.
There is an emerging literature in philosophical ethics that lends support to this
relational approach to inequality. If we think morality has a fundamentally ‘deontic
structure’, as Jay Wallace has compellingly suggested, then valuable relationships
lie at its core. On this approach, what explains the non-discretionary character of
moral reasons is that they describe a special relation that we have with others whom
we might wrong. Justification is interpersonal. The person adversely affected by my
action has a privileged basis for moral complaint against me. To wrong another is
to stand in a relation of mistreatment.
We can preserve the deontic character of our thinking about injustice by begin-
ning with the same question raised by relational accounts of morality: What is it to
wrong someone? The answer may be that injustice—whether interpersonal or insti-
tutional—is based on a dyadic moral relation among people. Perhaps what gives
injustice conceptual priority is that its absence characterises justice. In philosophi-
cal ethics, there is nothing surprising about this approach. Much more needs to be
said about what counts as a relationship and whether this approach can capture
paradigmatic cases of systematic injustice. I hope that I’ve said enough to suggest
the value of this line of thinking for the assessment of political institutions—work-
ing in the negative space of theories of injustice. Whether treating injustice as con-
ceptually prior to justice will be fruitful is an open question.
Assigning conceptual priority to injustice may help to clarify the position-
sensitivity of principles for responding to injustice. Earlier I described serviceable
principles for reformers as tailored to particular habitats of injustice. But there
is another way in which they can be location-sensitive. The way you stand in rela-
tion to an institutional injustice may bear strongly on how you should respond to
it. Whether you are a victim of that injustice, a bystander or a direct beneficiary

22 Timothy Williamson, Knowledge and its Limits (Oxford University Press, 2000) 2.
 Review Symposium: Amartya Sen, ‘The Idea of Justice’ 343

can gravely matter vis-à-vis your obligations towards an institution. Sen’s remark-
able work on position-sensitivity is a paradigm of the kind of inquiry that follows
from taking injustice as conceptually prior. For him, ‘the role of positionality may
be particularly crucial in interpreting systematic and persistent illusions that can
significantly influence—and distort—social understanding and the assessment of
public affairs’ (168). All-out theories of justice can be utopian precisely because
they are indifferent to the position of the individual towards a discrete injustice.
Their guidance is naïve in its failing to take seriously the reforming capacities of
the particular individual. I have suggested that our ethical thinking about injus-
tice should be ‘location-aware’ in two ways. It should take seriously the individual
reformer’s environment and their own bearing towards a candidate injustice.
In a world where we have considerable confidence that chronic injustices
persist, devoting serious attention to completing an all-out theory of justice can
Downloaded by [University of Exeter] at 05:44 01 April 2016

seem politically unnecessary and philosophically self-indulgent. For on-the-ground

agents, devoting scarce resources to answering questions of this kind may well be
counter-productive. We can make incremental moral progress quite independently
from its precise content and contours. I have argued that once we conceive of
transcendental and comparative theorising as mutually supportive, we can see the
full force of Sen’s distinctive approach to political theory. In his dedication to John
Rawls in Equality and Partiality, Thomas Nagel thanks him for changing the subject.
We owe the same to Sen’s The Idea of Justice.