Any topic (writer’s choice)

Reflection Essay:
John Rawls vs. Robert Nozick

It is clear that there are many approaches to studying ethics, concepts of justice, fairness, equity, and morality.  Few thinkers have had a more significant impact on contemporary politics and discourse than John Rawls and Robert Nozick.  Your reading mentions Rawls, but omits Nozick and I would like for you to take some time to read two essays (available in  the Power Points and other files tab) written by these scholars. You are likely to find some of this reading difficult.  Rawls is not an easy person to read because he tends to build long arguments that may seem like non sequiturs, but if you follow through you will realize his points are quite succinct.  Nozick is a little more direct but still requires careful attention. Needless to say, this assignment is not something most people can do the day it is due.

Your job is to write a reflection essay that is 5 full double-spaced pages in length, which does NOT included the Abstract, Cover page, or your References page. Your essay must include the following:

A brief summary of both authors’ major arguments and key points.
An explanation of where the authors disagree and where they may potentially agree.
Your conclusion as to which author provides a more workable framework for practical application to social problems and criminal justice.

Distributive Justice
Author(s): Robert Nozick
Source: Philosophy & Public Affairs, Vol. 3, No. 1 (Autumn, 1973), pp. 45-126
Published by: Wiley
Stable URL: http://www.jstor.org/stable/2264891 .
Accessed: 20/02/2014 16:22
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
Princeton University Press and Wiley are collaborating with JSTOR to digitize, preserve and extend access to
Philosophy &Public Affairs.
http://www.jstor.org
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
ROBERT NOZICK Distributive Justice
The term “distributive justice” is not a neutral one. Hearing the term
“distribution,” most people presume that some thing or mechanism
uses some principle or criterion to give out a supply of things. Into
this process of distributing shares some error may have crept. So it is
an open question, at least, whether redistribution should take place;
“Distributive Justice” from Anarchy, State and Utopia by Robert Nozick appears
by permission of the author and Basic Books, Inc., Publishers. Anarchy, State,
and Utopia will be published in Spring I974 ? Robert Nozick.
The essay here differs only slightly from Chapter 7 of the forthcoming book.
Apart from later stylistic revisions and the addition of several minor defenses
against possible objections, it is identical with the chapter of the manuscript written while I was a fellow at the Center for Advanced Study in the Behavioral
Sciences during 1971-1972. I am very grateful for the Center’s support. It will be
helpful to the reader of this essay to know its place in the book. The book’s central
concerns are the consequences for political philosophy of a far-reaching theory
of individual rights; in particular, the question of what activities, if any, may
be performed by the state or its agents without violating these rights. The first
half of the book, “State of Nature Theory, or How to Back into a State Without
Really Trying,” sets out the structure of these individual rights, and argues that
a minimal state would arise from anarchy even though no one intended or tried
to bring about that result, and argues that the “invisible-hand process” by which
the state would arise need not violate anyone’s rights. The present essay opens
the second part of the book, and functions there to rebut the possible claim that
a state more extensive than the minimal one justified in the first half would be
necessary or appropriate in order to achieve distributive justice.
This essay can stand alone. But it does not stand as solidly, I think, without
the material of the book’s first half that underlies it, the later material of the
second part that buttresses it by critically examining other reasons which purport to justify a state more extensive than the minimal one, and the book’s
last part on utopia and utopian theorizing, whose abstract model and whose
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
46 Philosophy & Public Affairs
whether we should do again what has already been done once, though
poorly. However, we are not in the position of children who have been
given portions of pie by someone who now makes last minute adjustments to rectify careless cutting. There is no central distribution, no
person or group entitled to control all the resources, (jointly) deciding
how they are to be doled out. What each person gets, he gets from
others who give to him in exchange for something, or as a gift. In
a free society, diverse persons control different resources, and new
holdings arise out of the voluntary exchanges and actions of persons.
There is no more a distributing or distribution of shares than there is
a distributing of mates in a society in which persons choose whom
they shall marry. The total result is the product of many individual
decisions which the different individuals involved are entitled to make.
Some uses of the term “distribution,” it is true, do not imply a previous
distributing appropriately judged by some criterion (e.g., “probability
distribution”); nevertheless, despite the title of this essay, it would
be best to use a terminology that clearly is neutral. We shall speak of
people’s holdings; a principle of justice in holdings describes (part
of) what justice tells us (requires) about holdings. I shall state first
what I take to be the correct view about justice in holdings, and then
turn to the discussion of alternative views.’
I. THE ENTITLEMENT THEORY
The subject of justice in holdings consists of three major topics. The
first is the original acquisition of holdings, the appropriation of unheld
things. This includes the issues of how unheld things may come to be
held, the process(es) by which unheld things may come to be held,
the things that may come to be held by these processes, the extent
discussion of filter devices, intertwine with it. I direct these remarks, of course,
especially to the readers this essay will leave unbudged. Very rarely does someone protest against a favorable reaction to the only part of his work another has
experienced, on the grounds that the work is an organic unity, no part of which
can be judged in isolation.
I. The reader who has looked ahead and seen that the second part of this essay
discusses Rawls’ theory, mistakenly may think that every remark or argument
in the first part against alternative theories of justice is meant to apply to or
anticipate a criticism of his theory. This is not so; there are other theories
also worth criticizing.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
47 Distributive Justice
of what comes to be held by a particular process, and so on. We shall
refer to the complicated truth about this topic, which we shall not
formulate here, as the principle of justice in acquisition. The second
topic concerns the transfer of holdings from one person to another.
By what processes may a person transfer holdings to another? How
may a person acquire a holding from another who holds it? Under
this topic come general descriptions of voluntary exchange, and gift,
and (on the other hand) fraud, as well as reference to particular conventional details fixed upon a given society. The complicated truth
about this subject (with placeholders for conventional details) we
shall call the principle of justice in transfer. (And we shall suppose
it also includes principles governing how a person may divest himself
of a holding, passing it into an unheld state.)
If the world were wholly just, the following inductive definition
would exhaustively cover the subject of justice in holdings.
(i) A person who acquires a holding in accordance with the principle of justice in acquisition is entitled to that holding.
(2) A person who acquires a holding in accordance with the principle of justice in transfer, from someone else entitled to the
holding, is entitled to the holding.
(3) No one is entitled to a holding except by (repeated) applications of (i) and (2).
The complete principle of distributive justice would say simply that a
distribution is just if everyone is entitled to the holdings they possess
under the distribution.
A distribution is just if it arises from another (just) distribution
by legitimate means. The legitimate means of moving from one distribution to another are specified by the principle of justice in transfer.
The legitimate first “moves” are specified by the principle of justice
in acquisition.2 Whatever arises from a just situation by just steps is
itself just. The means of change specified by the principle of justice
2. Applications of the principle of justice in acquisition, may also occur as
part of the move from one distribution to another. You may find an unheld
thing now, and appropriate it. Acquisitions also are to be understood as included
when, to simplify, I speak only of transitions by transfers.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
48 Philosophy & Public Affairs
in transfer, preserve justice. As correct rules of inference are truth
preserving, and any conclusion deduced via repeated application of
such rules from only true premisses is itself true, so the means of
transition from one situation to another specified by the principle of
justice in transfer are justice preserving, and any situation actually
arising from repeated transitions in accordance with the principle
from a just situation is itself just. The parallel between justice-preserving transformations and truth-preserving transformations illuminates where it fails as well as where it holds. That a conclusion
could have been deduced by truth-preserving means from premisses
that are true suffices to show its truth. That a situation could have arisen via justice-preserving means from a just situation does not suffice to
show its justice. The fact that a thief’s victims voluntarily could have
presented him with gifts, does not entitle the thief to his ill-gotten
gains. Justice in holdings is historical; it depends upon what actually
has happened. We shall return to this point below.
Not all actual situations are generated in accordance with the two
principles of justice in holdings: the principle of justice in acquisition
and the principle of justice in transfer. Some people steal from others,
or defraud them, or enslave them seizing their product and preventing
them from living as they choose, or forcibly exclude others from
competing in exchanges. None of these are permissible modes of
transition from one situation to another. And some persons acquire
holdings by means not sanctioned by the principle of justice in acquisition. The existence of past injustice (previous violations of the first
two principles of justice in holdings) raises the third major topic
under justice in holdings: the rectification of injustice in holdings.
If past injustice has shaped present holdings in various ways, some
identifiable and some not, what now, if anything, ought to be done to
rectify these injustices? What obligations are the performers of injustice under to their victims? What obligations do the beneficiaries of
injustice have to those whose position is worse than it would have been
had the injustice not been done? Or, than it would have been had
compensation been paid promptly? How, if at all, do things change if
the beneficiaries and those made worse off are not the direct parties
in the act of injustice, but, for example, their descendants? Is an injustice done to someone whose holding was itself based upon an unThis content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
49 Distributive Justice
rectified injustice? How far back must one go in wiping clean the
historical slate of injustices? What may victims of injustice permissibly do in order to rectify the injustices being done to them, including
the many injustices done by persons acting through their government?
I do not know of a thorough or theoretically sophisticated treatment
of such issues. Idealizing greatly, let us suppose theoretical investigation will produce a principle of rectification. This principle uses historical information about previous situations and injustices done in
them (as defined by the first two principles of justice, and rights
against interference), and information about the actual course of
events that flowed from these injustices, up until the present, and it
yields a description (or descriptions) of holdings in the society. The
principle of rectification presumably will make use of (its best estimate of) subjunctive information about what would have occurred
(or a probability distribution over what might have occurred, using
the expected value) if the injustice had not taken place. If the actual
description of holdings turns out not to be one of the descriptions
yielded by the principle, then one of the descriptions yielded must be
realized.3
The general outlines of the theory of justice in holdings are that
the holdings of a person are just if he is entitled to them by the principles of justice in acquisition and transfer, or by the principle of
rectification of injustice (as specified by the first two principles). If
each person’s holdings are just then the total set (distribution) of
holdings is just. To tum these general outlines into a specific theory
we would have to specify the details of each of the three principles of
justice in holdings: the principle of acquisition of holdings, the principle of transfer of holdings, and the principle of rectification of violations of the first two principles. I shall not attempt that task here.
(Locke’s principle of justice in acquisition is discussed below.)
3. If the principle of rectification of violations of the first two principles
yields more than one description of holdings, then some choice must be made
as to which of these is to be realized. Perhaps the sort of considerations about
distributive justice and equality I argue against play a legitimate role in this
subsidiary choice. Similarly, there may be room for such considerations in deciding which otherwise arbitrary features a statute will embody, when such
features are unavoidable because other considerations do not specify a precise
line, yet one must be drawn.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
50 Philosophy & Public Affairs
i. Historical Principles and End-Result Principles. The general
outlines of the entitlement theory illuminate the nature and defects
of other conceptions of distributive justice. The entitlement theory of
justice in distribution is historical; whether a distribution is just depends upon how it came about. In contrast, current time-slice principles of justice hold that the justice of a distribution is determined
by how things are distributed (who has what) as judged by some
structural principle (s) of just distribution. A utilitarian who judges between any two distributions by seeing which has the greater sum of utility and, if these tie, who applies some fixed equality criterion to choose
the more equal distribution, would hold a current time-slice principle
of justice. As would someone who had a fixed schedule of trade-offs
between the sum of happiness and equality. All that needs to be looked
at, in judging the justice of a distribution, according to a current timeslice principle, is who ends up with what; in comparing any two distributions one need look only at the matrix presenting the distributions. No further information need be fed into a principle of justice.
It is a consequence of such principles of justice that any two structurally identical distributions are equally just. (Two distributions are
structurally identical if they present the same profile, but [perhaps]
have different persons occupying the particular slots. My having ten
and your having five, and my having five and your having ten are
structurally identical distributions.) Welfare economics is the theory
of current time-slice principles of justice. The subject is conceived as
operating on matrices representing only current information about
distribution. This, as well as some of the usual conditions (e.g., the
choice of distribution is invariant under relabeling of columns), guarantees that welfare economics will be a current time-slice theory, with
all of its inadequacies.
Most persons do not accept current time-slice principles as constituting the whole story about distributive shares. They think it
relevant in assessing the justice of a situation to consider not only
the distribution it embodies, but also how that distribution came about.
If some persons are in prison for murder or war crimes, we do not
say that to assess the justice of the distribution in the society we
must look only at what this person has, and that person has, and
that person has . . . , at the current time. We think it relevant to
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
51 Distributive Justice
ask whether someone did something so that he deserved to be punished, deserved to have a lower share. Most will agree to the relevance
of further information with regard to punishments and penalties.
Consider also desired things. One traditional socialist view is that
workers are entitled to the product and full fruits of their labor; they
have earned it; a distribution is unjust if it does not give the workers
what they are entitled to. Such entitlements are based upon some past
history. No socialist holding this view would find it comforting to be
told that because the actual distribution A happens to coincide structurally with the one he desires D, A therefore is no less just than D;
it differs only in that the “<parasitic” owners of capital receive under A
what the workers are entitled to under D, and the workers receive
under A what the owners are entitled to (under D), namely very little.
Rightly in my view, this socialist holds onto the notions of earning,
producing, entitlement, desert, etc. and he rejects (current time-slice)
principles that look only to the structure of the resulting set of holdings. (The set of holdings resulting from what? Isn’t it implausible
that how holdings are produced and come to exist has no effect at all
on who should hold what?) His mistake lies in his view of what
entitlements arise out of what sorts of productive processes.
We construe the position we discuss too narrowly by speaking of
current time-slice principles. Nothing is changed if structural principles operate upon a time sequence of current time-slice profiles and,
for example, give someone more now to counterbalance the less he
has had earlier. A utilitarian or an egalitarian or any mixture of the
two over time will inherit the difficulties of his more myopic comrades.
He is not helped by the fact that some of the information others consider relevant in assessing a distribution is reflected, unrecoverably, in
past matrices. Henceforth, we shall refer to such unhistorical principles of distributive justice, including the current time-slice principles, as end-result principles or end-state principles.
In contrast to end-result principles of justice, historical principles
of justice hold that past circumstances or actions of people can create
differential entitlements or differential deserts to things. An injustice
can be worked by moving from one distribution to another structurally
identical one, for the second, in profile the same, may violate people’s
entitlements or deserts; it may not fit the actual history.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
52 Philosophy & Public Affairs
2. Patterning. The entitlement principles of justice in holdings that
we have sketched are historical principles of justice. To better understand their precise character, we shall distinguish them from another
subclass of the historical principles. Consider, as an example, the
principle of distribution according to moral merit. This principle requires total distributive shares to vary directly with moral merit; no
person should have a greater share than anyone whose moral merit
is greater. (If moral merit could be not merely ordered but measured
on an interval or ratio scale, stronger principles could be formulated.)
Or consider the principle that results by substituting “usefulness to
society” for “moral merit” in the previous principle. Or instead of
“distribute according to moral merit,” or “distribute according to usefulness to society,” we might consider “distribute according to the
weighted sum of moral merit, usefulness to society, and need,” with
the weights of the different dimensions equal. Let us call a principle
of distribution patterned if it specifies that a distribution is to vary
along with some natural dimension, weighted sum of natural dimensions, or lexicographic ordering of natural dimensions. And let us say
a distribution is patterned if it accords with some patterned principle.
(I speak of natural dimensions, admittedly without a general criterion
for them, because for any set of holdings some artificial dimensions
can be gimmicked up to vary along with the distribution of the set.)
The principle of distribution in accordance with moral merit is a
patterned historical principle, which specifies a patterned distribution. “Distribute according to I.Q.” is a patterned principle that looks
to information not contained in distributional matrices. It is not historical, however, in that it does not look to any past actions creating
differential entitlements to evalute a distribution; it requires only distributional matrices whose columns are labeled by I.Q. scores. The
distribution in a society, however, may be composed of such simple
patterned distributions, without itself being simply patterned. Different sectors may operate different patterns, or some combination
of patterns may operate in different proportions across a society. A
distribution composed in this manner, from a small number of patterned distributions, we also shall term patterned. And we extend the
use of “pattern” to include the overall designs put forth by combinations of end-state principles.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
53 Distributive Justice
Almost every suggested principle of distributive justice is patterned:
to each according to his moral merit, or needs, or marginal product,
or how hard he tries, or the weighted sum of the foregoing, and so on.
The principle of entitlement we have sketched is not patterned.4 There
is no one natural dimension or weighted sum or combination of (a
small number of) natural dimensions that yields the distributions
generated in accordance with the principle of entitlement. The set of
holdings that results when some persons receive their marginal products, others win at gambling, others receive a share of their mate’s
income, others receive gifts from foundations, others receive interest
on loans, others receive gifts from admirers, others receive returns on
investment, others make for themselves much of what they have,
others find things, and so on, will not be patterned. Heavy strands of
patterns will run through it; significant portions of the variance in
holdings will be accounted for by pattern variables. If most people
most of the time choose to transfer some of their entitlements to others
only in exchange for something from them, then a large part of what
many people hold will vary with what they held that others wanted.
More details are provided by the theory of marginal productivity. But
gifts to relatives, charitable donations, bequests to children, and the
like, are not best conceived, in the first instance, in this manner. Ignoring the strands of pattern, let us suppose for the moment that a
distribution actually gotten by the operation of the principle of entitlement is random with respect to any pattern. Though the resulting set
of holdings will be unpatterned, it will not be incomprehensible, for it
4. One might try to squeeze a patterned conception of distributive justice into
the framework of the entitlement conception, by formulating a gimmicky obligatory ‘principle of transfer’ that would lead to the pattern. For example, the
principle that if one has more than the mean income, one must transfer everything one holds above the mean to persons below the mean so as to bring them
up to (but not over) the mean. We can formulate a criterion for a ‘principle of
transfer’ to rule out such obligatory transfers, or we can say that no correct
principle of transfer, no principle of transfer in a free society will be like this.
The former is probably the better course, though the latter also is true.
Altematively, one might think to make the entitlement conception instantiate
a pattern, by using matrix entries that express the relative strength of a person’s
entitlements as measured by some real-valued function. But even if the limitation to natural dimensions failed to exclude this function, the resulting edifice
would not capture our system of entitlements to particular things.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
54 Philosophy & Public Affairs
can be seen as arising from the operation of a small number of principles. These principles specify how an initial distribution may arise
(the principle of acquisition of holdings) and how distributions may
be transformed into others (the principle of transfers of holdings).
The process whereby the set of holdings is generated will be intelligible, though the set of holdings itself that results from this process
will be unpatterned.
The writings of F. A. Hayek focus less than others’ upon what patterning distributive justice requires. Hayek argues that we cannot
know enough about each person’s situation to distribute to each according to his moral merit (but would justice demand we do so if
we did have this knowledge?); and he goes on to say, “our objection
is against all attempts to impress upon society a deliberately chosen
pattern of distribution, whether it be an order of equality or of inequality.”5 However, Hayek concludes that in a free society there will
be distribution in accordance with value rather than (moral) merit;
that is, in accordance with the perceived value of a person’s actions
and services to others. Despite his rejection of a patterned conception of
distributive justice, Hayek himself suggests a pattern he thinks justifiable: distribution in accordance with the (perceived) benefits given
to others, and so leaves room for the complaint that a free society does
not realize exactly this pattern. Stating this patterned strand of a free
capitalist society more precisely, we get: “To each according to how
much he benefits others who have the resources for benefitting those
who benefit them.” This will seem arbitrary unless some acceptable
initial set of holdings is specified, or unless it is held that the operation
of the system over time washes out any significant effects from the initial set of holdings. As an example of the latter, if almost anyone would
have bought a car from Henry Ford, the supposition that it was an arbitrary matter who held the money then (and so bought) would not place
Henry Ford’s earnings under a cloud. In any event, his coming to hold
it is not arbitrary. Distribution according to benefits to others is a
major patterned strand in a free capitalist society, as Hayek correctly
points out, but it is only a strand and does not constitute the whole
pattern of a system of entitlements (viz., inheritance, gifts for arbi5. F. A. Hayek, The Constitution of Liberty (Chicago, 1972), chap. 6: “Equality, Value, and Merit,” p. 87.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
55 Distributive Justice
trary reasons, charity, etc.) or a standard one should insist a society
fit. Will people tolerate for long a system yielding distributions that
(they believe) are unpatterned?6 No doubt people will not long accept
a distribution they believe is unjust. People want their society to be
and to look just. But must the look of justice reside in a resulting pattern rather than in the underlying generating principles? We are in
no position to conclude the inhabitants of a society embodying an
entitlement conception of justice in holdings will find it unacceptable.
Still, it must be granted that were people’s reasons for transferring
some of their holdings to others always irrational or arbitrary, we
would find this disturbing. (Suppose people always determined what
holdings they would transfer, and to whom, by using a random device.) We feel more comfortable upholding the justice of an entitlement system if most of the transfers under it are done for reasons.
This does not mean necessarily that all deserve what holdings they
receive. It means only that there is a purpose or point to someone’s
transferring a holding to one person rather than to another; that
usually we can see what the transferrer thinks he’s gaining, what
cause he thinks he’s serving, what goals he thinks he’s helping to
achieve, etc. Since often in a capitalist society people transfer holdings
to others in accordance with how much they perceive these others
benefitting them, the fabric constituted by the individual transactions
and transfers is largely reasonable and intelligible. (Gifts to loved
ones, bequests to children, charity to the needy also are nonarbitrary
components of the fabric.) In stressing the large strand of distribution
in accordance with benefit to others, Hayek shows the point of many
transfers, and so shows that the system of transfer of entitlements is
not just spinning its gears aimlessly. The system of entitlements is
defensible when constituted by the individual aims of individual trans6. This question does not imply that they will tolerate any and every patterned distribution. In discussing Hayek’s views, Irving Kristol has recently
speculated that people will not long tolerate a system that yields distributions
patterned in accordance with value rather than merit. (“‘When Virtue Loses
All Her Loveliness’-Some Reflections on Capitalism and ‘The Free Society,”‘
The Public Interest [Fall I970], pp. 3-15.) Kristol, following some remarks of
Hayek’s, equates the latter with justice. Since some case can be made for the
external standard of distribution in accordance with benefit to others, we ask
about a weaker (and therefore more plausible) hypothesis.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
56 Philosophy & Public Affairs
actions. No overarching aim is needed, no distributional pattern is
required.
To think that the task of a theory of distributive justice is to fill in
the blank in “to each according to his ,” is to be predisposed
to search for a pattern; and the separate treatment of “from each accordir(g to his ,” treats production and distribution as two
separate and independent issues. On an entitlement view these are
not two separate questions. Whomever makes something, having
bought or contracted for all other held resources used in the process
(transferring some of his holdings for these cooperating factors), is
entitled to it. The situation is not one of something’s getting made, and
there being an open question of who is to get it. Things come into the
world already attached to people having entitlements over them. From
the point of view of the historical entitlement conception of justice in
holdings, those who start afresh to complete “to each according to
his ,” treat objects as if they appeared from nowhere, out of
nothing. A complete theory of justice might cover this limit case as
well; here perhaps is a use for the usual conceptions of distributive
justice.7
So entrenched are maxims of the usual form that perhaps we should
present the entitlement conception as a competitor. Ignoring acquisition and rectification, we might say:
From each according to what he chooses to do, to each according
to what he makes for himself (perhaps with the contracted-for aid
of others) and what others choose to do for him and choose to give
him of what they’ve been given previously (under this maxim)
and haven’t yet expended or transferred.
This, the discerning reader will have noticed, has its defects as a
slogan. So as a summary (and not as a maxim with any independent
meaning) and great simplification we have:
From each as they choose, to each as they are chosen.
7. Varying situations continuously from that limit situation to our own would
force us to consider whether entitlement considerations lexicographically precede
the considerations of the usual theories of distributive justice, so that the slightest strand of entitlement outweighs the considerations of the usual theories of
distributive justice.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
57 Distributive Justice
3. How Liberty Upsets Patterns. It is not clear how those holding
alternative conceptions of distributive justice can reject the entitlement conception of justice in holdings. For suppose a distribution
favored by one of these nonentitlement conceptions is realized. Let us
suppose it is your favorite one and call this distribution D1; perhaps
everyone has an equal share, perhaps shares vary in accordance with
some dimension you treasure. Now suppose that Wilt Chamberlain
is greatly in demand by basketball teams, being a great gate-attraction.
(Also suppose contracts run only for a year, with players being free
agents.) He signs the following sort of contract with a team: In each
home game, twenty-five cents from the price of each ticket of admission goes to him. (We ignore the question of whether he is “gouging”
the owners, letting them look out for themselves.) The season starts,
and people cheerfully attend his team’s games; they buy their tickets,
each time dropping a separate twenty-five cents of their admission
price into a special box with Chamberlain’s name on it. They are excited about seeing him play; it is worth the total admission price to
them. Let us suppose that in one season one million persons attend
his home games, and Wilt Chamberlain winds up with $250,000, a
much larger sum than the average income and larger even than anyone else has. Is he entitled to this income? Is this new distribution D2
unjust? If so, why? There is no question about whether each of the
people was entitled to the control over the resources they held, in D1,
because that was the distribution (your favorite) that (for the purposes of argument) we assumed was acceptable. Each of these persons
chose to give twenty-five cents of their money to Chamberlain. They
could have spent it on going to the movies, or on candy bars, or on
copies of Dissent magazine, or of Monthly Review. But they all, at
least one million of them, converged on giving it to Wilt Chamberlain
in exchange for watching him play basketball. If D1 was a just distribution, and people voluntarily moved from it to D2, transferring parts
of their shares they were given under D1 (what was it for if not to do
something with?), isn’t D2 also just? If the people were entitled to
dispose of the resources to which they were entitled (under D1), didn’t
this include their being entitled to give it to, or exchange it with, Wilt
Chamberlain? Can anyone else complain on grounds of justice? Each
other person already has his legitimate share under D1. Under D1 there
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
58 Philosophy & Public Affairs
is nothing that anyone has that anyone else has a claim of justice
against. After someone transfers something to Wilt Chamberlain,
third parties still have their legitimate shares; their shares are not
changed. By what process could such a transfer among two persons
give rise to a legitimate claim of distributive justice on a portion of
what was transferred, by a third party who had no claim of justice on
any holding of the others before the transfer?8 To cut off objections
irrelevant here, we might imagine the exchanges occurring in a socialist society, after hours. After playing whatever basketball he does in
his daily work, or doing whatever other daily work he does, Wilt Chamberlain decides to put in overtime to earn additional money. (First his
work quota is set; he works time over that.) Or imagine it is a skilled
juggler people like to see, who puts on shows after hours.
Why might some people work overtime in a society in which it is
assumed their needs are satisfied? Perhaps because they care about
things other than needs. I like to write in books that I read, and to
have easy access to books for browsing at odd hours. It would be very
pleasant and convenient to have the resources of Widener Library
in my back yard. No society, I assume, will provide such resources
close to each person who would like them as part of his regular allotment (under D1). Thus, persons either must do without some extra
things that they want, or be allowed to do something extra to get
8. Might not a transfer have instrumental effects on a third party, changing
his feasible options? (But what if the two parties to the transfer independently
had used their holdings in this fashion?) I discuss this question elsewhere, but
note here that this question concedes the point for distributions of ultimate intrinsic noninstrumental goods (pure utility experiences, so to speak) that are
transferrable. It also might be objected that the transfer might make a third
party more envious because it worsens his position relative to someone else. I
find it incomprehensible how it can be thought that this involves a claim of
justice. On envy, see Anarchy, State, and Utopia, chap. 8.
Here and elsewhere in this essay, a theory which incorporates elements of
pure procedural justice might find what I say acceptable, if kept in its proper
place; that is, if background institutions exist to ensure the satisfaction of
certain conditions on distributive shares. But if these institutions are not themselves the sum or invisible-hand result of people’s voluntary (nonaggressive)
actions, the constraints they impose require justification. At no point does our
argument assume any background institutions more extensive than those of
the minimal night-watchman state, limited to protecting persons against
murder, assault, theft, fraud, etc.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
59 Distributive Justice
(some of) these things. On what basis could the inequalities that
would eventuate be forbidden? Notice also that small factories would
spring up in a socialist society, unless forbidden. I melt down some
of my personal possessions (under D1) and build a machine out of the
material. I offer you, and others, a philosophy lecture once a week in
exchange for your cranking the handle on my machine, whose products I exchange for yet other things, and so on. (The raw materials
used by the machine are given to me by others who possess them under
D1, in exchange for hearing lectures.) Each person might participate
to gain things over and above their allotment under D1. Some persons
even might want to leave their job in socialist industry, and work full
time in this private sector. I say something more about these issues
elsewhere. Here I wish merely to note how private property, even in
means of production, would occur in a socialist society that did not
forbid people to use as they wished some of the resources they are
given under the socialist distribution D1. The socialist society would
have to forbid capitalist acts between consenting adults.9
The general point illustrated by the Wilt Chamberlain example and
the example of the entrepreneur in a socialist society is that no end9. See the selection from John Henry MacKay’s novel, The Anarchists, reprinted in Leonard Krimmerman and Lewis Perry, eds., Patterns of Anarchy
(New York, I966), pp. I6-33, in which an individualist anarchist presses upon a
communist anarchist the question: “Would you, in the system of society which
you call ‘free Communism’ prevent individuals from exchanging their labor
among themselves by means of their own medium of exchange? And further:
Would you prevent them from occupying land for the purpose of personal use?”
The novel continues: “[the] question was not to be escaped if he answered ‘Yes!’
he admitted that society had the right of control over the individual and threw
overboard the autonomy of the individual which he had always zealously defended; if on the other hand, he answered ‘No!’ he admitted the right of private
property which he had just denied so emphatically…. Then he answered ‘In
Anarchy any number of men must have the right of forming a voluntary association, and so realizing their ideas in practice. Nor can I understand how any
one could justly be driven from the land and house which he uses and occupies
. . . every serious man must declare himself: for Socialism, and thereby for
force and against liberty, or for Anarchism, and thereby for liberty and against
force.'” In contrast, we find Noam Chomsky writing, “Any consistent anarchist
must oppose private ownership of the means of production,” and “the consistent
anarchist then . . . will be a socialist . . . of a particular sort” (Introduction to
Daniel Guerin, Anarchism: From Theory to Practice [New York, I9701, pp. Xiii
and xv).
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
6o Philosophy & Public Affairs
state principle or distributional pattern principle of justice can be
continuously realized without continuous interference into people’s
lives. Any favored pattern would be transformed into one unfavored
by the principle, by people choosing to act in various ways; e.g., by
people exchanging goods and services with other people, or giving
things to other people, things the transferrers are entitled to under
the favored distributional pattern. To maintain a pattern one must
either continuously interfere to stop people from transferring resources as they wish to, or continually (or periodically) interfere to
take from some persons resources that others for some reason chose
to transfer to them. (But if some time limit is to be set on how long
people may keep resources others voluntarily transfer to them, why
let them keep these resources for any period of time? Why not have
immediate confiscation?) It might be objected that all persons voluntarily will choose to refrain from actions which would upset the pattern. This presupposes unrealistically (a) that all will most want to
maintain the pattern (are those who don’t, to be “reeducated” or forced
to undergo “self-criticism”?); (b) that each can gather enough information about his own actions and the ongoing activities of others
to discover which of his actions will upset the pattern; and (c) that
diverse and farflung persons can coordinate their actions to dovetail
into the pattern. Compare the manner in which the market is neutral
among persons’ desires, as it reflects and transmits widely scattered
information via prices, and coordinates persons’ activities.
It puts things perhaps a bit too strongly to say that every patterned
(or end-state) principle is liable to be thwarted by the voluntary actions of the individual parties transferring some of their shares they
receive under the principle. For perhaps some very weak patterns are
not so thwarted.10 Any distributional pattern with any egalitarian comio. Is the patterned principle stable that requires merely that a distribution be
Pareto-optimal? One person might give another a gift or bequest that the second
could exchange with a third to their mutual benefit. Before the second makes
this exchange, there is not Pareto-optimality. Is a stable pattern presented by a
principle choosing that among the Pareto-optimal positions that satisfies some
further condition C? It may seem there cannot be a counterexample, for won’t
any voluntary exchange made away from a situation show that the first situation wasn’t Pareto-optimal? (Ignore the implausibility of this last claim for the
case of bequests.) But principles are to be satisfied over time, during which
new possibilities arise. A distribution that at one time satisfies the criterion of
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
6I Distributive Justice
ponent is overturnable by the voluntary actions of individual persons
over time; as is every patterned condition with sufficient content so
as actually to have been proposed as presenting the central core of
distributive justice. Still, given the possibility that some weak conditions or patterns may not be unstable in this way, it would be better
to formulate an explicit description of the kind of (interesting and
contentful) patterns under discussion, and to prove a theorem about
their instability. Since the weaker the patterning, the more likely it
is that the entitlement system itself satisfies it, a plausible conjecture
is that any patterning either is unstable or is satisfied by the entitlement system.
4. Sen’s Argument. Our conclusions are reinforced by considering a
recent general argument of Amartya K. Sen.11 Suppose individual rights
are interpreted as the right to choose which of two alternatives is to be
more highly ranked in a social ordering of the alternatives. Add the
weak condition that if one alternative unanimously is preferred to another then it is ranked higher by the social ordering. If there are two
different individuals each with individual rights, interpreted as above,
over different pairs of alternatives (having no members in common),
then for some possible preference rankings of the alternatives by the
individuals, there is no linear social ordering. For suppose that person
i has the right to decide among (X,Y) and person II has the right to
decide among (Z,W); and suppose their individual preferences are
as follows (and that there are no other individuals). Person I prefers
W to X to Y to Z, and person ii prefers Y to Z to W to X. By the unanimity condition, in the social ordering W is preferred to X (since each
individual prefers it to X), and Y is preferred to Z (since each individual prefers it to Z). Also in the social ordering, X is preferred to Y, by
person I’s right of choice among these two alternatives. Combining
Pareto-optimality might not do so when some new possibilities arise (Wilt
Chamberlain grows up and starts playing basketball); and though people’s
activities will tend to move then to a new Pareto-optimal position, this new one
need not satisfy the contentful condition C. Continual interference will be needed
to insure the continual satisfaction of C. (The theoretical possibility should be
investigated of a pattern’s being maintained by some invisible-hand process that
brings it back to an equilibrium that fits the pattern when deviations occur.)
ii. Collective Choice and Social Welfare (San Francisco, 1970), chaps. 6
and 6*.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
62 Philosophy & Public Affairs
these three binary rankings, we get W preferred to X preferred to Y
preferred to Z, in the social ordering. However, by person II’s right of
choice, Z must be preferred to W in the social ordering. There is no
transitive social ordering satisfying all these conditions, and the social
ordering, therefore, is nonlinear. Thus far, Sen.
The trouble stems from treating an individual’s right to choose
among alternatives as the right to determine the relative ordering of
these alternatives within a social ordering. The system is no better
that has individuals rank pairs of alternatives, and separately rank
the individual alternatives; their ranking of pairs feeds into some
method of amalgamating preferences to yield a social ordering of
pairs; and the choice among the alternatives in the highest ranked
pair in the social ordering is made by the individual with the right to
decide between this pair. This system also has the result that an alternative may be selected although everyone prefers some other alternative; e.g., I selects X over Y, where (X,Y) somehow is the highest
ranked pair in the social ordering of pairs, although everyone, including i, prefers W to X. (But the choice person I was given, however,
was only between X and Y.)
A more appropriate view of individual rights is as follows. Individual rights are co-possible; each person may exercise his rights as he
chooses. The exercise of these rights fixes some features of the world.
Within the constraints of these fixed features, a choice may be made
by a social choice mechanism based upon a social ordering, if there
are any choices left to make! Rights do not determine a social ordering but instead set the constraints within which a social choice is
to be made, by excluding certain alternatives, fixing others, and so
on. (If I have a right to choose to live in New York or in Massachusetts, and I choose Massachusetts, then alternatives involving my living in New York are not appropriate objects to be entered in a social
ordering.) Even if all possible alternatives are ordered first, apart
from anyone’s rights, the situation is not changed: for then the highest
ranked alternative that is not excluded by anyone’s exercise of their
rights is instituted. Rights do not determine the position of an alternative or the relative position of two alternatives in a social ordering;
they operate upon a social ordering to constrain the choice it can yield.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
63 Distributive Justice
If entitlements to holdings are rights to dispose of them, then social
choice must take place within the constraints of how people choose
to exercise these rights. If any patterning is legitimate, it falls within
the domain of social choice, and hence is constrained by people’s
rights. How else can one cope with Sen’s result? The alternative of
first having a social ranking with rights exercised within its constraints, is no alternative at all. Why not just select the top ranked
alternative and forget about rights? If that top ranked alternative
itself leaves some room for individual choice (and here is where
“rights” of choice is supposed to enter in) there must be something
to stop these choices from transforming it into another alternative.
Thus Sen’s argument leads us again to the result that patterning requires continuous interference with individuals’ actions and choices.12
5. Redistribution and Property Rights. Apparently patterned principles allow people to choose to expend upon themselves, but not upon
others, those resources they are entitled to (or rather, receive) under
some favored distributional pattern D1. For if each of several persons
chooses to expend some of his D1 resources upon one other person,
then that other person will receive more than his D1 share, disturbing
the favored distributional pattern. Maintaining a distributional pattern is individualism with a vengeance! Patterned distributional principles do not give people what entitlement principles do, only better
distributed. For they do not give the right to choose what to do with
what one has; they do not give the right to choose to pursue an end
involving (intrinsically, or as a means) the enhancement of another’s position. To such views, families are disturbing; for within a
family occur transfers that upset the favored distributional pattern.
Either families themselves become units to which distribution takes
place, the column occupiers (on what rationale?), or loving behavior
is forbidden. We should note in passing the ambivalent position of
radicals towards the family. Its loving relationships are seen as a
12. Oppression will be less noticeable if the background institutions do not
prohibit certain actions that upset the patterning (various exchanges or transfers of entitlement), but rather prevent them from being done, by nullifying
them.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
64 Philosophy & Public Affairs
model to be emulated and extended across the whole society, while it
is denounced as a suffocating institution to be broken, and condemned
as a focus of parochial concerns that interfere with achieving radical
goals. Need we say that it is not appropriate to enforce across the
wider society the relationships of love and care appropriate within a
family, relationships which are voluntarily undertaken?13 Incidentally,
love is an interesting instance of another relationship that is historical,
in that (like justice) it depends upon what actually occurred. An adult
may come to love another because of the other’s characteristics; but
it is the other person, and not the characteristics, that is loved. The
love is not transferable to someone else with the same characteristics,
even to one who ‘scores>> higher for these characteristics. And the love
endures through changes of the characteristics that gave rise to it. One
loves the particular person one actually encountered. Why love is historical, attaching to persons in this way and not to characteristics,
is an interesting and puzzling question.
Proponents of patterned principles of distributive justice focus upon
criteria for determining who is to receive holdings; they consider the
reasons for which someone should have something, and also the total
picture of holdings. Whether or not it is better to give than to receive,
proponents of patterned principles ignore giving altogether. In considering the distribution of goods, income, etc., their theories are
theories of recipient-justice; they completely ignore any right a person
might have to give something to someone. Even in exchanges where
each party is simultaneously giver and recipient, patterned principles
of justice focus only upon the recipient role and its supposed rights.
Thus discussions tend to focus on whether people (should) have a
right to inherit, rather than on whether people (should) have a right
13. One indication of the stringency of Rawls’s difference principle, which
we attend to in the second part of this essay, is its inappropriateness as a governing principle even within a family of individuals who love one another. Should
a family devote its resources to maximizing the position of its least well off and
talented child, holding back the other children or using resources for their
education and development only if they will follow a policy throughout their lifetimes of maximizing the position of their least fortunate sibling? Surely not.
How then can this even be considered as the appropriate policy for enforcement
in the wider society? (I discuss below what I think would be Rawls’s reply:
that some principles apply at the macro-level which do not apply to microsituations.)
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
65 Distributive Justice
to bequeath or on whether persons who have a right to hold also have
a right to choose that others hold in their place. I lack a good explanation of why the usual theories of distributive justice are so recipientoriented; ignoring givers and transferrers and their rights is of a piece
with ignoring producers and their entitlements. But why is it all
ignored?
Patterned principles of distributive justice necessitate redistributive
activities. The likelihood is small that any actual freely arrived at set
of holdings fits a given pattern; and the likelihood is nil that it will
continue to fit the pattern as people exchange and give. From the
point of view of an entitlement theory, redistribution is a serious matter indeed, involving, as it does, the violation of people’s rights. (An
exception is those takings that fall under the principle of the rectification of injustices.) From other points of view, also, it is serious.
Taxation of earnings from labor is on a par with forced labor.’4
Some persons find this claim obviously true: taking the earnings of n
hours labor is like taking n hours from the person; it is like forcing
the person to work n hours for another’s purpose. Others find the claim
absurd. But even these, if they object to forced labor, would oppose
forcing unemployed hippies to work for the benefit of the needy.’5
And they also would object to forcing each person to work five extra
hours each week for the benefit of the needy. But a system that takes
five hours’ wages in taxes does not seem to them like one that forces
someone to work five hours, since it offers the forcee a wider range of
choice in activities than does taxation in kind with the particular labor
specified. (But we can imagine a gradation of systems of forced labor,
from one that specifies a particular activity, to one that gives a choice
among two activities, to . . . ; and so on up.) Furthermore, people
14. I am unsure as to whether the arguments I present below show that such
taxation just is forced labor; so that “is on a par with” means “is one kind of.”
Or alternatively, whether the arguments emphasize the great similarities between such taxation and forced labor, to show it is plausible and illuminating
to view such taxation in the light of forced labor. This latter approach would
remind one of how John Wisdom conceives of the claims of metaphysicians.
15. Nothing hangs on the fact that here and elsewhere I speak loosely of
needs; since I go on, each time, to reject the criterion of justice which includes
it. If, however, something did depend upon the notion, one would want to examine it more carefully. For a skeptical view, see Kenneth Minogue, The Liberal
Mind (New York, I963), pp. 103-112.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
66 Philosophy & Public Affairs
envisage a system with something like a proportional tax on everything above the amount necessary for basic needs. Some think this
does not force someone to work extra hours, since there is no fixed
number of extra hours he is forced to work, and since he can avoid the
tax entirely by earning only enough to cover his basic needs. This is
a very uncharacteristic view of forcing for those who also think people
are forced to do something whenever the alternatives they face are
considerably worse. However, neither view is correct. The fact that
others intentionally intervene, in violation of a side-constraint against
aggression, to threaten force to limit the alternatives, in this case to
paying taxes or (presumably the worse alternative) bare subsistence,
makes the taxation system one of forced labor, and distinguishes it
from other cases of limited choices which are not forcings.16
The man who chooses to work longer to gain an income more than
sufficient for his basic needs prefers some extra goods or services to
the leisure and activities he could perform during the possible nonworking hours; whereas the man who chooses not to work the extra
time prefers the leisure activities to the extra goods or services he
could acquire by working more. Given this, if it would be illegitimate
for a tax system to seize some of a man’s leisure (forced labor) for
the purpose of serving the needy, how can it be legitimate for a tax
system to seize some of a man’s goods for that purpose? Why should
we treat the man whose happiness requires certain material goods
or services differently from the man whose preferences and desires
make such goods unnecessary for his happiness? Why should the man
who prefers seeing a movie (and who has to earn money for a ticket)
be open to the required call to aid the needy, while the person who
prefers looking at a sunset (and hence need earn no extra money) is
not? Indeed, isn’t it surprising that redistributionists choose to ignore
the man whose pleasures are so easily attainable without extra labor,
while adding yet another burden to the poor unfortunate who must
work for his pleasures? If anything, one would have expected the
reverse. Why is the person with the nonmaterial or nonconsumption
desire allowed to proceed unimpeded to his most favored feasible ali6. Further details that this statement should include are contained in my
essay, “Coercion,” in Philosophy, Science, and Method, eds. S. Morgenbesser,
P. Suppes, and M. White (New York, I969).
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
67 Distributive Justice
ternative, whereas the man whose pleasures or desires involve material
things and who must work for extra money (thereby serving whoever considers his activities valuable enough to pay him) is constrained in what he can realize? Perhaps there is no difference in
principle. And perhaps some think the answer concerns merely administrative convenience. (These questions and issues will not disturb
those who think forced labor to serve the needy or realize some favored
end-state pattern acceptable.) In a fuller discussion we would have
(and want) to extend our argument to include interest, entrepreneurial profits, etc. Those who doubt that this extension can be carried
through, and who draw the line here at taxation of income from
labor, will have to state rather complicated patterned historical principles of distributive justice; since end-state principles would not
distinguish sources of income in any way. It is enough for now to get
away from end-state principles and to make clear how various patterned principles are dependent upon particular views about the
sources or the illegitimacy or the lesser legitimacy of profits, interest,
etc.; which particular views may well be mistaken.
What sort of right over others does a legally institutionalized endstate pattern give one? The central core of the notion of a property
right in X, relative to which other parts of the notion are to be explained, is the right to determine what shall be done with X; the right
to choose which of the constrained set of options concerning X shall
be realized or attempted.17 The constraints are set by other principles
or laws operating in the society; in our theory by the Lockean rights
people possess (under the minimal state). My property rights in my
knife allow me to leave it where I will, but not in your chest. I may
choose which of the acceptable options involving the knife is to be
realized. This notion of property helps us to understand why earlier
theorists spoke of people as having property in themselves and their
labor. They viewed each person as having a right to decide what would
become of himself and what he would do, and as having a right to
reap the benefits of what he did.
This right of selecting the alternative to be realized from the constrained set of alternatives may be held by an individual or by a group
17. On the themes in this and the next paragraph, see the writings of Armen
Alchian.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
68 Philosophy & Public Affairs
with some procedure for reaching a joint decision; or the right may be
passed back and forth, so that one year I decide what’s to become of X,
and the next year you do (with the alternative of destruction, perhaps,
being excluded). Or, during the same time period, some types of decisions about X may be made by me, and others by you. And so on.
We lack an adequate, fruitful, analytical apparatus for classifying
the types of constraints on the set of options among which choices
are to be made, and the types of ways decision powers can be held,
divided, and amalgamated. A theory of property would, among other
things, contain such a classification of constraints and decision modes,
and from a small number of principles would follow a host of interesting statements about the consequences and effects of certain combinations of constraints and modes of decision.
When end-result principles of distributive justice are built into the
legal structure of a society, they (as do most patterned principles) give
each citizen an enforcible claim to some portion of the total social
product; that is, to some portion of the sum total of the individually
and jointly made products. This total product is produced by individuals laboring, using means of production others have saved to bring
into existence, by people organizing production or creating means to
produce new things or things in a new way. It is on this batch of individual activities that patterned distributional principles give each
individual an enforcible claim. Each person has a claim to the activities and the products of other persons, independently of whether the
other persons enter into particular relationships that give rise to these
claims, and independently of whether they voluntarily take these
claims upon themselves, in charity or in exchange for something.
Whether it is done through taxation on wages or on wages over a
certain amount, or through seizure of profits, or through there being a
big social pot so that it’s not clear what’s coming from where and
what’s going where, patterned principles of distributive justice involve
appropriating the actions of other persons. Seizing the results of someone’s labor is equivalent to seizing hours from him and directing him
to carry on various activities. If people force you to do certain work,
or unrewarded work, for a certain period of time, they decide what
you are to do and what purposes your work is to serve apart from your
decisions. This process whereby they take this decision from you
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
69 Distributive Justice
makes them a part owner of you; it gives them a property right in you.
Just as having such partial control and power of decision, by right,
over an animal or inanimate object would be to have a property right
in it.
End-state and most patterned principles of distributive justice institute (partial) ownership by others of people and their actions and
labor. These principles involve a shift from the classical liberals’
notion of self-ownership to a notion of (partial) property rights in
other people.
Considerations such as these confront end-state and other patterned
conceptions of justice with the question of whether the actions necessary to achieve the selected pattern don’t themselves violate moral
side-constraints. Any view holding that there are moral side-constraints
on actions, that not all moral considerations can be built into end-states
that are to be achieved,18 must face the possibility that some of its
goals are not achievable by any morally permissible available means.
An entitlement theorist will face such conflicts in a society that deviates from the principles of justice for the generation of holdings, if
and only if the only actions available to realize the principles themselves violate some moral constraints. Since deviation from the first
two principles of justice (in acquisition and transfer) will involve
other persons’ direct and aggressive intervention to violate rights, and
since moral constraints will not exclude defensive or retributive action
in such cases, the entitlement theorist’s problem rarely will be pressing. And whatever difficulties he has in applying the principle of
rectification to persons who did not themselves violate the first two
principles, are difficulties in balancing the conflicting considerations
so as correctly to formulate the complex principle of rectification itself; he will not violate moral side-constraints by applying the principle. Proponents of patterned conceptions of justice, however, often
will face head-on clashes (and poignant ones if they cherish each
party to the clash) between moral side-constraints on how individuals
may be treated on the one hand and, on the other, their patterned
conception of justice that presents an end-state or other pattern that
must be realized.
i8. See Anarchy, State, and Utopia, chap. 3.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
70 Philosophy & Public Affairs
May a person emigrate from a nation that has institutionalized
some end-state or patterned distributional principle? For some principles (e.g., Hayek’s) emigration presents no theoretical problem. But
for others it is a tricky matter. Consider a nation having a compulsory
scheme of minimal social provision to aid the neediest (or one organized so as to maximize the position of the worst off group); no one
may opt out of participating in it. (None may say, “don’t compel me
to contribute to others and don’t provide for me via this compulsory
mechanism if I am in need.”) Everyone above a certain level is forced
to contribute to aid the needy. But if emigration from the country
were allowed, anyone could choose to move to another country that
did not have compulsory social provision but otherwise was (as
much as possible) identical. In such a case, the person’s only motive
for leaving would be to avoid participating in the compulsory scheme
of social provision. And if he does leave, the needy in his initial country will receive no (compelled) help from him. What rationale yields
the result that the person be permitted to emigrate, yet forbidden to
stay and opt out of the compulsory scheme of social provision? If providing for the needy is of overriding importance, this does militate
against allowing internal opting out; but it also speaks against allowing external emigration. (Would it also support, to some extent, the
kidnapping of persons living in a place without compulsory social
provision, who could be forced to make a contribution to the needy
in your community?) Perhaps the crucial component of the position
that allows emigration solely to avoid certain arrangements, while not
allowing anyone internally to opt out of them, is a concern for fraternal feelings within the country. “We don’t want anyone here who
doesn’t contribute, who doesn’t care enough about the others to contribute.” That concern, in this case, would have to be tied to the view
that forced aiding tends to produce fraternal feelings between the
aided and the aider (or perhaps merely to the view that the knowledge
that someone or other voluntarily is not aiding produces unfraternal
feelings).
6. Locke’s Theory of Acquisition. Before we turn to consider another
theory of justice in detail, we must introduce an additional bit of
complexity into the structure of the entitlement theory. This is best
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
71 Distributive Justice
approached by considering Locke’s attempt to specify a principle of
justice in acquisition. Locke views property rights in an unowned object as originating through someone’s mixing his labor with it. This
gives rise to many questions. What are the boundaries of what labor
is mixed with? If a private astronaut clears a place on Mars, has he
mixed his labor with (so that he comes to own) the whole planet, the
whole uninhabited universe, or just a particular plot? Which plot
does an act bring under ownership? The minimal (possibly disconnected) area such that an act decreases entropy in that area, and not
elsewhere? Can virgin land (for the purposes of ecological investigation by high flying airplanes) come under ownership by a Lockean
process? Building a fence around a territory presumably would make
one the owner of only the fence (and the land immediately underneath it).
Why does mixing one’s labor with something make one the owner
of it? Perhaps because one owns one’s labor, and so one comes to own
a previously unowned thing that becomes permeated with what one
owns. Ownership seeps over into the rest. But why isn’t mixing what
I own with what I don’t own a way of losing what I own rather than
a way of gaining what I don’t? If I own a can of tomato juice, and
spill it in the sea so that its molecules (radioactive, so I can check
this) mingle evenly throughout the sea, do I thereby come to own
the sea, or have I foolishly dissipated my tomato juice? Perhaps the
idea, instead, is that laboring on something improves it and makes it
more valuable; and anyone is entitled to own a (thing whose) value
he has created. (Reinforcing this, perhaps, is the view that laboring
is unpleasant. If some people made things effortlessly, as the cartoon
characters in The Yellow Submarine trail flowers in their wake, would
they have lesser claim to their own products whose making didn’t cost
them anything?) Ignore the fact that laboring on something may
make it less valuable (spraying pink enamel paint on a found piece
of driftwood). Why should one’s entitlement extend to the whole
object rather than just to the added value one’s labor has produced?
(Such reference to value might also serve to delimit the extent of
ownership; e.g., substitute “increases the value of'” for “decreases entropy in” in the above entropy criterion.) No workable or coherent
value-added property scheme has yet been devised, and any such
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
72 Philosophy & Public Affairs
scheme presumably would fall to objections (similar to those) that
fell the theory of Henry George.
It will be implausible to view improving an object as giving full
ownership to it, if the stock of unowned objects that might be improved is limited. For an object’s coming under one person’s ownership changes the situation of all others. Whereas previously they were
at liberty (in Hohfeld’s sense) to use the object, they now no longer
are. This change in the situation of others (by removing their liberty
to act on a previously unowned object) need not worsen their situation. If I appropriate a grain of sand from Coney Island, no one else
may now do as they will with that grain of sand. But there are plenty
of others left for them to do the same with. Or if not grains of sand,
then other things. Alternatively, the things I do with the grain of sand
I appropriate might improve the position of others, counterbalancing
their loss of the liberty to use that grain. The crucial point is whether
appropriation of an unowned object worsens the situation of others.
Locke’s proviso that there be “enough and as good left in common
for others” (?27) is meant to ensure that the situation of others is not
worsened. (If this proviso is met, is there any motivation for his
further condition of non-waste?) It is often said that this proviso once
held but now no longer does. But there appears to be an argument for
the conclusion that if the proviso no longer holds, then it cannot ever
have held so as to yield permanent and inheritable property rights.
Consider the first person Z for whom there is not enough and as good
left to appropriate. The last person Y to appropriate left Z without his
previous liberty to act on an object, and so worsened Z’s situation.
So Y’s appropriation is not allowed under Locke’s proviso. Therefore
the next to last person X to appropriate left Y in a worse position, for
X’s act ended permissible appropriation. Therefore X’s appropriation
wasn’t permissible. But then the appropriator two from last, W, ended
permissible appropriation and so, since it worsened X’s position, W’s
appropriation wasn’t permissible. And so on back to the first appropriator A of a permanent property right.
This argument, however, proceeds too quickly. Someone may be
made worse off by another’s appropriation in two ways: first, by losing
the opportunity to improve his situation by a particular appropriation
or any one; and second, by no longer being able to use freely (without
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
73 Distributive Justice
appropriation) what he previously could. A stringent requirement
that another not be made worse off by an appropriation would exclude the first way if nothing else counterbalances the diminution in
opportunity, as well as the second. A weaker requirement would exclude the second way though not the first. With the weaker requirement, we cannot zip back so quickly from Z to A, as in the above argument; for though person Z can no longer appropriate, there may
remain some for him to use as before. In this case Y’s appropriation
would not violate the weaker Lockean condition. (With less remaining that people are at liberty to use, users might face more inconvenience, crowding, etc; in that way the situation of others might be
worsened, unless appropriation stopped far short of such a point.) It
is arguable that no one legitimately can complain if the weaker provision is satisfied. However, since this is less clear than in the case
of the more stringent proviso, Locke may have intended this stringent
proviso by “enough and as good” remaining, and perhaps he meant
the non-waste condition to delay the end point from which the argument zips back.
Is the situation of persons who are unable to appropriate (there
being no more accessible and useful unowned objects) worsened by
a system allowing appropriation and permanent property? Here enter the various familiar social considerations favoring private property: it increases the social product by putting means of production
in the hands of those who can use them most efficiently (profitably);
experimentation is encouraged, because with separate persons controlling resources, there is no one person or small group whom someone with a new idea must convince to try it out; private property enables people to decide on the pattern and types of risks they wish to
bear, leading to specialized types of risk bearing; private property
protects future persons by leading some to hold back resources from
current consumption for future markets; it provides alternate sources
of employment for unpopular persons who don’t have to convince any
one person or small group to hire them, and so on. These considerations enter a Lockean theory to support the claim that appropriation
of private property satisfies the intent behind the “enough and as good
left over” proviso, not as a utilitarian justification of property. They
enter to rebut the claim that because the proviso is violated, no natural
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
74 Philosophy & Public Affairs
right to private property can arise by a Lockean process. The difficulty
in working such an argument to show the proviso is satisfied is in fixing the appropriate baseline for comparison. Lockean appropriation
makes people no worse off than they would be how?19 This question
of fixing the baseline needs more detailed investigation than we are
able to give it here. It would be desirable to have an estimate of the general economic importance of original appropriation for a society, in
order to see how much leeway there is for differing theories of appropriation and of the location of the baseline. Perhaps this importance
can be measured by the percentage of all income that is based upon
untransformed raw materials and given resources (rather than human actions), mainly rental income representing the unimproved
value of the land, and the price of raw materials in situ, and by the
percentage of current wealth that represents such income in the
past.19a
We should note that it is not only persons favoring private property
who need a theory of how property rights legitimately originate. Those
believing in collective property-for example, those believing that a
group of persons living in an area jointly own the territory, or its
mineral resources-also must provide a theory of how such property
rights arise, of why the persons living there have rights to determine
what is done with the land and resources there that persons living
elsewhere don’t have (with regard to the same land and resources).
7. The Proviso. Whether or not Locke’s particular theory of appropriation can be spelled out so as to handle various difficulties, I assume
that any adequate theory of justice in acquisition will contain a proviso similar to the weaker of the ones we have attributed to Locke. A
process normally giving rise to a permanent bequeathable property
ig. Compare Section n of Robert Paul Wolff’s “A Refutation of Rawls’
Theorem on Justice,” Journal of Philosophy 63 (March I966): 179-190. Wolff’s
criticism does not apply to Rawls’ conception under which the baseline is fixed
by the difference principle.
iga. I have not seen a precise estimate. David Friedman discusses this issue
(The Machinery of Freedom [Harper and Row, 1973], pp. xiv, xv) and suggests
one twentieth (of national income) as an upper limit for the first two factors
mentioned. However, he does not attempt to estimate the percentage of current
wealth that is based upon such income in the past.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
75 Distributive Justice
right in a previously unowned thing, will not do so if the position of
others no longer at liberty to use the thing is thereby worsened. It is
important to specify this particular mode of worsening the situation
of others, for the proviso does not encompass other modes. It does not
include the worsening due to more limited opportunities to appropriate
(the first way above, corresponding to the more stringent condition),
and it does not include how I “worsen” a seller’s position if I appropriate materials to make some of what he is selling, and enter into competition with him. Someone whose appropriation (otherwise) would
violate the proviso still may appropriate provided he compensates the
others so that their situation is not thereby worsened; unless he does
compensate these others, his appropriation will violate the (proviso
of the) principle of justice in acquisition and will be an illegitimate
one.20 A theory of appropriation incorporating this Lockean proviso
will handle correctly the cases (objections to the theory lacking the
proviso) where someone appropriates the total supply of something
necessary for life.21
A theory that includes this proviso in its principle of justice in acquisition, also must contain a more complex principle of justice in
transfer. Some reflection of the proviso about appropriation constrains
later actions. If my appropriating all of a certain substance violates
the Lockean proviso, then so does my appropriating some and purchasing all the rest from others who obtained it without (otherwise)
violating the Lockean proviso. If the proviso excludes someone’s appropriating all the drinkable water in the world, it also excludes his
purchasing it all. (More weakly, and messily, it may exclude his
charging certain prices for some of his supply.) This proviso (almost?)
20. Fourier held that since the process of civilization had deprived the members of society of certain liberties (to gather, pasture, engage in the chase), a
socially guaranteed minimum provision for persons was justified as compensation for the loss (Alexander Gray, The Socialist Tradition [New York, i968],
p. i88). But this puts the point too strongly. This compensation would be due
those persons, if any, for whom the process of civilization was a net loss, for
whom the benefits of civilization did not counterbalance being deprived of these
particular liberties.
21. For example, Rashdall’s case of someone who comes upon the only water
in the desert several miles ahead of others who also will come to it, and appropriates it all. Hastings Rashdall, “The Philosophical Theory of Property,” in
Property, its Duties and Rights (London, I9I5).
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
76 Philosophy & Public Affairs
never will come into effect; the more someone acquires of a scarce
substance that others want, the higher the price of the rest will go,
and the more difficult it will become for him to acquire it all. But still,
we can imagine, at least, that something like this occurs: someone
makes simultaneous secret bids to the separate owners of a substance,
each of whom sells assuming he can easily purchase more from the
other owners; or some natural catastrophe destroys all of the supply
of something except that in one person’s possession. The total supply
could not be all permissibly appropriated by one person at the beginning. His later acquisition of it all does not show that the original
appropriation violated the proviso (even by a reverse argument similar to the one above that tried to zip back from Z to A). Rather, it is
the combination of the original appropriation plus all the later transfers and actions that violates the Lockean proviso.
Each owner’s title to his holding includes the historical shadow of
the Lockean proviso in appropriation. This excludes his transferring
it into an agglomeration that does violate the Lockean proviso, and
excludes his using it in a way, in coordination with others or independently of them, so as to violate the proviso by making the situation
of others worse than their baseline situation. Once it is known that
someone’s ownership runs afoul of the Lockean proviso, there are
stringent limits on what he may do with (what it is difficult any
longer unreservedly to call) “his property.” Thus a person may not
appropriate the only water hole in a desert and charge what he will.
Nor may he charge what he will if he possesses one, and unfortunately
it chances that all the water holes in the desert dry up, except for his.
This unfortunate circumstance, admittedly no fault of his, brings into
operation the Lockean proviso and limits his property rights.22 Similarly, an owner’s property right in the only island in an area does not
allow him to order a castaway from a shipwreck off his island as a
trespasser, for this would violate the Lockean proviso.
Notice that the theory does not say that owners do have these rights
22. The situation would be different if his water hole didn’t dry up, due to
special precautions he took to prevent this. Compare our discussion of the case
in the text with Hayek’s, The Constitution of Liberty, p. 136; and also with
Ronald Hamowy’s “Hayek’s Concept of Freedom; A Critique,” New Individualist
Review (April I96I): 28-3 I.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
77 Distributive Justice
but that the rights are overridden to avoid some catastrophe. (Overridden rights do not disappear; they leave a trace of a sort absent in
the cases under discussion).23 There is no such external (and ad hoc?’)
overriding. Considerations internal to the theory of property itself, to
its theory of acquisition and appropriation, provide the means for
handling such cases. The results, however, may be coextensive with
some condition about catastrophe, since the baseline for comparison
is so low as compared to the productiveness of a society with private
appropriation, that the question of the Lockean proviso being violated
arises only in the case of catastrophe (or a desert-island situation).
The fact that someone owns the total supply of something necessary
for others to stay alive, does not entail that his (or anyone’s) appropriation of anything left some people (immediately or later) in a situation worse than the baseline one. A medical researcher who synthesizes a new substance that effectively treats a certain disease and
who refuses to sell except on his terms, does not worsen the situation
of others by depriving them of whatever he has appropriated. The
others (easily can) possess the same materials he appropriated; the
researcher’s appropriation or purchase of chemicals didn’t make those
chemicals scarce in a way so as to violate the Lockean proviso. Nor
would someone else’s purchasing the total supply of the synthesized
substance from the medical researcher. The fact that the medical
researcher uses easily available chemicals to synthesize the drug no
more violates the Lockean proviso than does the fact that the only
surgeon able to perform a particular operation eats easily obtainable
food in order to stay alive and have the energy to work. This shows
that the Lockean proviso is not an “end-state principle”; it focuses on
a particular way that appropriative acts affect others, and not on
the structure of the situation that results.
Intermediate between someone who takes all of the public supply,
and someone who makes the total supply out of easily obtainable substances, is someone who appropriates the total supply of something
in a way that does not deprive the others of it. For example, someone
finds a new substance in an out of the way place. He discovers that
23. I discuss overriding and its moral traces in “Moral Complications and
Moral Structures,” Natural Law Forum I3 (I968): 1-50.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
78 Philosophy & Public Affairs
it effectively treats a certain disease, and appropriates the total supply.
He does not worsen the situation of others; if he did not stumble upon
the substance no one else would have, and the others would remain
without it. However, as time passes, the likelihood increases that
others would have come across the substance; upon this fact might
be based a limit to his property right in the substance so that others
are not below their baseline position, e.g., its bequest might be limited.
The theme of someone worsening another’s situation by depriving
him of something he otherwise would possess, may also illuminate
the example of patents. An inventor’s patent does not deprive others
of an object which would not exist if not for the inventor. Yet patents
would have this effect on others who independently invent the object.
Therefore, these independent inventors, upon whom the burden of
proving independent discovery may rest, should not be excluded from
utilizing their own invention as they wish (including selling it to
others). Furthermore, a known invention drastically lessens the
chances of actual independent invention. For persons who know of
an invention usually will not try to reinvent it, and the notion of independent discovery here would be murky at best. Yet we may assume that in the absence of the original invention, sometime later
someone else would have come up with it. This suggests placing a
time limit on patents, as a rough rule of thumb to approximate how
long it would have taken, in the absence of knowledge of the invention, for independent discovery.
I believe that the free operation of a market system will not actually
run afoul of the Lockean proviso. If this is correct, the proviso will not
provide a significant opportunity for future state action. Indeed, were
it not for the effects of previous illegitimate state action, people would
not think the possibility of the proviso’s being violated as of more
interest than any other logical possibility. (Here I make an empirical
historical claim; as does someone who disagrees with this.) This completes our indication of the complication in the entitlement theory
introduced by the Lockean proviso.
II. RAWLS’ THEORY
We can bring our discussion of distributive justice into sharper focus
by considering in some detail John Rawls’ recent contribution to the
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
79 Distributive Justice
subject. A Theory of Justice24 is a powerful, deep, subtle, wide-ranging,
systematic work in political and moral philosophy which has not seen
its like since the writings of John Stuart Mill, if then. It is a fountain
of illuminating ideas, integrated together into a lovely whole. 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 alternative conception. Even those who remain unconvinced
after wrestling with Rawls’ systematic vision will learn much from
its close study. I do not speak only of the Millian sharpening of one’s
views in combatting (what one takes to be) error. It is impossible
to read Rawls’ book without incorporating much, perhaps transmuted, into one’s own deepened view. And it is impossible to finish
his book without a new and inspiring vision of what a moral theory
may attempt to do and unite, of how beautiful a whole theory can be.
I permit myself to concentrate here on disagreements with Rawls’
theory only because I am confident that my readers will have discovered for themselves its many virtues.
i. Social Cooperation. In considering the role of the principles of
justice, Rawls says:
Let us assume, to fix ideas, that a society is a more or less selfsufficient association of persons who in their relations to one another recognize certain rules of conduct as binding and who for the
most part act in accordance with them. Suppose further that these
rules specify a system of cooperation designed to advance the good
of those taking part in it. Then, although a society is a cooperative
venture for mutual advantage, it is typically marked by a conflict
as well as by an identity of interests. There is an identity of interests since social cooperation makes possible a better life for all
than any would have if each were to live solely by his own efforts.
There is a conflict of interests since persons are not indifferent as
to how the greater benefits produced by their collaboration are
distributed, for in order to pursue their ends they each prefer a
24. (Cambridge, Mass., 1971). Otherwise unidentified references in the text
that follows are to this volume.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
8o Philosophy & Public Affairs
larger to a lesser share. A set of principles is required for choosing
among the various social arrangements which determine this division of advantages and for underwriting an agreement on the
proper distributive shares. These principles are the principles of
social justice: they provide a way of assigning rights and duties in
the basic institutions of society and they define the appropriate distribution of the benefits and burdens of social cooperation (p. 4).
Let us imagine n individuals who do not cooperate together and
who each live solely by their own efforts. Each person i receives a
payoff, return, income, etc. Si; the sum total of what all the separately
n
acting individuals got is S 2 Si. By cooperating together they can
j I
obtain a larger sum total T. The problem of distributive social justice,
according to Rawls, is how these benefits of cooperation are to be
distributed or allocated. This problem might be conceived of in two
ways: how is the total T to be allocated?; or, how is the incremental
amount due to social cooperation, that is the benefits of social cooperation T-S, to be allocated? The latter formulation assumes that each
individual i receives from the subtotal S of T, his share Si. The two
statements of the problem differ. When combined with the noncooperative distribution of S (each i getting Si), a “fair-looking” distribution of T-S under the second version may not yield a “fair-looking”
distribution of T (the first version). Alternatively, a fair-looking distribution of T may give a particular individual i less than his share St.
(The constraint Ti ? Si on the answer to the first formulation of the
problem, where Ti is the share in T of the ith individual, would exclude
this possibility.) Rawls, without distinguishing these two formulations
of the problem, writes as though his concern is the first one of how
the total sum T is to be distributed. One might claim, to support a
focus on the first issue, that due to the enormous benefits of social
cooperation, the non-cooperative shares Si are so small in comparison
to any cooperative ones T, that they may be ignored in setting up the
problem of social justice. Though we should note that this certainly
is not how people entering into cooperation with each other would
agree to conceive of the problem of dividing up cooperation’s benefits.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
8i Distributive Justice
Why does social cooperation create the problem of distributive justice? Would there be no problem of justice and no need for a theory
of justice, if there was no social cooperation at all, if each person got
his share solely by his own efforts? If we suppose, as Rawls seems to,
that this situation does not raise questions of distributive justice, then
in virtue of what facts about social cooperation do these questions of
justice emerge? What is it about social cooperation that gives rise to
issues of justice? It cannot be said that there will be conflicting claims
only where there is social cooperation; that individuals who produce
independently and (initially) fend for themselves will not make
claims of justice on each other. If there were ten Robinson Crusoes,
each working alone for two years on separate islands, who discovered
each other and the facts of their different allotments by radio communication via transmitters left twenty years earlier, could they not make
claims on each other, supposing it were possible to transfer goods from
one island to the next?25 Wouldn’t the one with least make a claim on
ground of need, or on the ground that his island was naturally poorest,
or on the ground that he was naturally least capable of fending for
himself? Mightn’t he say that justice demanded he be given some
more by the others, claiming it unfair that he should receive so much
less and perhaps be destitute, perhaps starving? He might go on to
say that the different individual noncooperative shares stem from differential natural endowments, which are not deserved, and that the
task of justice is to rectify these arbitrary facts and inequities. Rather
than its being the case that one will make such claims in the situation
lacking social cooperation, perhaps the point is that such claims clearly
would be without merit. Why would they clearly be without merit? In
the social noncooperation situation, it might be said, each individual
deserves what he gets unaided by his own efforts; or rather, no one else
can make a claim of justice against this holding. It is pellucidly clear in
this situation who is entitled to what, so no theory of justice is needed.
On this view social cooperation introduces a muddying of the waters
that makes it unclear or indeterminate who is entitled to what. Rather
than saying that no theory of justice applies to this noncooperative
case (wouldn’t it be unjust if someone stole another’s products in the
25. See Milton Friedman, Capitalism and Freedom (Chicago, I962), p. I65.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
82 Philosophy & Public Affairs
noncooperative situation?), I would say that it is a clear case of application of the correct theory of justice: the entitlement theory.
How does social cooperation change things so that the same entitlement principles that apply to the noncooperative cases become inapplicable or inappropriate to cooperative ones? It might be said that
one cannot disentangle the contributions of distinct individuals who
cooperate; everything is everyone’s joint product. On this joint product, or on any portion of it, each person plausibly will make claims of
equal strength; all have an equally good claim, or at any rate no person has a distinctly better claim than any other. Somehow (this line
of thought continues), it must be decided how this total product of
joint social cooperation (to which individual entitlements do not apply differentially) is to be divided up: this is the problem of distributive justice.
Don’t individual entitlements apply to parts of the cooperatively
produced product? First, suppose that social cooperation is based upon
division of labor, specialization, comparative advantage, and exchange; each person works singly to transform some input he receives,
contracting with others who further transform or transport his product until it reaches its ultimate consumer. People cooperate in making
things but they work separately; each person is a miniature firm.26 The
products of each person are easily identifiable, and exchanges are
made in open markets with prices set competitively, given informational constraints, etc. In such a system of social cooperation, what is
the task of a theory of justice? It might be said that whatever holdings
result will depend upon the exchange ratios or prices at which exchanges are made, and therefore that the task of a theory of justice
is to set criteria for “fair prices.” This is hardly the place to trace the
serpentine windings of theories of a just price. It is difficult to see why
these issues should even arise here. People are choosing to make exchanges with other people and to transfer entitlements, with no re26. On the question of why the economy contains firms (of more than one
person), and each individual does not contract and recontract with others, see
Ronald H. Coase, “The Nature of the Firm,” reprinted in Readings in Price
Theory, eds. George Stigler and Kenneth Boulding (Homewood, Ill., 1952); and
Armen A. Alchian and Harold Demsetz, “Production, Information Costs and
Economic Organization,” American Economic Review, 1972.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
83 Distributive Justice
strictions on their freedom to trade with any other party at any mutually acceptable ratio.27 Why does such sequential social cooperation,
linked together by people’s voluntary exchanges, raise any special
problems about how things are to be distributed? Why isn’t the appropriate (a not inappropriate) set of holdings just the one which actually
occurs via this process of mutually agreed to exchanges whereby
people choose to give to others what they are entitled to give or hold?
Let us now drop our assumption that people work independently,
cooperating only in sequence via voluntary exchanges, and instead
consider people who work together jointly to produce something. Is
it now impossible to disentangle people’s respective contributions?
The question here is not whether marginal productivity theory is an
appropriate theory of fair or just shares, but whether there is some
coherent notion of identifiable marginal product. It seems unlikely
that Rawls’ theory rests on the strong claim that there is no such
reasonably serviceable notion. Anyway, once again we have a situation
of a large number of bilateral exchanges: owners of resources reaching separate agreements with entrepreneurs about the use of their
resources, entrepreneurs reaching agreements with individual workers, or groups of workers first reaching some joint agreement and
then presenting a package to an entrepreneur, etc. People transfer
their holdings or labor in free markets, with the exchange ratios
(prices) determined in the usual manner. If marginal productivity
theory is reasonably adequate, people will be receiving, in these voluntary transfers of holdings, (roughly) their marginal products.28
27. We do not, however, assume here or elsewhere the satisfaction of those
conditions specified in economists’ artificial model of “perfect competition.” One
appropriate mode of analysis is presented in Israel M. Kirzner, Market Theory
and the Price System (Princeton, N.J., I963).
28. Receiving this, we should note, is not the same as receiving the equivalent
of what the person causes to exist, or produces. The marginal product of a unit
of F1 with respect to factor F2, . . . , F. is a subjunctive notion; it is the difference
between the total product of F1, . . . , F. used most efficiently (as efficiently as
known how, given prudence about many costs in finding out the most efficient
use of factors), and the total product of the most efficient use of F2, . . ., F,
along with a unit less of F1. But these two different most efficient uses of F2,
. . . , F. along with a unit less of F1 (one with the additional unit of F1, the other
without it) will use them differently. And F1’s marginal product (with respect
to the other factors), what everyone reasonably would pay for an additional
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
84 Philosophy & Public Affairs
But if the notion of marginal product were so ineffective that factors’ marginal products in actual situations of joint production could
not be identified by hirers or purchasers of the factors, then the resulting distribution to factors would not be patterned in accordance with
marginal product. Someone who viewed marginal productivity theory,
where it was applicable, as a patterned theory of justice, might think
such situations of joint production and indeterminate marginal product provided an opportunity for some theory of justice to enter to determine appropriate exchange ratios. But an entitlement theorist
would find acceptable whatever distribution resulted from the party’s
voluntary exchanges.29 The questions about the workability of marginal productivity theory are intricate ones.30 Let us merely note here the
strong personal incentive for owners of resources to converge to the
marginal product, and the strong market pressures tending to produce
this result. Employers of factors of productions are not all dolts who
don’t know what they’re doing, transferring holdings they value to
others on an irrational and arbitrary basis. Indeed, Rawls’ position on
inequalities requires that separate contributions to joint products be
isolable, to some extent. For Rawls goes out of his way to argue that
inequalities are justified if they serve to raise the position of the worstoff group in the society; if without the inequalities the worst-off group
unit of F1, will not be what it causes (it causes) combined with F2, . . . , F and
the other units of F1, but rather the difference it makes, the difference there
would be if this unit of F1 were absent and the remaining factors were organized
most efficiently to cope with its absence. Thus marginal productivity theory is
not best thought of as a theory of actual produced product, of those things whose
causal pedigree includes the unit of the factor; but rather as a theory of the
difference (subjunctively defined) made by the presence of a factor. If such a
view were connected with justice, it would seem to fit best with an entitlement
conception.
29. Readers who believe that Marx’s analysis of exchange relations between
owners of capital and laborers undercuts the view that the set of holdings which
results from voluntary exchange is legitimate, or who believe it a distortion to
term such exchanges “voluntary,” will find some relevant considerations adduced
in Anarchy, State, and Utopia, chap. 8.
30. See Marc Blaug, Economic Theory in Retrospect, chapter ii, and the
references cited therein. For a recent survey of issues about the marginal productivity of capital, see G. C. Harcourt, “Some Cambridge Controversies in the
Theory of Capital,” Journal of Economic Literature 7, no. 2 (June I969):
369-405.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
85 Distributive Justice
would be even more worse off. These serviceable inequalities stem, at
least in part, from the necessity to provide incentives to certain people
to perform various activities or fill various roles that not everyone can
do equally well. (Rawls is not imagining that inequalities are needed
to fill positions that everyone can do equally well, or that the most
drudgery-filled positions that require the least skill will command the
highest income.) But to whom are the incentives to be paid? To which
performers of what activities? When it is necessary to provide incentives to some to perform their productive activities, there is no talk
of a joint social product from which no individual’s contribution can
be disentangled. If the product was all that inextricably joint, it
couldn’t be known that the extra incentives were going to the crucial
persons; and it couldn’t be known that the additional product produced
by these now motivated people is greater than the expenditure to them
in incentives. So it couldn’t be known whether the provision of incentives was efficient or not, whether it involved a net gain or a net loss.
But Rawls’ discussion of justifiable inequalities presupposes that
these things can be known. And so the claim we have imagined about
the indivisible nonpartitionable nature of the joint product is seen
to dissolve, leaving the reasons for the view that social cooperation
creates special problems of distributive justice otherwise not present,
unclear if not mysterious.
2. Terms of Cooperation and the Difference Principle. Another entry into the issue of the connection of social cooperation with distributive shares brings us to grips with Rawls’ actual discussion. Rawls
imagines rational, mutually disinterested, individuals meeting in a
certain situation, or abstracted from their other features not provided
for in this situation. In this hypothetical situation of choice, which
Rawls calls “the original position,” they choose the first principles of
a conception of justice that is to regulate all subsequent criticism and
reform of their institutions. While making this choice, no one knows
his place in society, his class position or social status, or his natural
assets and abilities, his strength, intelligence, etc.
The principles of justice are chosen behind a veil of ignorance. This
ensures that no one is advantaged or disadvantaged in the choice
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
86 Philosophy & Public Affairs
of principles by the outcome of natural chance or the contingency
of social circumstances. Since all are similarly situated and no one
is able to design principles to favor his particular condition,
the principles of justice are the result of a fair agreement or
bargain (?3).
What would persons in the original position agree to?
Persons in the initial situation would choose two . . . principles:
the first requires equality in the assignment of basic rights and
duties, while the second holds that social and economic inequalities,
for example, inequalities of wealth and authority are just only if
they result in compensating benefits for everyone, and in particular
for the least advantaged members of society. These principles rule
out justifying institutions on the grounds that the hardships of
some are offset by a greater good in the aggregate. It may be expedient but it is not just that some should have less in order that
others may prosper. But there is no injustice in the greater benefits
earned by a few provided that the situation of persons not so fortunate is thereby improved. The intuitive idea is that since everyone’s
well-being depends upon a scheme of cooperation without which
no one could have a satisfactory life, the division of advantages
should be such as to draw forth the willing cooperation of everyone
taking part in it, including those less well situated. Yet this can be
expected only if reasonable terms are proposed. The two principles
mentioned seem to be a fair agreement on the basis of which those
better endowed, or more fortunate in their social position, neither
of which we can be said to deserve, could expect the willing cooperation of others when some workable scheme is a necessary condition
of the welfare of all ( ?3 ).
This second principle, which Rawls specifies as the difference principle, holds that the institutional structure is to be so designed that
the worst-off group under it is at least as well off as the worst-off
group (not necessarily the same group) would be under any alternative institutional structure. If persons in the original position follow
the minimax policy in making the significant choice of principles of
justice, Rawls argues, they will choose the difference principle. Our
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
87 Distributive Justice
concern here is not whether persons in the position Rawls describes
actually would minimax and actually would choose the particular
principles Rawls specifies. Still, we should question why individuals
in the original position would choose a principle that focuses upon
groups, rather than individuals. Won’t application of the minimax
principle lead each person in the original position to favor maximizing
the position of the worst-off individual? To be sure, this principle
would reduce questions of evaluating social institutions to the issue
of how the unhappiest depressive fares. Yet avoiding this by moving
the focus to groups (or representative individuals) seems ad hoc, and
is inadequately motivated for those in the original position (see
p. 98 and ? i 6 generally). Nor is it clear which groups are appropriately considered; why exclude the group of depressives or alcoholics
or the representative paraplegic?
If the difference principle is not satisfied by some institutional structure J, then under J some group G is worse off than it would be under
another institutional structure I that satisfies the principle. If another
group F is better off under J than it would be under the I favored by
the difference principle, is this sufficient to say that under J “some . . .
have less in order that others may prosper”? (Here one would have in
mind that G has less in order that F prosper. Could one also make the
same statement about I? Does F have less under I in order that G may
prosper?) Suppose that in a society
(i) Group G has amount A and group F has amount B, with B
greater than A. Also things could be arranged differently so
that G would have more than A, and F would have less than B.
(The different arrangement might involve a mechanism to
transfer some holdings from F to G.)
Is this sufficient to say
(2) G is poorly off because F is well off; G is poorly off in order that
F be well off; F’s being well off makes G poorly off; G is poorly
off on account of F’s being well off; G is not better off because
of how well off F is?
If so, does the truth of (2) depend on G’s being in a worse position than
F? There is yet another possible institutional structure K that transfers
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
88 Philosophy & Public Affairs
holdings from the worse-off group G to F, making G even more worse
off. Does the possibility of K make it true to say that under J, F is not
(even) better off because of how well off G is?
We do not normally hold that the truth of a subjunctive as in (i)
is alone sufficient for the truth of some indicative causal statement as
in (2). It would improve my life in various ways if you were to choose
to become my devoted slave, supposing I could get over the initial
discomfort. Is the cause of my present state your not becoming my
slave? Because your enslaving yourself to a poorer person would improve his lot and worsen yours, are we to say that that poor person
is badly off because you are as well off as you are; has he less in order
that you may prosper? From
(3) If P were to do act A then Q would not be in situation S,
we will conclude
(4) P’s not doing A is responsible for Q’s being in situation S; P’s
not doing A causes Q to be in S
only if we also believe that
(5) P ought to do act A, or P has a duty to do act A, or P has an
obligation to do act A, etc.3′
Thus the inference from (3) to (4), in this case, presupposes (5).
One cannot argue from (3) to (4) as one step in order to get to (5).
The statement that in a particular situation some have less in order
that others may prosper is often based upon the very evaluation of
a situation or an institutional framework that it is introduced to support. Since this evaluation does not follow merely from the subjunctive
(e.g., [I] or [31) an independent argument must be produced for it.32
31. Here we simplify the content of (5), but not to the detriment of our present discussion. Also, of course, beliefs other than (5), when conjoined with (3)
would justify the inference to (4); for example belief in the material conditional
“If (3) then (a).” It is something like (5), though, that is relevant to our discussion here.
32. Though Rawls does not clearly distinguish (2) from (i) and (4) from
(3), I do not claim that he makes the illegitimate step of sliding from the latter
subjunctive to the former indicative. Even so, the mistake is worth pointing out
because it is an easy one to fall into, and it might appear to prop up positions we
argue against.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
89 Distributive Justice
Rawls holds, as we have seen, that
since everyone’s well-being depends upon a scheme of cooperation
without which no one could have a satisfactory life, the division of
advantages should be such as to draw forth the willing cooperation
of everyone taking part in it, including those less well situated. Yet
this can be expected only if reasonable terms are proposed. The two
principles mentioned seem to be a fair agreement on the basis of
which those better endowed or more fortunate in their social position … could expect the willing cooperation of others when some
workable scheme is a necessary condition of the welfare of all (?3).
No doubt, the difference principle presents terms on the basis of which
those less well endowed would be willing to cooperate. (What better
terms could they propose for themselves? ) But is this a fair agreement
on the basis of which those worse endowed could expect the willing
cooperation of others? With regard to the existence of gains from
social cooperation, the situation is symmetrical. The better endowed
gain by cooperating with the worse endowed, and the worse endowed
gain by cooperating with the better endowed. Yet the difference principle is not neutral between the better and the worse endowed. Whence
the asymmetry?
Perhaps the symmetry is upset if one asks how much each gains
from the social cooperation. This question might be understood in two
ways: How much do people benefit from social cooperation, as compared to their individual holdings in a non-cooperative scheme? That
is, how much is TrSt, for each individual i? Or, alternatively, how
much does each individual gain from general social cooperation, as
compared (not with no cooperation but) with more limited cooperation? The latter is the more appropriate question with regard to general social cooperation. For failing general agreement on the principles
to govern how the benefits of general social cooperation are to be held,
not everyone will remain in a noncooperative situation if there is some
other beneficial cooperative arrangement involving some (but not all)
people, whose participants can agree. These people will participate in
this more narrow cooperative arrangement. To focus upon the benefits
of the better and the worse endowed cooperating together, we must try
to imagine less extensive schemes of partitioned social cooperation in
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
go Philosophy & Public Affairs
which the better endowed cooperate only among themselves and the
worse endowed cooperate only among themselves, with no crosscooperation. The members of both groups gain from the internal
cooperation within their respective groups, and have larger shares
than they would if there were no social cooperation at all. An individual benefits from the wider system of extensive cooperation between
the better and the worse endowed, to the extent of his incremental
gain from this wider cooperation; the amount by which his share
under a scheme of general cooperation is greater than it would be
under one of limited intragroup (but not cross-group) cooperation.
General cooperation will be of more benefit to the better or to the worse
endowed if (to pick a simple criterion) the mean incremental gain
from general cooperation (over against limited intragroup cooperation) is greater in one group than it is in the other.
One might speculate about whether there is an inequality between
the groups’ mean incremental gains and, if so, which way it goes. If the
better endowed includes those who manage to accomplish something
of great economic advantage to others, such as new inventions, new
ideas about production or ways of doing things, skill at economic
tasks, etc.,33 it is difficult to avoid concluding that the less well endowed gain more than the better endowed do from the scheme of
33. They needn’t be better endowed, from birth. In the context in which Rawls
uses it, all “better endowed” means is: accomplishes more of economic value,
able to do this, has a high marginal product, etc. (The role played in this by
unpredictable factors complicates imagining a prior partitioning of the two
groups.) The text follows Rawls in categorizing persons as “better” and “worse”
endowed only in order to criticize the considerations he adduces for his theory.
The entitlement theory does not rest upon any assumption that the classification
is an important one or even a possible one, or upon any elitist presupposition.
Since the entitlement theorist does not accept the patterned principle “to each
according to his natural endowment,” he can easily grant that what an exercised
endowment brings in the market will depend upon the endowments of others
and how they choose to exercise them, upon the market-expressed desires of
buyers, upon the alternate supply of what he offers and of what others may substitute for what he offers, and upon other circumstances summing the myriad
choices and actions of others. Similarly, we saw earlier that the similar considerations Rawls adduces about upon what social factors the marginal product
of labor depends (p. 308) will not faze an entitlement theorist, even though
they might undercut the rationale put forth by a proponent of the patterned
principle of distribution according to marginal product.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
9I Distributive Justice
general cooperation. What follows from this conclusion? I do not
mean to imply that the better endowed should get even more than
they get under the entitlement system of general social cooperation.34
What does follow from the conclusion is a deep suspicion of imposing,
in the name of fairness, constraints upon voluntary social cooperation
(and the set of holdings that arises from it) so that those already
benefitting most from this general cooperation benefit even more!
Rawls would have us imagine the worse endowed persons saying
something like the following: “Look, better endowed, you gain by cooperating with us. If you want our cooperation you’ll have to accept
reasonable terms. We suggest these terms: We’ll cooperate with you
only if we get as much as possible. That is, the terms of our cooperation should give us that maximal share such that, if it was tried to
give us more, we’d end up with less.” How generous these proposed
terms are might be seen by imagining that the better endowed make
the (almost) symmetrical opposite proposal: “Look, worse endowed:
you gain by cooperating with us. If you want our cooperation you’ll
have to accept reasonable terms. We propose these terms: We’ll cooperate with you so long as we get as much as possible. That is, the
terms of our cooperation should give us the maximal share such that,
if it was tried to give us more, we’d end up with less.” If these terms
seem outrageous, as they are, why don’t the terms proposed by those
34. Supposing they could identify themselves and each other, they might try
to exact a larger share by banding together as a group and bargaining jointly
with the others. Given the large numbers of persons involved and the incentive
for some of the better endowed individuals to break ranks and reach separate
agreements with the worse endowed, if such a coalition of the better endowed
is unable to impose sanctions on its defectors it will dissolve. The better endowed
remaining in the coalition may use boycott as a “sanction,” and refuse to cooperate with a defector. To break the coalition, those less well endowed would
have to (be able to) offer someone better endowed sufficient incentive to defect
to make up for his loss through no longer being able to cooperate with the other
better endowed persons. Perhaps it would pay for someone to defect from the
coalition only as part of a sizable group of defectors, which defecting group
the initial coalition might try to keep small by special offers to individuals to
defect from it, etc. The problem is a complicated one, further complicated by
the obvious fact (despite our use of Rawls’ classificatory terminology) that there
is no sharp line of cleavage between the endowments of people, to determine
which groups would form.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
92 Philosophy & Public Affairs
worse endowed seem the same? Why shouldn’t the better endowed
treat this latter proposal as beneath consideration, supposing someone
to have the nerve explicitly to state it?
Rawls devotes much attention to explaining why those less well
favored should not complain at receiving less. His explanation, simply
put, is that because the inequality works for his advantage, someone
less well favored shouldn’t complain about it; he receives more in the
unequal system than he would in an equal one. (Though he might
receive still more in another unequal system that placed someone else
below him.) But Rawls discusses the question of whether those more
favored will or should find the terms satisfactory only in the following
passage, where A and B are any two representative men with A being
the more favored:
The difficulty is to show that A has no grounds for complaint. Perhaps he is required to have less than he might since his having
more would result in some loss to B. Now what can be said to the
more favored man? To begin with, it is clear that the well-being
of each depends on a scheme of social cooperation without which
no one could have a satisfactory life. Secondly, we can ask for the
willing cooperation of everyone only if the terms of the scheme are
reasonable. The difference principle, then, seems to be a fair basis
on which those better endowed, or more fortunate in their social
circumstances, could expect others to collaborate with them when
some workable arrangement is a necessary condition of the good
of all (p. I03).
What Rawls imagines being said to the more favored men does not
show that these men have no grounds for complaint, nor does it at
all diminish the weight of whatever complaints they have. That the
well-being of all depends on social cooperation without which no one
could have a satisfactory life could also be said to the less well endowed by someone proposing any other principle, including that of
maximizing the position of the best endowed. Similarly for the fact
that we can ask for the willing cooperation of everyone only if the
terms of the scheme are reasonable. The question is: what terms
would be reasonable? What Rawls imagines being said thus far merely
sets up his problem; it doesn’t distinguish his proposed difference
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
93 Distributive Justice
principle from the (almost) symmetrical counterproposal that we
imagined the better endowed making, or from any other proposal.
Thus, when Rawls continues, “The difference principle, then, seems
to be a fair basis on which those best endowed, or more fortunate in
their social circumstances, could expect others to collaborate with
them when some workable arrangement is a necessary condition of
the good of all,” the presence of the “then” in his sentence is puzzling.
Since the sentences which precede it are neutral between his proposal
and any other proposal, the conclusion that the difference principle
presents a fair basis for cooperation cannot follow from what precedes
it in this passage. Rawls is merely repeating that it seems reasonable;
hardly a convincing reply to anyone to whom it doesn’t seem reasonable.35 Rawls has not shown that the more favored man A has no
35. I treat Rawls’ discussion here as one concerning better and worse endowed individuals who know who they are. Alternatively, one might imagine
that these considerations are to be weighed by someone in the original position.
(“If I turn out to be better endowed then .. .; if I turn out to be worse endowed
then….”) But this construal will not do. Why would Rawls bother saying
“The two principles seem to be a fair agreement on the basis of which those
better endowed or more fortunate in their social position could expect the willing
cooperation of others” (?3). Who is doing the expecting when? How is this to
be translated into subjunctives to be contemplated by someone in the original
position? Similarly, questions arise about Rawls’ saying, “The difficulty is to
show that A has no grounds for complaint. Perhaps he is required to have less
than he might since his having more would result in some loss to B. Now what
can be said to the more favored man? . . . The difference principle then seems
to be a fair basis on which those better endowed . . . could expect others to
collaborate with them . . .” (p. I03, my italics). Are we to understand this as:
someone in the original position wonders what to say to himself as he then
thinks of the possibility that he will turn out to be one of the better endowed?
And does he then say that the difference principle then seems a fair basis for
cooperation despite the fact that and even while he is contemplating the possibility that he is better endowed? Or does he say then that even later, if and
when he knows he is better endowed, the difference principle will seem fair to
him at that later time? And when are we to imagine him possibly complaining?
Not while in the original position, for then he is agreeing to the difference
principle. Nor does he worry, while in the process of deciding in the original
position, that he will complain later. For he knows that he will have no cause
to complain later at the effects of whatever principle he himself rationally will
choose soon in the original position. Are we to imagine him complaining against
himself? And isn’t the answer to any later complaint, “You agreed to it (or you
would have agreed to it if so originally positioned)”? What “difficulty” does
Rawls concern himself with here? Trying to squeeze it into the original position
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
94 Philosophy & Public Affairs
grounds for complaint at being required to have less in order that
another B might have more than he otherwise would. And he can’t
have shown this, since A does have grounds for complaint. Doesn’t he?
3. The Original-Position and End-Result Principles. How can it
have been supposed that these terms offered by the less well endowed
are fair? Imagine a social pie somehow appearing so that no one has
any claim at all on any portion of it, no one has any more of a claim
than any other person; yet there must be unanimous agreement on
how it is to be divided. Undoubtedly, apart from threats or holdouts
in bargaining, an equal distribution would be suggested and found
plausible as a solution. (It is, in Schelling’s sense, a focal point solution.) If somehow the size of the pie wasn’t fixed, and it was realized
that pursuing an equal distribution somehow would lead to a smaller
total pie than otherwise might occur, the people might well agree to
an unequal distribution which raised the size of the least share. But in
any actual situation, wouldn’t this realization reveal something about
differential claims on parts of the pie? Who is it that could make the
makes it completely mysterious. And what is thinking of what is a “fair agreement” (?3) or a “fair basis” (p. I03) doing here anyway, in the midst of the
rational self-interested calculations of persons in the original position, who do
not then knowingly possess, or at any rate utilize, particular moral notions?
I see no coherent way to incorporate how Rawls treats and speaks of the issue
of the terms of cooperation between the better and the worse endowed, into the
structure and perspective of the original position. Therefore my discussion
considers Rawls here as addressing himself to individuals outside the original
position, either to better endowed individuals or to his readers, to convince them
that the difference principle which Rawls extracts from the original position is
fair. It is instructive to compare how Rawls imagines justifying the social order
to a person in the worst-off group in an unequal society. Rawls wants to tell
this person that the inequalities work out to his advantage. This is told to someone who knows who he is. (“The social order can be justified to everyone, and
in particular to those who are least favored” [p. 103].) Rawls does not want
to say “You would have gambled, and you lost” or any such thing, even “you
chose it then in the original position”; nor does he wish merely to address
someone in the original position. He also wants a consideration apart from the
original position that will convince someone who knows of his inferior position
in an unequal society. To say “you have less in order that I may prosper” would
not convince someone who knows of his inferior position, and Rawls rightly
rejects it, even though its subjunctive analogue for someone in the original
position, if we could make sense of this, would not be without force.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
95 Distributive Justice
pie larger, and would do it if given a larger share, but not if given an
equal share under the scheme of equal distribution? To whom is an
incentive to be provided to make this larger contribution? (There’s
no talk here of inextricably entangled joint product; it’s known to
whom incentives are to be offered, or at least, to whom a bonus is to
be paid after the fact.) Why doesn’t this identifiable differential contribution lead to some differential entitlement?
If things fell from heaven like manna, and no one had any special
entitlement to any portion of it, and unless all agreed to a particular
distribution no manna would fall, and somehow the quantity varied
depending on the distribution, then it is plausible to claim that persons
placed so that they couldn’t make threats, or hold out for specially
large shares, would agree to the difference principle rule of distribution. But is this the appropriate model for thinking about how the
things people produce are to be distributed? Why think the same
results should obtain for situations where there are differential entitlements as for situations where there are not?
A procedure that founds principles of distributive justice on what
rational persons who know nothing about themselves or their histories
would agree to, guarantees that end-state principles of justice will be
taken as fundamental. Perhaps some historical principles of justice
are derivable from end-state principles, as the utilitarian tries to derive
individual rights, prohibitions on punishing the innocent, etc., from
his end-state principle; perhaps such arguments even can be constructed for the entitlement principle. But no historical principle, it
seems, could be agreed to in the first instance by the participants in
Rawls’ original position. For people meeting together behind a veil
of ignorance to decide who gets what, knowing nothing about any
special entitlements people may have, will treat anything to be distributed as manna from heaven.36
Suppose there were a group of students who have studied during a
36. Do the people in the original position ever wonder whether they have the
right to decide how everything is to be divided up? Perhaps they reason that
since they are deciding this question, they must assume they are entitled to
do so; and so particular people can’t have particular entitlements to holdings
(for then they wouldn’t have the right to decide together on how all holdings
are to be divided); and hence everything legitimately may be treated like manna
from heaven.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
96 Philosophy & Public Affairs
year, taken examinations, and received grades between o and ioo
which they have not yet learned of. They are now gathered together,
having no idea of the grade any one of them has received, and they
are asked to allocate grades among themselves so that the grades
total to a given sum (which is determined by the sum of the grades
they actually have received from the teacher). First, let us suppose
they are to decide jointly upon a particular distribution of grades; they
are to give a particular grade to each identifiable one of them present
at the meeting. Here, given sufficient restrictions on their ability to
threaten each other, they probably would agree to each person receiving the same grade, to each person’s grade being equal to the total
divided by the number of people to be graded. Surely they would not
chance upon the particular set of grades they already have received.
Suppose next that there is posted on a bulletin board at their meeting
a paper headed ENTITLEMENTS, that lists each person’s name with a
grade next to it, the listing being identical to the instructor’s gradings.
Still, this particular distribution will not be agreed to by those having
done poorly. Even if they know what “entitlement” means (which
perhaps we must suppose they don’t, in order to match the absence
of moral factors in the calculations of persons in Rawls’ original
position), why should they agree to the instructor’s distribution? What
self-interested reason to agree to it would they have?
Next suppose that they are unanimously to agree not to a particular
distribution of grades, but rather to general principles to govern the
distribution of grades. What principle would be selected? The equality
principle, which gives each person the same grade, would have a
prominent chance. And if it turned out that the total was variable depending upon how they divided it, depending on which of them got
what grade, and a higher grade was desirable though they were not
competing among each other (e.g., each of them was competing for
some position with the members of separate distinct groups), then the
principle of distributing grades so as to maximize the lowest grades
might seem a plausible one. Would these people agree to the non-endstate historical principle of distribution: give people grades according
to how their examinations were evaluated by a qualified and impartial
observer?37 If all the people deciding knew the particular distribution
37. I do not mean to assume that all teachers are such, nor even that learnThis content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
97 Distributive Justice
that would be yielded by this historical principle, they wouldn’t agree
to it. For the situation then would be equivalent to the earlier one of
their deciding upon a particular distribution, in which we already have
seen they would not agree to the entitlement distribution. Suppose
then that the people do not know the particular distribution actually
yielded by this historical principle. They cannot be led to select this
historical principle because it looks just, or fair, to them; for no such
notions are allowed to be at work in the original position. (Otherwise
people would argue there, like here, about what justice requires.) Each
person engages in a calculation to decide whether it will be in his
own interests to accept this historical principle of distribution. Grades,
under the historical principle, depend upon nature and developed
intelligence, how hard the people have worked, accident, etc., about
which people in the original position know almost nothing. (It would
be risky for someone to think that since he is reasoning so well in
thinking about the principles, he must be one of the intellectually
better endowed. Who knows what dazzling argument the others are
reasoning their way through, and perhaps keeping quiet about for
strategic reasons.) Each person in the original position will do something like assigning probability distributions to his place along these
various dimensions. It seems unlikely that each person’s probability
calculations would lead to the historical-entitlement principle, in
preference to every other principle. Consider the principle we may
call the reverse-entitlement principle. It recommends drawing up a
list in order of magnitude of the historical entitlements, and giving
the most anyone is entitled to, to the person entitled to the least; the
second most to the person entitled to the second least, and so on.38 Any
probability calculations of self-interested persons in Rawls’ original
ing in universities should be graded. All I need is some example of entitlement,
the details of which the reader will have some familiarity with, to use to examine
decision-making in the original position. Grading is a simple example, though
not a perfect one, entangled as it is with whatever ultimate social purposes the
ongoing practice serves. We may ignore this complication, for their selecting
the historical principle on the grounds that it effectively serves those purposes
would illustrate our point below that their fundamental concerns and fundamental principles are end-state ones.
38. But recall the reasons why using magnitudes of entitlement does not
capture accurately the entitlement principle, presented in footnote 4 above.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
98 Philosophy & Public Affairs
position, or any probability calculations of the students we have considered, will lead them to view the entitlement and the reverse-entitlement principles as ranked equally insofar as their own self-interest is
concerned! (What calculations could lead them to view one of the
principles as superior to the other?) Their calculations will not lead
them to select the entitlement principle.
The nature of the decision problem facing persons deciding upon
principles in an original position behind a veil of ignorance, limits
them to end-state principles of distribution. The self-interested person
evaluates any non-end-state principle on the basis of how it works out
for him; his calculations about any principle focus on how he ends up
under the principle. (These calculations include consideration of the
labor he is yet to do, which does not appear in the grading example
except as the sunk cost of the labor already done.) Thus for any
principle an occupant of the original position will focus on the distribution D of goods that it leads to, or a probability distribution
over the distributions D1 .. . Dn it may lead to, and upon his probabilities of occupying each position in each Di profile supposing it to obtain. The point would remain the same if, rather than using personal
probabilities, he used some other decision rule of the sort discussed
by decision theorists. In these calculations, the only role played by the
principle is that of generating a distribution of goods (or whatever else
they care about) or of generating a probability distribution over distributions of goods. Different principles are compared solely by comparing the alternative distributions they generate. Thus the principles
drop out of the picture, and each self-interested person makes a choice
among alternative end-state distributions. People in the original position either directly agree to an end-state distribution or they agree to
a principle; if they agree to a principle, they do it solely on the basis
of considerations about end-state distributions. The fundamental principles they agree to, the ones they can all converge in agreeing upon,
must be end-state principles.
Rawls’ construction is incapable of yielding an entitlement or historical conception of distributive justice. The end-state principles of
justice yielded by his procedure might be used in an attempt to derive,
when conjoined with factual information, historical-entitlement principles, as derivative principles falling under a nonentitlement conThis content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
99 Distributive Justice
ception of justice.39 It is difficult to see how such attempts could derive
and account for the particular convolutions of historical-entitlement
principles. And any derivations from end-state principles of approximations of the principles of acquisition, transfer, and rectification,
would strike one as similar to utilitarian contortions in trying to derive
(approximations of) usual precepts of justice; they do not yield the
particular result desired, and they produce the wrong reasons for the
sort of result they try to get. If historical-entitlement principles are
fundamental, then Rawls’ construction at best will yield approximations of them; it will produce the wrong sorts of reasons for them, and
its derived results sometimes will conflict with the precisely correct
principles. The whole procedure of persons choosing principles in
Rawls’ original position presupposes that no historical-entitlement
conception of justice is correct.
It might be objected to our argument that Rawls’ procedure is designed to establish all facts about justice; there is no independent
notion of entitlement, not provided by his theory, to stand on in criticizing his theory. But we do not need any particular developed historical-entitlement theory as a basis from which to criticize Rawls’ construction. If any such fundamental historical-entitlement view is
correct, then Rawls’ theory is not. We are thus able to make this structural criticism of the type of theory Rawls presents and the type of principles it must yield, without first having formulated fully a particular
historical-entitlement theory as an alternative to his. We would be illadvised to accept Rawls’ theory and his construal of the problem as
one of which principles would be chosen by rational self-interested
individuals behind a veil of ignorance, unless we were sure that no
adequate historical-entitlement theory was to be gotten.
Since Rawls’ construction doesn’t yield an historical or entitlement
conception of justice, there will be some feature(s) of his construction
in virtue of which it doesn’t. Have we done anything other than focus
upon the particular feature(s), and say that this makes Rawls’ con39. Some years ago, Hayek argued (The Constitution of Liberty, chap. 3:
“The Common Sense of Progress”) that a free capitalist society, over time, raises
the position of those worst off more than any alternative institutional structure;
to use present terminology, he argued that it best satisfies the end-state principle
of justice formulated by the difference principle.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
IOO Philosophy & Public Affairs
struction incapable in principle of yielding an entitlement or historical
conception of justice? This would be a criticism without any force at
all, for in this sense we would have to say that the construction is incapable in principle of yielding any conception other than the one it
actually yields. It seems clear that our criticism goes deeper than this
(and I hope it is clear to the reader); but it is difficult to formulate the
requisite criterion of depth. Lest this appear lame, let us add that as
Rawls states the root idea underlying the veil of ignorance, that feature
which is the most prominent in excluding agreement to an entitlement
conception, it is to prevent someone from tailoring principles to his
own advantage, from designing principles to favor his particular condition. But not only does the veil of ignorance do this; it ensures that
no shadow of entitlement considerations will enter the rational calculations of ignorant nonmoral individuals constrained to decide in
a situation reflecting some formal conditions of morality.40 Perhaps,
in a Rawls-like construction, some condition weaker than the veil of
ignorance could serve to exclude the special tailoring of principles, or
perhaps some other “structural-looking” feature of the choice situation
could be formulated to mirror entitlement considerations. But as it
stands there is no reflection of entitlement considerations in any form
in the situation of those in the original position; these considerations
do not enter even to be overridden or outweighed or otherwise put
aside. Since no glimmer of entitlement principles is built into the
structure of the situation of persons in the original position, there is
no way these principles could be selected; and Rawls’ construction
is incapable in principle of yielding them. This is not to say, of course,
that the entitlement principle (or “the principle of natural liberty”)
couldn’t be written on the list of principles to be considered by those
in the original position. Rawls doesn’t do even this, perhaps because
it is so transparently clear that there would be no point in including
it to be considered there.
40. Someone might think entitlement principles count as specially tailored
in a morally objectionable way, and so reject my claim that the veil of ignorance
accomplishes more than its stated purpose. Since to specially tailor principles
is to tailor them unfairly for one’s own advantage, and since the question of
the fairness of the entitlement principle is precisely the issue, it is difficult to
decide which begs the question: my criticism of the strength of the veil of ignorance, or the defense against this criticism which I imagine in this note.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
iOI Distributive Justice
4. Macro and Micro. We noted above the objection which doubted
whether there is any independent notion of entitlement. This connects
with Rawls’ insistence that the principles he formulates are to be applied only to the fundamental macro-structure of the whole society,
and that no micro-counterexample to them will be admissible. The difference principle is, on the face of it, unfair (though that will be of no
concern to anyone deciding in the original position); and a wide gamut
of counterexamples to it can be produced that focus on small situations
that are easy to take in and manage. But Rawls does not claim the
difference principle is to apply to every situation; only to the basic
structure of the society. How are we to decide if it applies to that?
Since we may have only weak confidence in our intuitions and judgments about the justice of the whole structure of society, we may
attempt to aid our judgment by focusing on micro-situations that we
do have a firm grasp of. For many of us, an important part of the
process of arriving at what Rawls calls “reflective equilibrium” will
consist of thought experiments in which we try out principles in hypothetical micro-situations. If, in our considered judgment, they do not
apply there, then they are not universally applicable. And we may
think that since correct principles of justice are universally applicable,
principles that fail for micro-situations cannot be correct. Since Plato,
at any rate, that has been our tradition; principles may be tried out in
the large and in the small. Plato thought that writ large the principles
are easier to discern; others may think the reverse.
Rawls, however, proceeds as though distinct principles apply to
macro and micro contexts, to the basic structure of society and to
the situations we can take in and understand. Are the fundamental
principles of justice emergent in this fashion, applying (only) to the
largest social structure yet not to its parts? Perhaps one thinks of
the possibility that a whole social structure is just, even though none
of its parts are, because the injustice in each part somehow balances
out or counteracts another one, and the total injustice ends up being
balanced out or nullified. But can a part satisfy the most fundamental
principle of justice yet still clearly be unjust, apart from its failure
to perform any supposed task of counterbalancing another existing
injustice? Perhaps so, if a part involves some special domain. But
surely a regular ordinary every day part, possessing no very unusual
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
102 Philosophy & Public Affairs
features, should turn out to be just when it satisfies the fundamental
principles of justice; else special explanations must be offered. One
cannot say merely that one is speaking of principles to apply only
to the fundamental structure, so that micro-counterexamples do not
tell. In virtue of what features of the basic structure, features not
possessed by micro-cases, do special moral principles apply that would
be unacceptable elsewhere?
There are special disadvantages to proceeding by focusing only
on the intuitive justice of described complex wholes. For complex
wholes are not easily scannable; we cannot easily keep track of everything that is relevant. The justice of a whole society may depend on
its satisfying a number of distinct principles. These principles, though
individually compelling (witness their application to a wide range
of particular micro-cases), may yield surprising results when combined together. That is, one may be surprised at which, and only
which, institutional forms satisfy all the principles. (Compare the
surprise at discovering what, and only what, satisfies a number of
distinct and individually compelling conditions of adequacy; and how
illuminating such discoveries are.) Or perhaps it is one simple principle which is to be writ large, and what things look like when this is
done is very surprising, at first. I am not claiming that new principles
emerge in the large, but that how the old micro-principles turn out
to be satisfied in the large may surprise. If this is so, then one should
not depend upon judgments about the whole as providing the only
or even the major body of data against which to check one’s principles.
One major path to changing one’s intuitive judgments about some
complex whole is through seeing the larger and often surprising implications of principles solidly founded at the micro-level. Similarly,
discovering that one’s judgnents are wrong or mistaken often surely
will involve overturning them by stringent applications of principles
grounded on the micro-level. For these reasons it is undesirable to
attempt to protect principles by excluding micro-tests of them.
The only reason I have thought of for discounting micro-tests of
the fundamental principles is that micro-situations have particular
entitlements built into them. Of course, continues the argument, the
fundamental principles under consideration will run afoul of these
entitlements, for the principles are to operate at a deeper level than
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
103 Distributive Justice
such entitlements. Since they are to operate at the level that underlies
such entitlements, no micro-situation that includes entitlements can
be introduced as an example by which to test these fundamental principles. Note that this reasoning grants that Rawls’ procedure assumes
that no fundamental entitlement view is correct; that it assumes there
is some level so deep that no entitlements operate that far down.
May all entitlements be relegated to relatively superficial levels?
For example, people’s entitlements to the parts of their own bodies?
An application of the principle of maximizing the position of those
worst off might well involve forcible redistribution of bodily parts
(“You’ve been sighted for all these years; now one [or even both] of
your eyes is to be transplanted to others”), or killing some people
early to use their bodies in order to provide material necessary to
save the lives of those who otherwise would die young.41 To bring up
such cases is to sound slightly hysterical. But we are driven to such
extreme examples in examining Rawls’ prohibition on micro-counterexamples. That not all entitlements in micro-cases are plausibly construed as superficial, and hence as illegitimate material by which to
test out suggested principles, is made especially clear if we focus on
those entitlements and rights that most clearly are not socially or
institutionally based. On what grounds are such cases, whose detailed
specifications I leave to the ghoulish reader, ruled inadmissible? On
what grounds can it be claimed that the fundamental principles of
justice need only apply to the fundamental institutional structure of
a society? (And couldn’t we build such redistributive practices concerning bodily parts or the ending of people’s lives, into the fundamental structure of a society?)
It is ironic that we criticize Rawls’ theory for its fundamental incompatibility with historical-entitlement conceptions of justice. For
Rawls’ theory itself describes a process (abstractly conceived) with
a result. He does not present a direct deductive argument for his two
principles of justice from other statements that entail them. Any
deductive formulation of Rawls’ argument would contain meta-statements, statements about principles, such as: Any principles agreed
41. This is especially serious in view of the weakness of Rawls’ reasons, as
presented in ?82, for placing the liberty principle prior to the difference principle in a lexicographic ordering.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
I04 Philosophy & Public Affairs
to by persons in a certain situation are correct. Combined with an
argument showing that persons in that situation would agree to principles P, one can deduce that P is correct, and then deduce that P.
At some places in the argument, “P” appears in quotes, distinguishing
the argument from a direct deductive argument for the truth of P.
Instead of a direct deductive argument, a situation and process are
specified, and any principles that would emerge from that situation
and process are held to constitute the principles of justice. (Here I
ignore the complicated interplay between which principles of justice
one wants to derive, and which initial situation one specifies.) Just
as for an entitlement theorist any set of holdings that emerges from
a legitimate process (specified by the principle of transfer) is just, so
for Rawls any set of principles that emerges from the original position
by the constrained process of unanimous agreement is the set of
(correct) principles of justice. Each theory specifies starting points
and processes of transformation, and each accepts whatever comes out.
According to each theory, whatever comes out is to be accepted because of its pedigree, its history. Any theory which gets to a process
must start with something which is not itself justified by being the
outcome of a process (otherwise, it should start farther back); namely,
either general statements arguing for the fundamental priority of the
process, or with the process itself. Entitlement theory and Rawls’
theory each get to a process. Entitlement theory specifies a process
for generating sets of holdings. The three principles of justice (in
acquisition, transfer, and rectification) that underlie this process,
having this process as their subject matter, are themselves processprinciples rather than end-state principles of distributive justice. They
specify an ongoing process, without fixing how it is to turn out, without
providing some external patterned criterion it must meet. Rawls’
theory arrives at a process P for generating principles of justice. This
process P involves people in the original position agreeing to principles
of justice behind a veil of ignorance. According to Rawls, any principles emerging from this process P will be the principles of justice.
But this process P for generating principles of justice cannot, we already have argued, itself generate process-principles as the fundamental principles of justice. P must generate end-state or end-result prinThis content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
I05 Distributive Justice
ciples. Even though the difference principle, in Rawls’ theory, is to
apply to an ongoing and continuing institutional process (one that
includes derived entitlements based upon institutional expectations
under the principle, and derived elements of pure procedural justice,
etc.), it is an end-result principle (but not a current time-slice principle). The difference principle fixes how the ongoing process is to
turn out and provides an external patterned criterion it must meet; any
process is rejected which fails to meet the test of the criterion. The
mere fact that a principle regulates an ongoing institutional process
does not make it a process-principle. If it did, the utilitarian principle
would also be a process-principle, rather than the end-result principle
it is.
The structure of Rawls’ theory thus presents a dilemma. If processes are so great, Rawls’ theory is defective because incapable of
yielding process-principles of justice. If processes are not so great,
then insufficient support has been provided for the principles yielded
by Rawls’ process P for arriving at principles. Contract arguments
embody the assumption that anything that emerges from a certain
process is just. Upon the force of this fundamental assumption rests
the force of a contract argument. Surely then no contract argument
should be structured so as to preclude process principles being the
fundamental principles of distributive justice by which to judge the
institutions of a society; no contract argument should be structured
so as to make it impossible that its results be of the same sort as the
assumptions upon which it rests.42 If processes are good enough to
found a theory upon, they are good enough to be the possible result
of the theory. One can’t have it both ways.
We should note that the difference principle is an especially strong
kind of patterned end-state principle. Let us say that a principle of
distribution is organic if an unjust distribution, according to the principle, can be gotten from one the principle deems just, by deleting (in
imagination) some people and their distributive shares. Organic principles focus on features dependent upon the overall pattern. In con42. “The idea of the original position is to set up a fair procedure so that any
principle agreed to will be just. The aim is to use the notion of pure procedural
justice as a basis for theory” (?24, p. 136).
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
io6 Philosophy & Public Affairs
trast, pattemed principles of the form “to each according to his score
on a particular natural dimension D” are not organic principles. If
a distribution satisfies this principle, it will continue to do so when
some people and their holdings are deleted, for this deletion will not
affect the ratios of the remaining people’s holdings, or the ratios of
their scores along the dimension D. These unchanged ratios will continue to be the same, and will continue to satisfy the principle.
The difference principle is organic. If the least well-off group and
their holdings are deleted from a situation, there is no guarantee that
the resulting situation and distribution will maximize the position of
the new least well-off group. Perhaps that new bottom group could
have more if the top group had even less (though there was no way
to transfer from the top group to the previous bottom group).43
Failure to satisfy the deletion condition (that a distribution remains
just under deletion of people and their holdings) marks off organic
principles. Consider also the addition condition, which holds that if
two distributions (over disjoint sets of individuals) are just, then so
is the distribution which consists in the combination of these two
just distributions. (If the distribution on earth is just, and that on
some planet of a distant star is just, then so is the sum distribution of
the two.) Principles of distribution of the form “to each according
to his score on natural dimension D” violate this condition, and therefore (let us say) are nonaggregative. For though within each group,
all ratios of shares match ratios of scores on D, they needn’t match
between the groups.44 The entitlement principle of justice in holdings
satisfies both the deletion and the addition conditions; the entitlement
principle is inorganic and aggregative.
43. The difference principle thus creates two conflicts of interest: between
those at the top and those at bottom; and between those in the middle and those
at bottom, for if those at bottom were gone, the difference principle might apply
to improve the position of those in the middle, who would become the new
(bottom) group whose position is to be maximized.
44. Let the second group have individuals who score half as much on D and
have shares twice as large as the corresponding individuals in the first group,
where in the first group the ratios between any two individuals’ shares and their
scores on D are the same. It follows that within the second group, the ratio of
any two individuals’ shares will be the same as the ratio of their scores. Yet between groups this identity of ratios will not hold.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
107 Distributive Justice
5. Natural Assets and Arbitrariness. Rawls comes closest to considering the entitlement system in his discussion of what he terms
the system of natural liberty:
The system of natural liberty selects an efficient distribution
roughly as follows. Let us suppose that we know from economic
theory that under the standard assumptions defining a competitive
market economy, income and wealth will be distributed in an
efficient way, and that the particular efficient distribution which
results in any period of time is determined by the initial distribution
of assets, that is, by the initial distribution of income and wealth,
and of natural talents and abilities. With each initial distribution,
a definite efficient outcome is arrived at. Thus it turns out that if
we are to accept the outcome as just, and not merely as efficient,
we must accept the basis upon which over time the initial distribution of assets is determined.
In the system of natural liberty the initial distribution is regulated by the arrangements implicit in the conception of careers open
to talents. These arrangements presuppose a background of equal
liberty (as specified by the first principle) and a free market economy. They require a formal equality of opportunity in that all have
at least the same legal rights of access to all advantaged social
positions. But since there is no effort to preserve an equality or
similarity, of social conditions, except insofar as this is necessary
to preserve the requisite background institutions, the initial distribution of assets for any period of time is strongly influenced by
natural and social contingencies. The existing distribution of income and wealth, say, is the cumulative effect of prior distributions
of natural assets-that is, natural talents and abilities-as these have
been developed or left unrealized, and their use favored or disfavored over time by social circumstances and such chance contingencies as accident and good fortune. Intuitively, the most
obvious injustice of the system of natural liberty is that it permits
distributive shares to be improperly influenced by these factors so
arbitrary from a moral point of view” (? 12, p. 72) .45
45. Rawls goes on to discuss what he calls a liberal interpretation of his two
principles of justice, which is designed to eliminate the influence of social
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
io8 Philosophy & Public Affairs
Here we have Rawls’ reason for rejecting a system of natural liberty:
it “permits distributive shares to be improperly influenced by these
factors so arbitrary from a moral point of view.” These factors are:
“prior distribution . . . of natural talents and abilities as these have
been developed over time by social circumstances and such chance
contingencies as accident and good fortune.” Notice that there is no
mention at all of how persons have chosen to develop their own natural assets. Why is that simply left out? Perhaps because such choices
also are viewed as being the products of factors outside the person’s
control, and hence as “arbitrary from a moral point of view.” “The
assertion that a man deserves the superior character that enables him
to make the effort to cultivate his abilities is equally problematic; for
his character depends in large part upon fortunate family and social
circumstances for which he can claim no credit” (p. 104). (What view
is presupposed here of character and its relation to action?) “The
initial endowment of natural assets and the contingencies of their
growth and nurture in early life are arbitrary from a moral point of
view . . . the effort a person is willing to make is influenced by his
natural abilities and skills and the alternatives open to him. The better endowed are more likely, other things equal, to strive conscientiously . . .” (pp. 3II-3I2). This line of argument can succeed in
blocking the introduction of persons’ autonomous choices and actions
(and their results) only by attributing everything noteworthy about
the person completely to (certain sorts of) “external” factors. So denigrating a person’s autonomy and prime responsibility for his actions
is a risky line to take for a theory that otherwise wishes to buttress the
dignity and self-respect of autonomous beings; especially for a theory
that founds so much (including a theory of the good) upon persons’
choices. One doubts that the unexalted picture of human beings
contingencies, but which “intuitively, still appears defective . . . [for] it still
permits the distribution of wealth and income to be determined by the natural
distribution of abilities and talents . . . distributive shares are decided by the
outcome of the natural lottery; and this outcome is arbitrary from a moral perspective. There is no more reason to permit the distribution of income and wealth
to be settled by the distribution of natural assets than by historical and social
fortune” (pp. 73-74).
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
I09 Distributive Justice
Rawls’ theory presupposes and rests upon can be made to fit together
with the view of human dignity it is designed to lead to and embody.
Before we investigate Rawls’ reasons for rejecting the system of
natural liberty, we should note the situation of those in the original
position. The system of natural liberty is one interpretation of a principle that (according to Rawls) they do accept: social and economic
inequalities are to be arranged so that they both are reasonably expected to be to everyone’s advantage, and attached to positions and
offices open to all. It is left unclear whether the persons in the original
position explicitly consider and choose among all the various interpretations of this principle, though this would seem to be the most
reasonable construal. (Rawls’ chart on page I24 listing the conceptions of justice considered in the original position does not include
the system of natural liberty.) Certainly they explicitly consider one
interpretation, the difference principle. Rawls does not state why
persons in the original position who considered the system of natural
liberty would reject it. Their reason cannot be that it makes the resulting distribution depend upon a morally arbitrary distribution of
natural assets. What we must suppose, as we have seen before, is
that the self-interested calculation of persons in the original position
does not (and cannot) lead them to adopt the entitlement principle.
We, however, and Rawls, base our evaluations on different considerations.
Rawls has explicitly designed the original position and its choice
situation so as to embody and realize his (negative) reflective evaluation of allowing shares in holdings to be affected by natural assets:
“Once we decide to look for a conception of justice that nullifies the
accidents of natural endowment and the contingencies of social circumstance . . .” (p. I5). (Rawls makes many scattered references to
this theme of nullifying the accidents of natural endowment and the
contingencies of social circumstance.) This quest crucially shapes
Rawls’ theory, and it underlies his delineation of the original position.
It is not that persons who did deserve their natural endowments
would choose differently if placed in Rawls’ original position; but
rather that, presumably, for such persons, Rawls would not hold that
the principles of justice to govern their mutual relations were fixed
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
IIO Philosophy & Public Affairs
by what they would choose in the original position. It is useful to remember how much of Rawls’ construction rests upon this foundation. For example, Rawls argues (,?8i) that certain egalitarian demands are not motivated by envy but rather, because they are in
accord with his two principles of justice, are motivated by resentment
of injustice. This argument can be undercut, as Rawls realizes,46 if
the very considerations which underlie the original position (yielding
Rawls’ two principles of justice) themselves embody or are based upon envy. So in addition to wanting to understand Rawls’ rejection of
alternative conceptions and to assess how powerful a criticism he
makes of the entitlement conception, reasons internal to his theory
provide motivation to explore the basis of the requirement that a conception of justice be geared to nullify differences in social circumstances, and in natural assets (and any differences in social circumstances they result in).
Why shouldn’t holdings partially depend upon natural endowments? (They will also depend on how these are developed, and the
uses to which they are put.) Rawls’ reply is that these natural endowments and assets, being undeserved, are “arbitrary from a moral point
of view.” There are two ways to understand the relevance of this reply:
it might be part of an argument to establish that the distributive effects
of natural differences ought to be nullified, which I shall call the positive reply; or it might be part of an argument to rebut a possible
counterargument holding that the distributive effects of natural differences oughtn’t to be nullified, which I shall call the negative reply.
Whereas the positive argument attempts to establish that the distributive effects of natural differences ought to be nullified, the negative reply, by merely rebutting one argument that the differences
oughtn’t to be nullified, leaves open the possibility that (for other
reasons) the differences oughtn’t to be nullified. (The negative reply
also leaves it possibly a matter of moral indifference whether the distributive effects of natural differences are to be nullified; note the
difference between saying that something ought to be the case and
saying that it’s not that it oughtn’t to be the case.)
46. “In order to show that the principles of justice are based in part on envy
it would have to be established that one or more of the conditions of the original
position arose from this propensity” (p. 538).
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
III Distributive Justice
6. The Positive Connection. We shall begin with the positive reply.
How might the point that differences in natural endowments are
arbitrary from a moral point of view, function in an argument meant
to establish that differences in holdings stemming from differences in
natural assets ought to be nullified. We shall consider four possible
arguments; the first, the following argument A:
(i) Any person should morally deserve the holdings he has; it
shouldn’t be that persons have holdings they don’t deserve.
(2) People do not morally deserve their natural assets.
(3) If a person’s X partially determines his Y, and his X is undeserved then so is his Y.
Therefore,
(4) People’s holdings shouldn’t be partially determined by their
natural assets.
This argument will serve as a surrogate for other similar, more complicated ones.47 But Rawls explicitly and emphatically rejects distribution according to moral desert:
There is a tendency for common sense to suppose that income and
wealth, and the good things in life generally, should be distributed
according to moral desert. Justice is happiness according to virtue.
While it is recognized that this ideal can never be fully carried out,
it is the appropriate conception [according to common sense] of
distributive justice, at least as a prima facie principle, and society
should try to realize it as circumstances permit. Now justice as fairness rejects this conception. Such a principle would not be chosen
in the original position (?48). (The rest of ?48 goes on to criticize
the conception of distribution according to moral desert.)
47. For example,
(I) Differences between any two persons’ holdings should be morally deserved; morally undeserved differences should not exist.
(2) Differences between persons in natural assets are morally undeserved.
(3) Differences between persons partially determined by other differences
that are undeserved, are themselves undeserved.
Therefore,
(4) Differences between persons’ holdings shouldn’t be partially determined
by differences in their natural assets.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
II2 Philosophy & Public Affairs
Rawls could not, therefore, accept any premiss like the first premiss
in argument A, and so no variant of this argument underlies his rejection of differences in distributive shares stemming from (undeserved) differences in natural assets. Not only does Rawls reject
premiss (i), his theory is not coextensive with it. He favors giving
incentives to persons if this most improves the lot of the least well off,
and it often will be because of their natural assets that these persons
will receive incentives and have larger shares. We noted earlier that
the entitlement conception of justice in holdings, not being a patterned conception of justice, does not accept distribution in accordance with moral desert either. Anyone may give any holding they are
entitled to, to anyone else, independently of whether the recipient
morally deserves to be the recipient. To each according to the legitimate entitlements that legitimately have been transferred to him, is
not a patterned principle.
If argument A and its first premiss are rejected, it is not obvious
how to construct the positive argument. Consider next argument B:
(i) Holdings ought to be distributed according to some pattern
that is not arbitrary from a moral point of view.
(2) That persons have different natural assets is arbitrary from
a moral point of view.
Therefore,
(3) Holdings ought not to be distributed according to natural
assets.
But differences in natural assets might be correlated with other differences that are not arbitrary from a moral point of view and that
are clearly of some possible moral relevance to distributional questions. For example, Hayek argued that, under capitalism, distribution
generally is in accordance with perceived service to others. Since differences in natural assets will produce differences in ability to serve
others, there will be some correlation of differences in distribution
with differences in natural assets. The principle of the system is not
distribution in accordance with natural assets; but differences in
natural assets will lead to differences in holdings under a system
whose principle is distribution according to perceived service to others.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
II3 Distributive Justice
If the conclusion (3) above is to be interpreted in extension so as to
exclude this, it should be made explicit. But to add the premiss that
any pattern that has some roughly coextensive description that is
arbitrary from a moral point of view is itself arbitrary from a moral
point of view, would be far too strong, as it would yield the result
that every pattern is arbitrary from a moral point of view. Perhaps
the crucial thing to be avoided is not mere coextensiveness, but rather
some morally arbitrary feature’s giving rise to differences in distributive shares. Thus consider argument c:
(i) Holdings ought to be distributed according to some pattern
that is not arbitrary from a moral point of view.
(2) That persons have different natural assets is arbitrary from
a moral point of view.
(3) If part of the explanation of why a pattern contains differences
in holdings is that other differences in persons give rise to
these differences in holdings, and if these other differences
are arbitrary from a moral point of view, then the pattern also
is arbitrary from a moral point of view.
Therefore,
(4) Differences in natural assets should not give rise to differences
in holdings among persons.
Premiss (3) of this argument holds that any moral arbitrariness that
underlies a pattern infects the pattern and makes it too morally
arbitrary. But any pattern will have some morally arbitrary facts as
part of the explanation of how it arises, including the pattern proposed by Rawls. The difference principle operates to give some persons
larger distributive shares than others; which persons receive these
larger shares will depend (partially) on differences between these
persons and others that are arbitrary from a moral point of view, for
some persons with special natural assets will be offered larger shares
as an incentive to use these assets in certain ways. Perhaps some
premiss similar to (3) can be formulated so as to exclude what Rawls
wishes to exclude while not excluding his own view. Still, the resulting
argument would assume that the set of holdings should realize some
pattern.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
II4 Philosophy & Public Affairs
Why should the set of holdings be patterned? Patterning is not
intrinsic to a theory of justice, as we have seen in our presentation of
the entitlement theory: a theory that focuses upon the underlying
principles that generate sets of holdings rather than upon the pattern
a set of holdings realizes. If it be denied that the theory of these underlying principles is a separate theory of distributive justice, rather than
merely a collection of diverse considerations from other areas, then
the question becomes one of whether there is any separate subject of
distributive justice which requires a separate theory.
On the manna from heaven model, there might be a more compelling reason to search for a pattern. But since things come into being
already held (or with agreements already made about how they are
to be held), there is no need to search for some pattern for unheld
holdings to fit; and since the process whereby holdings actually come
into being or are shaped, itself needn’t realize any particular pattern,
there is no reason to expect any pattern to result. The situation is not
an appropriate one for wondering, “after all, what is to become of
these things; what are we to do with them.” In the non-manna-fromheaven world in which things have to be made or produced or transformed by people, there is no separate process of distribution for a
theory of distribution to be a theory of. The reader will recall our
earlier argument that (roughly) any set of holdings realizing a particular pattern may be transformed by the voluntary exchanges, gifts,
etc., of the persons having the holdings under the pattern, into another set of holdings that does not fit the pattern. The view that holdings must be patterned perhaps will seem less plausible when it is
seen to have the consequence that people may not choose to do (even
with things they legitimately hold) acts that upset the patterning.
There is another route to a patterned conception of justice that,
perhaps, should be mentioned. Suppose that each morally legitimate
fact has a “unified” explanation that shows it is morally legitimate;
and that conjunctions fall into the domain of facts to be explained as
morally legitimate. If p, and q are each morally legitimate facts, with
their respective explanations as morally legitimate being P, and Q,
then if p Aq is also to be explained as morally legitimate, and if P AQ
does not constitute a “unified” explanation (but is a mere conjunction
of different explanations), then some further explanation will be
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
I I,5 Distributive Justice
needed. Applying this to holdings, suppose there are separate entitlement explanations showing the legitimacy of my having my holdings,
and of your having yours, and the question is asked: “Why is it legitimate that I hold what I do and you hold what you do; why is that joint
fact and all the relations contained within it legitimate?” If the conjunction of the two separate explanations will not be held to explain
in a unified manner the legitimacy of the joint fact (whose legitimacy
is not viewed as being constituted by the legitimacy of its constituent
parts), then some patterned principles of distribution would appear
to be necessary to show its legitimacy, and to legitimate any non-unit
set of holdings.
With scientific explanation of particular facts, the usual practice
is to consider some conjunctions of explained facts as not requiring
separate explanation, but as being explained by the conjunctions
of the explanations of the conjuncts. (If E1 explains el and E2 explains e2 then E1AE2 explains el Ae2.) If we required that any two
conjuncts and any n-place conjunction had to be explained in some
unified fashion, and not merely by the conjunction of separate and
disparate explanations, then we would be driven to reject most of the
usual explanations and to search for an underlying pattern to explain
what appear to be separate facts. (Scientists, of course, often do offer
a unified explanation of apparently separate facts.) It would be well
worth exploring the interesting consequences of refusing to treat, even
in the first instance, any two facts as legitimately separable, as having
separate explanations whose conjunction is all there is to the explanation of them. What would our theories of the world look like if we required unified explanations of all conjunctions? Perhaps an extrapolation of how the world looks to paranoid persons. Or, to put it
undisparagingly, the way it appears to persons having certain sorts of
dope experiences. (For example, the way it sometimes appears to me
with marijuana.) Such a vision of the world differs fundamentally
from the way we normally look at it; it is surprising at first that a simple condition on the adequacy of explanations of conjunctions leads to
it, until we realize that such a condition of adequacy must lead to a
view of the world as deeply and wholly patterned.
A similar condition of adequacy on explanations of the moral legitimacy of conjunctions of separate morally legitimate facts would lead
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
II6 Philosophy & Public Affairs
to a view that requires sets of holdings to exhibit an overall patterning.
It seems unlikely that there will be compelling arguments for imposing such a principle of adequacy. Some may find such a unified vision
plausible for only one realm; e.g., in the moral realm concerning sets
of holdings, but not in the realm of ordinary nonmoral explanation,
or vice versa. For the case of explaining nonmoral facts, the challenge
would be to produce such a unified theory. Were one produced that
introduced novel considerations and explained no new facts (other
than conjunctions of old ones) the decision as to its acceptability
might be a difficult one, and would depend largely on how explanatorily satisfying was the new way we saw the old facts. In the case of
moral explanations and accounts which show the moral legitimacy of
various facts, the situation is somewhat different. First, there is even
less reason (I believe) to suppose a unified explanation appropriate
and necessary. There is less need for a greater degree of explanatory
unity than that provided when the same underlying principles for
generating holdings appear in different explanations. (Rawls’ theory,
which contains elements of what he calls pure procedural justice, does
not satisfy a strong condition of adequacy for explaining conjunctions, and entails that such a condition cannot be satisfied.) Secondly,
there is more danger than in the scientific case that the demand for
a unified explanation will shape the “moral facts” to be explained.
(“It can’t be that both of those are facts for there’s no unified patterned
explanation that would yield them both.”) Hence success in finding a
unified explanation of such seriously primed facts will leave it unclear
how well supported the explanatory theory is.
I turn now to our final positive argument (which purports) to derive
the conclusion that distributive shares shouldn’t depend upon natural
assets, from the statement that the distribution of natural assets is
morally arbitrary. This argument focuses on the notion of equality.
Since a large part of Rawls’ argument serves to justify or show acceptable a particular deviation from equal shares (some may have
more if this serves to improve the position of those worst off ), perhaps
a reconstruction of his underlying argument that places equality at
its center will be illuminating. Differences between persons (the argument runs) are arbitrary from a moral point of view if there is no
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
II7 Distributive Justice
moral argument for the conclusion that there ought to be the differences. Not all such differences will be morally objectionable. That
there is no such moral argument will seem important only in the case
of those differences we believe oughtn’t to obtain unless there is a
moral reason establishing that they ought to obtain. There is, so to
speak, a presumption against certain differences that can be overridden (can it merely be neutralized?) by moral reasons; in the absence of any such moral reasons of sufficient weight, there ought to be
equality. Thus we have argument D:
(i) Holdings ought to be equal, unless there is a (weighty) moral
reason why they ought to be unequal.
(2) People do not deserve the ways in which they differ from other
persons in natural assets; there is no moral reason why people
ought to differ in natural assets.
(3) If there is no moral reason why people differ in certain traits,
then their actually differing in these traits does not provide
and cannot give rise to a moral reason why they should differ
in other traits (e.g., in holdings).
Therefore,
(4) People’s differing in natural assets is not a reason why holdings
ought to be unequal.
Therefore,
(5) People’s holdings ought to be equal unless there is some other
moral reason (such as, e.g., raising the position of those worst
off) why their holdings ought to be unequal.
Statements similar to the third premiss will occupy us shortly. Here
let us focus on the first premiss, the equality premiss. Why ought
people’s holdings to be equal, in the absence of special moral reason
to deviate from equality? (Why think there ought to be any particular
pattern in holdings?) Why is equality the rest (or rectilinear motion)
position of the system, deviation from which may be caused only by
moral forces? Many “arguments” for equality merely assert that differences between persons are arbitrary and must be justified. Often
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
ii8 Philosophy & Public Affairs
writers state a presumption in favor of equality, in some form such
as: Differences in treatment of persons need to be justified.48 The
most favored situation for this sort of assumption is one in which
there is one person (or group) treating everyone, and having no
right or entitlement to bestow the particular treatment as they wish
or even whim. But if I go to one movie theater rather than another
adjacent to it, need I justify my different treatment of the two theater
owners? Isn’t it enough that I felt like going to one of them? That
differences in treatment need to be justified does fit contemporary
governments. Here there is a centralized process treating all, with no
entitlement to bestow treatment according to whim. The major portion
of distribution in a free society does not, however, come through the
actions of the government, nor does failure to overturn the results of
the localized individual exchanges constitute “state action.” When
there is no one doing the treating, and all are entitled to bestow their
holdings as they wish, it is not clear why the maxim that differences
in treatment must be justified, should be thought to have extensive
application. Why must differences between persons be justified? Why
think that we must change, or remedy, or compensate for any inequality which can be changed, remedied, or compensated for? Perhaps here is where social cooperation enters in: though there is no
presumption of equality (in say, primary goods, or things people
care about) among all persons, perhaps there is one among persons
cooperating together. But it is difficult to see an argument for this;
surely not all persons who cooperate together explicitly agree to this
presumption as one of the terms of their mutual cooperation. And its
48. “No reason need be given for . . . an equal distribution of benefits-for
that is ‘natural’-self-evidently right and just, and needs no justification, since
it is in some sense conceived as being self-justified…. The assumption is that
equality needs no reasons, only inequality does so; that uniformity, regularity,
similarity, symmetry, . . . need not be specially accounted for, whereas differences, unsystematic behavior, changes in conduct, need explanation and, as a
rule, justification. If I have a cake and there are ten persons among whom I
wish to divide it, then if I give exactly one tenth to each, this will not, at any
rate automatically, call for justification; whereas if I depart from this principle
of equal division I am expected to produce a special reason. It is some sense of
this, however latent, that makes equality an idea which has never seemed intrinsically eccentric….” Isaiah Berlin, “Equality,” reprinted in Frederick A. Olafson,
Justice and Social Policy (New York I96I1), p. I3I.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
II9 Distributive Justice
acceptance would provide an unfortunate incentive for well-off persons to refuse to cooperate with or allow any of their number to cooperate with some distant people, less well off than any among them.
For entering into such social cooperation, beneficial to those less well
off, would seriously worsen the position of the well-off group by creating relations of presumptive equality between themselves and the
worse-off group. Chapter 8 of the forthcoming Anarchy, State, and
Utopia includes a consideration of the major recent argument for
equality, one which turns out to be unsuccessful. Here we need only
note that the connection argument D forges between not deserving natural assets and some conclusion about distributive shares assumes
equality as a norm (that can be deviated from with, and only with,
moral reason); and hence D itself cannot be used to establish any such
conclusion about equality.
7. The Negative Connection. Unsuccessful in our quest for a convincing positive argument to connect the claim that people don’t
deserve their natural assets with the conclusion that differences in
holdings ought not to be based upon differences in natural assets, we
now turn to what we called the negative argument: the use of the
claim that people don’t deserve their natural assets to rebut a possible
counterargument to Rawls’ view. (If the equality argument D were
acceptable, the negative task of rebutting possible counterconsiderations would form part of the positive task of showing that a presumption for equality holds unoverridden in a particular case.) Consider
the following possible counterargument E to Rawls:
(i) People deserve their natural assets.
(2) If people deserve X, they deserve any Y that flows from X.
(3) People’s holdings flow from their natural assets.
Therefore,
(4) People deserve their holdings.
(5) If people deserve something, then they ought to have it (and
this overrides any presumption of equality there may be about
that thing).
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
120 Philosophy & Public Affairs
Rawls would rebut this counterargument to his position by denying its
first premiss. And so we see some connection between the claim that
the distribution of natural assets is arbitrary, and the statement that
distributive shares should not depend upon natural assets. However,
no great weight can be placed upon this connection. For there are other
counterarguments, in a similar vein; for example the argument F that
begins:
(i) If people have X, and their having X (whether or not they deserve to have it) does not violate anyone else’s (Lockean)
right or entitlement to X, and Y flows from (arises out of, etc.)
X by a process that does not itself violate anyone’s (Lockean)
rights or entitlements,49 then the person is entitled to Y.
(2) People’s having the natural assets they do, does not violate
anyone else’s (Lockean) entitlements or rights,
and goes on to argue that people are entitled to what they make, to
the products of their labor, to what others give them or exchange with
them. It is not true, for example, that a person earns Y (a right to keep
a painting he’s made, praise for writing A Theory of Justice, etc.) only
if he’s earned (or otherwise deserves) whatever he used (including natural assets) in the process of earning Y. Some of the things he uses he
just may have, not illegitimately. It needn’t be that the foundations
underlying desert are themselves deserved, all the way down.
At the very least, we can parallel these statements about desert
with ones about entitlements. And if, correctly, we describe people as
entitled to their natural assets even if it’s not the case that they can be
said to deserve them, then the argument parallel to E above, with “are
entitled to” replacing “deserve” throughout, will go through. This gives
us the acceptable argument G:
49. A process, we might strengthen the antecedent by adding, of the sort that
would create an entitlement to Y if the person were entitled to X. I use “Lockean”
rights and entitlements to refer to those against force, fraud, etc., which are to
be recognized in the minimal state. Since I believe these are the only rights and
entitlements people possess (apart from those they specially acquire), I needn’t
have included the specification to Lockean rights. One who believes some have
a right to the fruits of others’ labor, will deny the truth of the first premiss as
stated. If the Lockean specffication were not included, he might grant the truth
of (I), while denying that of (2) or of later steps.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
I 21 Distributive Justice
( i ) People are entitled to their natural assets.
(2) If people are entitled to something, they are entitled to whatever flows from it (via specified types of processes).
(3) People’s holdings flow from their natural assets.
Therefore,
(4) People are entitled to their holdings.
(5) If people are entitled to something then they ought to have it
(and this overrides any presumption of equality there may be
about holdings).
Whether or not people’s natural assets are arbitrary from a moral point
of view, they are entitled to them, and to what flows from them.50
A recognition of people’s entitlements to their natural assets (the
first premiss of argument G) might be necessary to avoid the stringent
application of the difference principle which would lead, we already
have seen, to even stronger property rights in other persons than redistributive theories usually yield. Rawls thinks to avoid this by people
in his original position ranking the principle of liberty as lexicographically prior to the difference principle, applied not only to economic
well-being but to health, length of life, etc.5′
We have found no cogent argument to (help) establish that differences in holdings arising from differences in natural assets should
be eliminated or minimized. Can the theme that people’s natural
assets are arbitrary from a moral point of view, be used differently, for
example, to justify a certain shaping of the original position? Clearly
if the shaping is designed to nullify differences in holdings due to dif50. If nothing of moral significance could flow from what was arbitrary, then
no particular person’s existence could be of moral significance; since of the many
sperm cells, which one succeeds in fertilizing the egg cell is (so far as we know)
arbitrary from a moral point of view. This suggests another, vaguer remark
directed to the spirit of Rawls’ position rather than its letter. Each existing
person is the product of a process wherein the one sperm cell which succeeds
is no more deserving than the millions that fail. Should we wish that process
had been “fairer” as judged by Rawls’ standards, that all “inequities” in it had
been nullified? We should be apprehensive about any principle that would condemn morally the very sort of process that brought us to be, and that therefore
would undercut the legitimacy of our very existing.
5I. But see our discussion below of Rawls’ view of natural abilities as a collective asset; and see also footnote 4I above.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
122 Philosophy & Public Affairs
ferences in natural assets, we need an argument for this goal, and we
are back to our unsuccessful quest for the route to the conclusion that
such differences in holdings ought to be nullified. Instead, the shaping
might take place, by excluding the participants in the original position
from knowing of their own natural endowments. In this way the fact
that natural endowments are arbitrary from a moral point of view
would help to impose and to justify the veil of ignorance. But how
does it do this; why should knowledge of natural endowments be excluded from the original position? Presumably the underlying principle would be that if any particular features are arbitrary from a
moral point of view then persons in the original position should not
know they possess them. But this would exclude their knowing anything about themselves, for each of their features (including rationality, the ability to make choices, having a life span of more than three
days, having a memory, being able to communicate with other organisms like themselves) will be based upon the fact that the sperm and
ovum which produced them contained particular genetic material. The
physical fact that those particular gametes contained particular organized chemicals (the genes for people rather than for muskrats or
trees) is arbitrary from a moral point of view; it is, from a moral point
of view, an accident. Yet the persons in the original position are to
know some of their attributes.
Perhaps we are too quick when we suggest excluding knowledge or
rationality, etc., merely because these features arise from morally
arbitrary facts. For these features also have moral significance; that
is, moral facts depend upon or arise from them. Here we see an ambiguity in saying a fact is arbitrary from a moral point of view. It
might mean that there is no moral reason why the fact ought to be
that way, or it might mean that the fact’s being that way is of no moral
significance and has no moral consequences. Rationality, the ability
to make choices, etc., are not morally arbitrary in this second sense.
But if they escape exclusion on this ground, now the problem is that
the natural assets, knowledge of which Rawls wishes to exclude from
the original position, are not morally arbitrary in this sense either. At
any rate, the entitlement theory’s claim that moral entitlements may
arise from or be partially based upon such facts, is what is now at
issue. Thus, in the absence of an argument to the effect that differThis content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
123 Distributive Justice
ences in holdings due to differences in natural assets ought to be
nullified, it is not clear how anything about the original position can
be based upon the (ambiguous) claim that differences in natural assets are arbitrary from a moral point of view.
8. Collective Assets. Rawls’ view seems to be that everyone has
some entitlement or claim on the total of natural assets (viewed as a
pool), with no one having differential claims. The distribution of natural abilities is viewed as a “collective asset” (p. I79).
We see then that the difference principle represents, in effect, an
agreement to regard the distribution of natural talents as a common
asset and to share in the benefits of this distribution whatever it
turns out to be. Those who have been favored by nature, whoever
they are, may gain from their good fortune only on terms that improve the situation of those who have lost out. . . . No one deserves
his greater natural capacity nor merits a more favorable starting
place in society. But it does not follow that one should eliminate
these distinctions. There is another way to deal with them. The
basic structure can be arranged so that these contingencies work
for the good of the least fortunate (pp. ioi-io2).
People will differ in how they view regarding natural talents as a
common asset. Some will complain, echoing Rawls against utilitarianism (p. 27), that this “does not take seriously the distinction between
persons”; and they will wonder whether any reconstruction of Kant
can be adequate that treats people’s abilities and talents as resources
for others. “The two principles of justice . . . rule out even the tendency to regard men as means to one another’s welfare” (p. I83). Only
if one presses very hard on the distinction between men and their
talents, assets, abilities and special traits. Whether any coherent conception of a person remains when the distinction is so pressed is an
open question. Why we, thick with particular traits, should be cheered
that (only) the thus purified men within us are not regarded as means,
is also unclear.
People’s talents and abilities are an asset to a free community;
others in the community benefit from their presence and are better
off because they are there rather than elsewhere or nowhere. (OtherThis content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
124 Philosophy & Public Affairs
wise they wouldn’t choose to deal with them.) Life, over time, is not
a constant-sum game, wherein if greater ability or effort leads to some
getting more, that means that others must lose. In a free society,
people’s talents do benefit others, and not only themselves. Is it the
extraction of even more benefit to others that is supposed to justify
treating people’s natural assets as a collective resource? What justifies
this extraction?
No one deserves his greater natural capacity nor merits a more
favorable starting place in society. But it does not follow that one
should eliminate these distinctions. There is another way to deal
with them. The basic structure can be arranged so that these contingencies work for the good of the least fortunate (p. I02).
And if there weren’t another “way to deal with them”? Would it then
follow that one should eliminate these distinctions? What exactly
would be contemplated in the case of natural assets? If people’s assets
and talents couldn’t be harnessed to serve others, would something be
done to remove these exceptional assets and talents, or to forbid them
from being exercised for the person’s own benefit or that of someone
else he chose; even though this limitation wouldn’t improve the absolute position of those somehow unable to harness the talents and abilities of others for their own benefit? Is it so implausible to claim that
envy underlies this conception of justice, forming part of its root
notion?52
52. Wil the lexicographic priority that Rawls claims for liberty in the original
position, prevent the difference principle from requiring a head tax on assets
and abilities? The legitimacy of a head tax is suggested by Rawls’ speaking of
“collective assets” and “common assets.” Those underutilizing their assets and
abilities are misusing a public asset. (Squandering public property?) Rawls may
intend no such strong inferences from his terminology, but we need to hear more
about why those in the original position wouldn’t accept the strong interpretation
and “agree to share one another’s fate” (p. I02). The notion of liberty needs elaboration which is to exclude a head tax yet allow the other taxation schemes. Assets and abilities can be harnessed without a head tax; and “harnessing” is an appropriate term, as it would be for a horse in hamess to a wagon which horse
doesn’t have to move ever, but if it does, it must draw the wagon along.
With regard to envy, the difference principle, applied to the choice between
A having ten and B having five; and A having eight and B having five, would
favor the latter. Thus, despite Rawls’ remarks on pages 79-80, it is inefficient
in that it sometimes will favor a status quo against a Pareto-better but more
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
125 Distributive Justice
We have used our entitlement conception of justice in holdings to
probe Rawls’ theory, sharpening our understanding of what the entitlement conception involves by bringing it to bear upon an alternative
conception of distributive justice, one that is deep and elegant. Also,
I believe, we have probed deep lying inadequacies in Rawls’ theory.
I am mindful of Rawls’ reiterated point that a theory cannot be evaluated by focusing upon a single feature or part of it; instead the whole
theory must be assessed (the reader will not know how whole a theory
can be until he has read all of Rawls’ book), and a perfect theory is
not to be expected. However, we have examined an important part of
Rawls’ theory, and its crucial underlying assumptions. I am as well
aware as anyone of how sketchy my discussion of the entitlement
conception of justice in holdings has been. But I no more believe we
need to have formulated a complete alternative theory in order to
reject Rawls’ undeniably great advance over utilitarianism, than
Rawls needed a complete alternative theory before he could reject
utilitarianism. What more does one need or can one have, in order
to begin progressing towards a better theory, than a sketch of a plausible alternative view, which from its very different perspective highlights the inadequacies of the best existing well worked-out theory?
Here, as in so many things, we learn from Rawls.
We close by considering the claim that a state more extensive than
the minimal state can be justified on the grounds that it is necessary,
or the most appropriate instrument, to achieve distributive justice.
On the entitlement conception of justice in holdings that we have
presented, there is no argument based upon the first two principles of
distributive justice, the principles of acquisition and of transfer, for
such a more extensive state. If the set of holdings is properly genunequal distribution. The inefficiency could be removed by shifting from the
simple difference principle to a staggered difference principle, which recommends the maximization of the position of the least well-off group, and subject
to that constraint the maximization of the position of the next least well-off
group, etc. This point also is made by A. K. Sen, ibid., page 138, footnote, and
is acknowledged by Rawls on page 83. But such a staggered principle does not
embody a presumption in favor of equality of the sort used by Rawls. How then
could Rawls justify an inequality special to the staggered principle to someone
in the least well-off group? Perhaps these issues underlie the unclarity (see p.
83) as to whether Rawls accepts the staggered principle.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions
I26 Philosophy & Public Affairs
erated, there is no argument for a more extensive state based upon
distributive justice. (Nor, we have claimed, will the Lockean proviso actually provide occasion for a more extensive state.) If, however, these principles were violated, the principle of rectification
comes into play. Perhaps it is best to view some patterned principles of distributive justice as rough rules of thumb meant to approximate the general results of applying the principle of rectification of
injustice. For example, lacking much historical information, and assuming that victims of injustice generally do worse than they otherwise would, and (another assumption) that those from the least welloff group in the society have the highest probabilities of being the
(descendants of) victims of the most serious injustice who are owed
compensation by those who benefitted from the injustices (assumed
to be those better off, though sometimes the perpetrators will be others
in the worst-off group), then a rough rule of thumb for rectifying injustices might seem to be: organize society so as to maximize the
position of whatever group ends up least well off in the society. This
particular example may well be implausible, but an important question for each society will be: given its particular history, what operable
rule of thumb best approximates the results of a detailed application in
that society of the principle of rectification? These issues are very
complex, and are best left to a full treatment of the principle of rectification. In the absence of such a treatment applied to a particular
society, one cannot use the analysis and theory presented here, to
condemn any particular scheme of transfer payments, unless it is
clear that no considerations of rectification of injustice could apply to
justify it. While to introduce socialism as the punishment for our sins
would be to go too far, past injustices might be so great as to make
a more extensive state necessary in the short run in order to rectify
them.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:22:57 PM
All use subject to JSTOR Terms and Conditions

Justice as Fairness: Political not Metaphysical
Author(s): John Rawls
Source: Philosophy & Public Affairs, Vol. 14, No. 3 (Summer, 1985), pp. 223-251
Published by: Wiley
Stable URL: http://www.jstor.org/stable/2265349 .
Accessed: 20/02/2014 16:19
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
Princeton University Press and Wiley are collaborating with JSTOR to digitize, preserve and extend access to
Philosophy &Public Affairs.
http://www.jstor.org
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
JOHN RAWLS Justice as Fairness:
Political not
Metaphysical
In this discussion I shall make some general remarks about how I now
understand the conception of justice that I have called “justice as fairness” (presented in my book A Theory of Justice)., I do this because it
may seem that this conception depends on philosophical claims I should
like to avoid, for example, claims to universal truth, or claims about the
essential nature and identity of persons. My aim is to explain why it does
not. I shall first discuss what I regard as the task of political philosophy
at the present time and then briefly survey how the basic intuitive ideas
drawn upon in justice as fairness are combined into a political conception
of justice for a constitutional democracy. Doing this will bring out how
and why this conception of justice avoids certain philosophical and metaphysical claims. Briefly, the idea is that in a constitutional democracy the
public conception of justice should be, so far as possible, independent of
controversial philosophical and religious doctrines. Thus, to formulate
such a conception, we apply the principle of toleration to philosophy itself:
the public conception of justice is to be political, not metaphysical. Hence
the title.
I want to put aside the question whether the text of A Theory of Justice
supports different readings than the one I sketch here. Certainly on a
Beginning in November of I983, different versions of this paper were presented at New
York University, the Yale Law School Legal Theory Workshop, the University of Illinois,
and the University of California at Davis. I am grateful to many people for clarifying
numerous points and for raising instructive difficulties; the paper is much changed as a
result. In particular, I am indebted to Arnold Davidson, B. J. Diggs, Catherine Elgin, Owen
Fiss, Stephen Holmes, Norbert Hornstein, Thomas Nagel, George Priest, and David Sachs;
and especially to Burton Dreben who has been of very great help throughout. Indebtedness
to others on particular points is indicated in the footnotes.
i. Cambridge, MA: Harvard University Press, I97I.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
224 Philosophy & Public Affairs
number of points I have changed my views, and there are no doubt others
on which my views have changed in ways that I am unaware of.2 J
recognize further that certain faults of exposition as well as obscure and
ambiguous passages in A Theory ofJustice invite misunderstanding; but
I think these matters need not concern us and I shan’t pursue them
beyond a few footnote indications. For our purposes here, it suffices first,
to show how a conception of justice with the structure and content of
justice as fairness can be understood as political and not metaphysical,
and second, to explain why we should look for such a conception of justice
in a democratic society.
I
One thing I failed to say in A Theory of Justice, or failed to stress sufficiently, is that justice as fairness is intended as a political conception of
justice. While a political conception of justice is, of course, a moral conception, it is a moral conception worked out for a specific kind of subject,
namely, for political, social, and economic institutions. In particular, justice as fairness is framed to apply to what I have called the “basic structure” of a modern constitutional democracy.3 (I shall use “constitutional
2. A number of these changes, or shifts of emphasis, are evident in three lectures entitled
“Kantian Constructivism in Moral Theory,” Journal of Philosophy 77 (September I980).
For example, the account of what I have called “primary goods” is revised so that it clearly
depends on a particular conception of persons and their higher-order interests; hence this
account is not a purely psychological, sociological, or historical thesis. See pp. 526f. There
is also throughout those lectures a more explicit emphasis on the role of a conception of
the person as well as on the idea that the justification of a conception of justice is a practical
social task rather than an epistemological or metaphysical problem. See pp. 5I8f. And in
this connection the idea of “Kantian constructivism” is introduced, especially in the third
lecture. It must be noted, however, that this idea is not proposed as Kant’s idea: the adjective
“Kantian” indicates analogy not identity, that is, resemblance in enough fundamental respects so that the adjective is appropriate. These fundamental respects are certain structural
features of justice as fairness and elements of its content, such as the distinction between
what may be called the Reasonable and the Rational, the priority of right, and the role of
the conception of the persons as free and equal, and capable of autonomy, and so on.
Resemblances of structural features and content are not to be mistaken for resemblances
with Kant’s views on questions of epistemology and metaphysics. Finally, I should remark
that the title of those lectures, “Kantian Constructivism in Moral Theory,” was misleading;
since the conception of justice discussed is a political conception, a better title would have
been “Kantian Constructivism in Political Philosophy.” Whether constructivism is reasonable for moral philosophy is a separate and more general question.
3. Theory, Sec. 2, and see the index; see also “The Basic Structure as Subject,” in Values
and Morals, eds. Alvin Goldman and Jaegwon Kim (Dordrecht: Reidel, I978), pp. 47-7I.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
225 Justice as Fairness
democracy” and “democratic regime,” and similar phrases interchangeably.) By this structure I mean such a society’s main political, social, and
economic institutions, and how they fit together into one unified system
of social cooperation. Whether justice as fairness can be extended to a
general political conception for different kinds of societies existing under
different historical and social conditions, or whether it can be extended
to a general moral conception, or a significant part thereof, are altogether
separate questions. I avoid prejudging these larger questions one way or
the other.
It should also be stressed that justice as fairness is not intended as the
application of a general moral conception to the basic structure of society,
as if this structure were simply another case to which that general moral
conception is applied.4 In this respect justice as fairness differs from
traditional moral doctrines, for these are widely regarded as such general
conceptions. Utilitarianism is a familiar example, since the principle of
utility, however it is formulated, is usually said to hold for all kinds of
subjects ranging from the actions of individuals to the law of nations.
The essential point is this: as a practical political matter no general moral
conception can provide a publicly recognized basis for a conception of
justice in a modern democratic state. The social and historical conditions
of such a state have their origins in the Wars of Religion following the
Reformation and the subsequent development of the principle of toleration, and in the growth of constitutional government and the institutions
of large industrial market economies. These conditions profoundly affect
the requirements of a workable conception of political justice: such a
conception must allow for a diversity of doctrines and the plurality of
conflicting, and indeed incommensurable, conceptions of the good affirmed by the members of existing democratic societies.
Finally, to conclude these introductory remarks, since justice as fairness is intended as a political conception of justice for a democratic
society, it tries to draw solely upon basic intuitive ideas that are embedded
in the political institutions of a constitutional democratic regime and the
public traditions of their interpretation. Justice as fairness is a political
conception in part because it starts from within a certain political tradition. We 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
4. See “Basic Structure as Subject,” ibid., pp. 48-50.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
226 Philosophy & Public Affairs
doctrines likely to persist and to gain adherents in a more or less just
constitutional democratic society.5
II
There are, of course, many ways in which political philosophy may be
understood, and writers at different times, faced with different political
and social circumstances, understand their work differently. Justice as
fairness I would now understand as a reasonably systematic and practicable conception of justice for a constitutional democracy, a conception
that offers an alternative to the dominant utilitarianism of our tradition
of political thought. Its first task is to provide a more secure and acceptable
basis for constitutional principles and basic rights and liberties than utilitarianism seems to allow.6 The need for such a political conception arises
in the following way.
There are periods, sometimes long periods, in the history of any society
during which certain fundamental questions give rise to sharp and divisive political controversy, and it seems difficult, if not impossible, to
find any shared basis of political agreement. Indeed, certain questions
may prove intractable and may never be fully settled. One task of political
philosophy in a democratic society is to focus on such questions and to
examine whether some underlying basis of agreement can be uncovered
and a mutually acceptable way of resolving these questions publicly established. Or if these questions cannot be fully settled, as may well be
the case, perhaps the divergence of opinion can be narrowed sufficiently
so that political cooperation on a basis of mutual respect can still be
maintained.7
5. This idea was introduced in Theory, pp. 387f., as a way to weaken the conditions for
the reasonableness of civil disobedience in a nearly just democratic society. Here and later
in Secs. VI and VII it is used in a wider context.
6. Theory, Preface, p. viii.
7. Ibid., pp. 582f. On the role of a conception of justice in reducing the divergence of
opinion, see pp. 44f., 53, 3I4, and 564. At various places the limited aims in developing a
conception of justice are noted: see p. 364 on not expecting too much of an account of
civil disobedience; pp. 200f. on the inevitable indeterminacy of a conception of justice in
specifying a series of points of view from which questions of justice can be resolved; pp.
89f. on the social wisdom of recognizing that perhaps only a few moral problems (it would
have been better to say: problems of political justice) can be satisfactorily settled, and thus
of framing institutions so that intractable questions do not arise; on pp. 53, 87ff., 320f. the
need to accept simplifications is emphasized. Regarding the last point, see also “Kantian
Constructivism,” pp. 560-64.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
227 Justice as Fairness
The course of democratic thought over the past two centuries or so
makes plain that there is no agreement on the way basic institutions of
a constitutional democracy should be arranged if they are to specify and
secure the basic rights and liberties of citizens and answer to the claims
of democratic equality when citizens are conceived as free and equal
persons (as explained in the last three paragraphs of Section III). A deep
disagreement exists as to how the values of liberty and equality are best
realized in the basic structure of society. To simplify, we may think of
this disagreement as a conflict within the tradition of democratic thought
itself, between the tradition associated with Locke, which gives greater
weight to what Constant called “the liberties of the moderns,” freedom
of thought and conscience, certain basic rights of the person and of
property, and the rule of law, and the tradition associated with Rousseau,
which gives greater weight to what Constant called “the liberties of the
ancients,” the equal political liberties and the values of public life. This
is a stylized contrast and historically inaccurate, but it serves to fix ideas.
Justice as fairness tries to adjudicate between these contending traditions first, by proposing two principles of justice to serve as guidelines
for how basic institutions are to realize the values of liberty and equality,
and second, by specifying a point of view from which these principles
can be seen as more appropriate than other familiar principles of justice
to the nature of democratic citizens viewed as free and equal persons.
What it means to view citizens as free and equal persons is, of course, a
fundamental question and is discussed in the following sections. What
must be shown is that a certain arrangement of the basic structure,
certain institutional forms, are more appropriate for realizing the values
of liberty and equality when citizens are conceived as such persons, that
is (very briefly), as having the requisite powers of moral personality that
enable them to participate in society viewed as a system of fair cooperation
for mutual advantage. So to continue, the two principles of justice (mentioned above) read as follows:
i. Each person has an equal right to a fully adequate scheme of equal
basic rights and liberties, which scheme is compatible with a similar
scheme 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 greatest benefit of the least advantaged members of society.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
228 Philosophy & Public Affairs
Each of these principles applies to a different part of the basic structure;
and both are concerned not only with basic rights, liberties, and opportunities, but also with the claims of equality; while the second part of
the second principle underwrites the worth of these institutional guarantees.8 The two principles together, when the first is given priority over
the second, regulate the basic institutions which realize these values.9
But these details, although important, are not our concern here.
We must now ask: how might political philosophy find a shared basis
for settling such a fundamental question as that of the most appropriate
institutional forms for liberty and equality? Of course, it is likely that the
most that can be done is to narrow the range of public disagreement. Yet
even firmly held convictions gradually change: religious toleration is now
accepted, and arguments for persecution are no longer openly professed;
similarly, slavery is rejected as inherently unjust, and however much the
aftermath of slavery may persist in social practices and unavowed attitudes, no one is willing to defend it. We collect such settled convictions
as the belief in religious toleration and the rejection of slavery and try to
organize the basic ideas and principles implicit in these convictions into
a coherent conception of justice. We can regard these convictions as
provisional fixed points which any conception of justice must account
for if it is to be reasonable for us. We look, then, to our public political
culture itself, including its main institutions and the historical traditions
of their interpretation, as the shared fund of implicitly recognized basic
ideas and principles. The hope is that these ideas and principles can be
formulated clearly enough to be combined into a conception of political
justice congenial to our most firmly held convictions. We express this by
saying that a political conception of justice, to be acceptable, must be in
accordance with our considered convictions, at all levels of generality, on
due reflection (or in what I have called “reflective equilibrium”).’0
The public political culture may be of two minds even at a very deep
8. The statement of these principles differs from that given in Theory and follows the
statement in “The Basic Liberties and Their Priority,” Tanner Lectures on Human Values,
Vol. III (Salt Lake City: University of Utah Press, i982), p. 5. The reasons for the changes
are discussed at pp. 46-55 of that lecture. They are important for the revisions made in
the account of the basic liberties found in Theory in the attempt to answer the objections
of H.L.A. Hart; but they need not concem us here.
9. The idea of the worth of these guarentees is discussed ibid., pp. 40f.
io. Theory, pp. 20f., 48-51, and I20f.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
229 Justice as Fairness
level. Indeed, this must be so with such an enduring controversy as that
concerning the most appropriate institutional forms to realize the values
of liberty and equality. This suggests that if we are to succeed in finding
a basis of public agreement, we must find a new way of organizing familiar
ideas and principles into a conception of political justice so that the claims
in conflict, as previously understood, are seen in another light. A political
conception need not be an original creation but may only articulate familiar intuitive ideas and principles so that they can be recognized as
fitting together in a somewhat different way than before. Such a conception may, however, go further than this: it may organize these familiar
ideas and principles by means of a more fundamental intuitive idea within
the complex structure of which the other familiar intuitive ideas are then
systematically connected and related. In justice as fairness, as we shall
see in the next section, this more fundamental idea is that of society as
a system of fair social cooperation between free and equal persons. The
concern of this section is how we might find a public basis of political
agreement. The point is that a conception of justice will only be able to
achieve this aim if it provides a reasonable way of shaping into one
coherent view the deeper bases of agreement embedded in the public
political culture of a constitutional regime and acceptable to its most
firmly held considered convictions.
Now suppose justice as fairness were to achieve its aim and a publicly
acceptable political conception of justice is found. Then this conception
provides a publicly recognized point of view from which all citizens can
examine before one another whether or not their political and social
institutions are just. It enables them to do this by citing what are recognized among them as valid and sufficient reasons singled out by that
conception itself. Society’s main institutions and how they fit together
into one scheme of social cooperation can be examined on the same basis
by each citizen, whatever that citizen’s social position or more particular
interests. It should be observed that, on this view, justification is not
regarded simply as valid argument from listed premises, even should
these premises be true. Rather, justification is addressed to others who
disagree with us, and therefore it must always proceed from some consensus, that is, from premises that we and others publicly recognize as
true; or better, publicly recognize as acceptable to us for the purpose of
establishing a working agreement on the fundamental questions of political justice. It goes without saying that this agreement must be inThis content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
230 Philosophy & Public Affairs
formed and uncoerced, and reached by citizens in ways consistent with
their being viewed as free and equal persons.”
Thus, the aim of justice as fairness as a political conception is practical,
and not metaphysical or epistemological. That is, it presents itself not as
a conception of justice that is true, but one that can serve as a basis of
informed and willing political agreement between citizens viewed as free
and equal persons. This agreement when securely founded in public
political and social attitudes sustains the goods of all persons and associations within a just democratic regime. To secure this agreement we
try, so far as we can, to avoid disputed philosophical, as well as disputed
moral and religious, questions. We do this not because these questions
are unimportant or regarded with indifference,12 but because we think
them too important and recognize that there is no way to resolve them
politically. The only alternative to a principle of toleration is the autocratic
use of state power. Thus, justice as fairness deliberately stays on the
surface, philosophically speaking. Given the profound differences in belief
and conceptions of the good at least since the Reformation, we must
recognize that, just as on questions of religious and moral doctrine, public
agreement on the basic questions of philosophy cannot be obtained without the state’s infringement of basic liberties. Philosophy as the search
for truth about an independent metaphysical and moral order cannot, I
believe, provide a workable and shared basis for a political conception of
justice in a democratic society.
We try, then, to leave aside philosophical controversies whenever possible, and look for ways to avoid philosophy’s longstanding problems.
Thus, in what I have called “Kantian constructivism,” we try to avoid the
problem of truth and the controversy between realism and subjectivism
about the status of moral and political values. This form of constructivism
neither asserts nor denies these doctrines.13 Rather, it recasts ideas from
the tradition of the social contract to achieve a practicable conception of
objectivity and justification founded on public agreement in judgment
on due reflection. The aim is free agreement, reconciliation through public reason. And similarly, as we shall see (in Section V), a conception of
the person in a political view, for example, the conception of citizens as
i i. Ibid., pp. 580-83.
I2. Ibid., pp. 2I4f.
I3. On Kantian constructivism, see especially the third lecture referred to in footnote 2
above.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
23I Justice as Fairness
free and equal persons, need not involve, so I believe, questions of philosophical psychology or a metaphysical doctrine of the nature of the self.
No political view that depends on these deep and unresolved matters can
serve as a public conception of justice in a constitutional democratic state.
As I have said, we must apply the principle of toleration to philosophy
itself. The hope is that, by this method of avoidance, as we might call it,
existing differences between contending political views can at least be
moderated, even if not entirely removed, so that social cooperation on the
basis of mutual respect can be maintained. Or if this is expecting too
much, this method may enable us to conceive how, given a desire for
free and uncoerced agreement, a public understanding could arise consistent with the historical conditions and constraints of our social world.
Until we bring ourselves to conceive how this could happen, it can’t
happen.
III
Let’s now survey briefly some of the basic ideas that make up justice as
fairness in order to show that these ideas belong to a political conception
of justice. As I have indicated, the overarching fundamental intuitive
idea, within which other basic intuitive ideas are systematically connected, is that of society as a fair system of cooperation between free and
equal persons. Justice as fairness starts from this idea as one of the basic
intuitive ideas which we take to be implicit in the public culture of a
democratic society.14 In their political thought, and in the context of
public discussion of political questions, citizens do not view the social
order as a fixed natural order, or as an institutional hierarchy justified
by religious or aristocratic values. Here it is important to stress that from
other points of view, for example, from the point of view of personal
morality, or from the point of view of members of an association, or of
one’s religious or philosophical doctrine, various aspects of the world and
one’s relation to it, may be regarded in a different way. But these other
points of view are not to be introduced into political discussion.
We can make the idea of social cooperation more specific by noting
three of its elements:
I4. Although Theory uses this idea from the outset (it is introduced on p. 4), it does not
emphasize, as I do here and in “Kantian Constructivism,” that the basic ideas of justice as
faimess are regarded as implicit or latent in the public culture of a democratic society.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
232 Philosophy & Public Affairs
i. Cooperation is distinct from merely socially coordinated activity, for
example, from activity coordinated by orders issued by some central
authority. Cooperation is guided by publicly recognized rules and
procedures which those who are cooperating accept and regard as
properly regulating their conduct.
2. Cooperation involves the idea of fair terms of cooperation: these are
terms that each participant may reasonably accept, provided that
everyone else likewise accepts them. Fair terms of cooperation specify an idea of reciprocity or mutuality: all who are engaged in cooperation and who do their part as the rules and procedures require,
are to benefit in some appropriate way as assessed by a suitable
benchmark of comparison. A conception of political justice characterizes the fair terms of social cooperation. Since the primary
subject of justice is the basic structure of society, this is accomplished in justice as fairness by formulating principles that specify
basic rights and duties within the main institutions of society, and
by regulating the institutions of background justice over time so
that the benefits produced by everyone’s efforts are fairly acquired
and divided from one generation to the next.
3. The idea of social cooperation requires an idea of each participant’s
rational advantage, or good. This idea of good specifies what those
who are engaged in cooperation, whether individuals, families, or
associations, or even nation-states, are trying to achieve, when the
scheme is viewed from their own standpoint.
Now consider the idea of the person.15 There are, of course, many
aspects of human nature that can be singled out as especially significant
depending on our point of view. This is witnessed by such expressions
as homo politicus, homo oeconomicus, homo faber, and the like. Justice
as fairness starts from the idea that society is to be conceived as a fair
I5. It should be emphasized that a conception of the person, as I understand it here, is
a normative conception, whether legal, political, or moral, or indeed also philosophical or
religious, depending on the overall view to which it belongs. In this case the conception
of the person is a moral conception, one that begins from our everyday conception of persons
as the basic units of thought, deliberation and responsibility, and adapted to a political
conception of justice and not to a comprehensive moral doctrine. It is in effect a political
conception of the person, and given the aims of justice as fairness, a conception of citizens.
Thus, a conception of the person is to be distinguished from an account of human nature
given by natural science or social theory. On this point, see “Kantian Constructivism,” pp.
534f.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
233 Justice as Fairness
system of cooperation and so it adopts a conception of the person to go
with this idea. Since Greek times, both in philosophy and law, the concept
of the person has been understood as the concept of someone who can
take part in, or who can play a role in, social life, and hence exercise and
respect its various rights and duties. Thus, we say that a person is someone who can be a citizen, that is, a fully cooperating member of society
over a complete life. We add the phrase “over a complete life” because
a society is viewed as a more or less complete and self-sufficient scheme
of cooperation, making room within itself for all the necessities and activities of life, from birth until death. A society is not an association for
more limited purposes; citizens do not join society voluntarily but are
born into it, where, for our aims here, we assume they are to lead their
lives.
Since we start within the tradition of democratic thought, we also think
of citizens as free and equal persons. The basic intuitive idea is that in
virtue of what we may call their moral powers, and the powers of reason,
thought, and judgment connected with those powers, we say that persons
are free. And in virtue of their having these powers to the requisite degree
to be fully cooperating members of society, we say that persons are
equal.’6 We can elaborate this conception of the person as follows. Since
persons can be full participants in a fair system of social cooperation, we
ascribe to them the two moral powers connected with the elements in
the idea of social cooperation noted above: namely, a capacity for a sense
of justice and a capacity for a conception of the good. A sense of justice
is the capacity to understand, to apply, and to act from the public conception of justice which characterizes the fair terms of social cooperation.
The capacity for a conception of the good is the capacity to form, to revise,
and rationally to pursue a conception of one’s rational advantage, or good.
In the case of social cooperation, this good must not be understood narrowly but rather as a conception of what is valuable in human life. Thus,
a conception of the good normally consists of a more or less determinate
scheme of final ends, that is, ends we want to realize for their own sake,
as well as of attachments to other persons and loyalties to various groups
and associations. These attachments and loyalties give rise to affections
and devotions, and therefore the flourishing of the persons and associations who are the objects of these sentiments is also part of our coni6. Theory, Sec. 77.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
234 Philosophy & Public Affairs
ception of the good. Moreover, we must also include in such a conception
a view of our relation to the world-religious, philosophical, or moralby reference to which the value and significance of our ends and attachments are understood.
In addition to having the two moral powers, the capacities for a sense
of justice and a conception of the good, persons also have at any given
time a particular conception of the good that they try to achieve. Since
we wish to start from the idea of society as a fair system of cooperation,
we assume that persons as citizens have all the capacities that enable
them to be normal and fully cooperating members of society. This does
not imply that no one ever suffers from illness or accident; such misfortunes are to be expected in the ordinary course of human life; and provision for these contingencies must be made. But for our purposes here
I leave aside permanent physical disabilities or mental disorders so severe
as to prevent persons from being normal and fully cooperating members
of society in the usual sense.
Now the conception of persons as having the two moral powers, and
therefore as free and equal, is also a basic intuitive idea assumed to be
implicit in the public culture of a democratic society. Note, however, that
it is formed by idealizing and simplifying in various ways. This is done
to achieve a clear and uncluttered view of what for us is the fundamental
question of political justice: namely, what is the most appropriate conception of justice for specifying the terms of social cooperation between
citizens regarded as free and equal persons, and as normal and fully
cooperating members of society over a complete life. It is this question
that has been the focus of the liberal critique of aristocracy, of the socialist
critique of liberal constitutional democracy, and of the conflict between
liberals and conservatives at the present time over the claims of private
property and the legitimacy (in contrast to the effectiveness) of social
policies associated with the so-called welfare state.
IV
I now take up the idea of the original position.’7 This idea is introduced
in order to work out which traditional conception of justice, or which
variant of one of those conceptions, specifies the most appropriate prin17. Ibid., Sec. 4, Ch. 3, and the index.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
235 Justice as Fairness
ciples for realizing liberty and equality once society is viewed as a system
of cooperation between free and equal persons. Assuming we had this
purpose in mind, let’s see why we would introduce the idea of the original
position and how it serves its purpose.
Consider again the idea of social cooperation. Let’s ask: how are the
fair terms of cooperation to be determined? Are they simply laid down by
some outside agency distinct from the persons cooperating? Are they, for
example, laid down by God’s law? Or are these terms to be recognized
by these persons as fair by reference to their knowledge of a prior and
independent moral order? For example, are they regarded as required by
natural law, or by a realm of values known by rational intuition? Or are
these terms to be established by an undertaking among these persons
themselves in the light of what they regard as their mutual advantage?
Depending on which answer we give, we get a different conception of
cooperation.
Since justice as fairness recasts the doctrine of the social contract, it
adopts a form of the last answer: the fair terms of social cooperation are
conceived as agreed to by those engaged in it, that is, by free and equal
persons as citizens who are born into the society in which they lead their
lives. But their agreement, like any other valid agreement, must be entered into under appropriate conditions. In particular, these conditions
must situate free and equal persons fairly and must not allow some persons greater bargaining advantages than others. Further, threats of force
and coercion, deception and fraud, and so on, must be excluded.
So far so good. The foregoing considerations are familiar from everyday
life. But agreements in everyday life are made in some more or less clearly
specified situation embedded within the background institutions of the
basic structure. Our task, however, is to extend the idea of agreement to
this background framework itself. Here we face a difficulty for any political conception of justice that uses the idea of a contract, whether social
or otherwise. The difficulty is this: we must find some point of view,
removed from and not distorted by the particular features and circumstances of the all-encompassing background framework, from which a
fair agreement between free and equal persons can be reached. The
original position, with the feature I have called “the veil of ignorance,”
is this point of view.”8 And the reason why the original position must
i8. On the veil of ignorance, see ibid., Sec. 24, and the index.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
236 Philosophy & Public Affairs
abstract from and not be affected by the contingencies of the social world
is that the conditions for a fair agreement on the principles of political
justice between free and equal persons must eliminate the bargaining
advantages which inevitably arise within background institutions of any
society as the result of cumulative social, historical, and natural tendencies. These contingent advantages and accidental influences from the
past should not influence an agreement on the principles which are to
regulate the institutions of the basic structure itself from the present into
the future.
Here we seem to face a second difficulty, which is, however, only
apparent. To explain: from what we have just said it is clear that the
original position is to be seen as a device of representation and hence
any agreement reached by the parties must be regarded as both hypothetical and nonhistorical. But if so, since hypothetical agreements cannot
bind, what is the significance of the original position?I9 The answer is
ig. This question is raised by Ronald Dworkin in the first part of his very illuminating,
and to me highly instructive, essay “Justice and Rights” (I973), reprinted in Taking Rights
Seriously (Cambridge, MA: Harvard University Press, I977). Dworkin considers several
ways of explaining the use of the original position in an account of justice that invokes the
idea of the social contract. In the last part of the essay (pp. I73-83), after having surveyed
some of the constructivist features of justice as fairness (pp. I59-68) and argued that it is
a right-based and not a duty-based or a goal-based view (pp. I68-77), he proposes that the
original position with the veil of ignorance be seen as modeling the force of the natural
right that individuals have to equal concern and respect in the design of the political
institutions that govern them (p. i8o). He thinks that this natural right lies as the basis of
justice as fairness and that the original position serves as a device for testing which principles
of justice this right requires. This is an ingenious suggestion but I have not followed it in
the text. I prefer not to think of justice as fairness as a right-based view; indeed, Dworkin’s
classification scheme of right-based, duty-based and goal-based views (pp. I7If.) is too
narrow and leaves out important possibilities. Thus, as explained in Sec. II above, I think
of justice as fairness as working up into idealized conceptions certain fundamental intuitive
ideas such as those of the person as free and equal, of a well-ordered society and of the
public role of a conception of political justice, and as connecting these fundamental intuitive
ideas with the even more fundamental and comprehensive intuitive idea of society as a fair
system of cooperation over time from one generation to the next. Rights, duties, and goals
are but elements of such idealized conceptions. Thus, justice as fairness is a conceptionbased, or as Elizabeth Anderson has suggested to me, an ideal-based view, since these
fundamental intuitive ideas reflect ideals implicit or latent in the public culture of a democratic society. In this context the original position is a device of representation that models
the force, not of the natural right of equal concern and respect, but of the essential elements
of these fundamental intuitive ideas as identified by the reasons for principles of justice
that we accept on due reflection. As such a device, it serves first to combine and then to
focus the resultant force of all these reasons in selecting the most appropriate principles
of justice for a democratic society. (In doing this the force of the natural right of equal
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
237 Justice as Fairness
implicit in what has already been said: it is given by the role of the various
features of the original position as a device of representation. Thus, that
the parties are symmetrically situated is required if they are to be seen
as representatives of free and equal citizens who are to reach an agreement under conditions that are fair. Moreover, one of our considered
convictions, I assume, is this: the fact that we occupy a particular social
position is not a good reason for us to accept, or to expect others to accept,
a conception of justice that favors those in this position. To model this
conviction in the original position the parties are not allowed to know
their social position; and the same idea is extended to other cases. This
is expressed figuratively by saying that the parties are behind a veil of
ignorance. In sum, the original position is simply a device of representation: it describes the parties, each of whom are responsible for the
essential interests of a free and equal person, as fairly situated and as
reaching an agreement subject to appropriate restrictions on what are to
count as good reasons.20
Both of the above mentioned difficulties, then, are overcome by viewing
the original position as a device of representation: that is, this position
models what we regard as fair conditions under which the representatives
concern and respect will be covered in other ways.) This account of the use of the original
position resembles in some respects an account Dworkin rejects in the first part of his
essay, especially pp. I53f. In view of the ambiguity and obscurity of Theory on many of
the points he considers, it is not my aim to criticize Dworkin’s valuable discussion, but
rather to indicate how my understanding of the original position differs from his. Others
may prefer his account.
20. The original position models a basic feature of Kantian constructivism, namely, the
distinction between the Reasonable and the Rational, with the Reasonable as prior to the
Rational. (For an explanation of this distinction, see “Kantian Constructivism,” pp. 528-
32, and passim.) The relevance of this distinction here is that Theory more or less consistently speaks not of rational but of reasonable (or sometimes of fitting or appropriate)
conditions as constraints on arguments for principles of justice (see pp. i8f., 20f., I20f.,
I3of., I38, 446, 5i6f., 578, 584f.). These constraints are modeled in the original position
and thereby imposed on the parties: their deliberations are subject, and subject absolutely,
to the reasonable conditions the modeling of which makes the original position fair. The
Reasonable, then, is prior to the Rational, and this gives the priority of right. Thus, it was
an error in Theory (and a very misleading one) to describe a theory of justice as part of the
theory of rational choice, as on pp. i 6 and 583. What I should have said is that the conception
of justice as fairness uses an account of rational choice subject to reasonable conditions to
characterize the deliberations of the parties as representives of free and equal persons; and
all of this within a political conception of justice, which is, of course, a moral conception.
There is no thought of trying to derive the content of justice within a framework that uses
an idea of the rational as the sole normative idea. That thought is incompatible with any
kind of Kantian view.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
238 Philosophy & Public Affairs
of free and equal persons are to specify the terms of social cooperation
in the case of the basic structure of society; and since it also models
what, for this case, we regard as acceptable restrictions on reasons available to the parties for favoring one agreement rather than another, the
conception of justice the parties would adopt identifies the conception
we regard-here and now-as fair and supported by the best reasons.
We try to model restrictions on reasons in such a way that it is perfectly
evident which agreement would be made by the parties in the original
position as citizens’ representatives. Even if there should be, as surely
there will be, reasons for and against each conception of justice available,
there may be an overall balance of reasons plainly favoring one conception
over the rest. As a device of representation the idea of the original position
serves as a means of public reflection and self-clarification. We can use
it to help us work out what we now think, once we are able to take a
clear and uncluttered view of what justice requires when society is conceived as a scheme of cooperation between free and equal persons over
time from one generation to the next. The original position serves as a
unifying idea by which our considered convictions at all levels of generality are brought to bear on one another so as to achieve greater mutual
agreement and self-understanding.
To conclude: we introduce an idea like that of the original position
because there is no better way to elaborate a political conception of justice
for the basic structure from the fundamental intuitive idea of society as
a fair system of cooperation between citizens as free and equal persons.
There are, however, certain hazards. As a device of representation the
original position is likely to seem somewhat abstract and hence open to
misunderstanding. The description of the parties may seem to presuppose
some metaphysical conception of the person, for example, that the essential nature of persons is independent of and prior to their contingent
attributes, including their final ends and attachments, and indeed, their
character as a whole. But this is an illusion caused by not seeing the
original position as a device of representation. The veil of ignorance, to
mention one prominent feature of that position, has no metaphysical
implications concerning the nature of the self; it does not imply that the
self is ontologically prior to the facts about persons that the parties are
excluded from knowing. We can, as it were, enter this position any time
simply by reasoning for principles of justice in accordance with the enuThis content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
239 Justice as Fairness
merated restrictions. When, in this way, we simulate being in this position, our reasoning no more commits us to a metaphysical doctrine
about the nature of the self than our playing a game like Monopoly
commits us to thinking that we are landlords engaged in a desperate
rivalry, winner take all.21 We must keep in mind that we are trying to
show how the idea of society as a fair system of social cooperation can
be unfolded so as to specify the most appropriate principles for realizing
the institutions of liberty and equality when citizens are regarded as free
and equal persons.
V
I just remarked that the idea of the original position and the description
of the parties may tempt us to think that a metaphysical doctrine of the
person is presupposed. While I said that this interpretation is mistaken,
it is not enough simply to disavow reliance on metaphysical doctrines,
for despite one’s intent they may still be involved. To rebut claims of this
2I. Theory, pp. I38f., I47. The parties in the original position are said (p. I47) to be
theoretically defined individuals whose motivations are specified by the account of that
position and not by a psychological view about how human beings are actually motivated.
This is also part of what is meant by saying (p. I2I) that the acceptance of the particular
principles of justice is not conjectured as a psychological law or probability but rather
follows from the full description of the original position. Although the aim cannot be perfectly
achieved, we want the argument to be deductive, “a kind of moral geometry.” In “Kantian
Constructivism” (p. 532) the parties are described as merely artificial agents who inhabit
a construction. Thus I think R. B. Brandt mistaken in objecting that the argument from
the original position is based on defective psychology. See his A Theory of the Good and
the Right (Oxford: Clarendon Press, I979), pp. 239-42. Of course, one might object to the
original position that it models the conception of the person and the deliberations of the
parties in ways that are unsuitable for the purposes of a political conception of justice; but
for these purposes psychological theory is not directly relevant. On the other hand, psychological theory is relevant for the account of the stability of a conception of justice, as
discussed in Theory, Pt. III. See below, footnote 33. Similarly, I think Michael Sandel
mistaken in supposing that the original position involves a conception of the self “. . . shorn
of all its contingently-given attributes,” a self that “assumes a kind of supra-empirical status,
… and given prior to its ends, a pure subject of agency and possession, ultimately thin.”
See Liberalism and the Limits of Justice (Cambridge: Cambridge University Press, I982),
pp. 93-95. I cannot discuss these criticisms in any detail. The essential point (as suggested
in the introductory remarks) is not whether certain passages in Theory call for such an
interpretation (I doubt that they do), but whether the conception of justice as fairness
presented therein can be understood in the light of the interpretation Isketch in this article
and in the earlier lectures on constructivism, as I believe it can be.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
240 Philosophy & Public Affairs
nature requires discussing them in detail and showing that they have no
foothold. I cannot do that here.22
I can, however, sketch a positive account of the political conception of
the person, that is, the conception of the person as citizen (discussed in
Section III), involved in the original position as a device of representation.
To explain what is meant by describing a conception of the person as
political, let’s consider how citizens are represented in the original position as free persons. The representation of their freedom seems to be
one source of the idea that some metaphysical doctrine is presupposed.
I have said elsewhere that citizens view themselves as free in three
respects, so let’s survey each of these briefly and indicate the way in
which the conception of the person used is political.23
First, citizens are free in that they conceive of themselves and of one
another as having the moral power to have a conception of the good. This
is not to say that, as part of their political conception of themselves, they
view themselves as inevitably tied to the pursuit of the particular con22. Part of the difficulty is that there is no accepted understanding of what a metaphysical
doctrine is. One might say, as Paul Hoffman has suggested to me, that to develop a political
conception of justice without presupposing, or explicitly using, a metaphysical doctrine,
for example, some particular metaphysical conception of the person, is already to presuppose
a metaphysical thesis: namely, that no particular metaphysical doctrine is required for this
purpose. One might also say that our everyday conception of persons as the basic units of
deliberation and responsibility presupposes, or in some way involves, certain metaphysical
theses about the nature of persons as moral or political agents. Following the method of
avoidance, I should not want to deny these claims. What should be said is the following.
If we look at the presentation of justice as fairness and note how it is set up, and note the
ideas and conceptions it uses, no particular metaphysical doctrine about the nature of
persons, distinctive and opposed to other metaphysical doctrines, appears among its premises, or seems required by its argument. If metaphysical presuppositions are involved,
perhaps they are so general that they would not distinguish between the distinctive metaphysical views-Cartesian, Leibnizian, or Kantian; realist, idealist, or materialist-with
which philosophy traditionally has been concerned. In this case, they would not appear to
be relevant for the structure and content of a political conception of justice one way or the
other. I am grateful to Daniel Brudney and Paul Hoffman for discussion of these matters.
23. For the first two respects, see “Kantian Constructivism,” pp. 544f. (For the third
respect, see footnote 26 below.) The account of the first two respects found in those lectures
is further developed in the text above and I am more explicit on the distinction between
what I call here our “public” versus our “nonpublic or moral identity.” The point of the
term “moral” in the latter phrase is to indicate that persons’ conceptions of the (complete)
good are normally an essential element in characterizing their nonpublic (or nonpolitical)
identity, and these conceptions are understood as normally containing important moral
elements, although they include other elements as well, philosophical and religious. The
term “moral” should be thought of as a stand-in for all these possibilities. I am indebted to
Elizabeth Anderson for discussion and clarification of this distinction.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
241 Justice as Fairness
ception of the good which they affirm at any given time. Instead, as
citizens, they are regarded as capable of revising and changing this conception on reasonable and rational grounds, and they may do this if they
so desire. Thus, as free persons, citizens claim the right to view their
persons as independent from and as not identified with any particular
conception of the good, or scheme of final ends. Given their moral power
to form, to revise, and rationally to pursue a conception of the good, their
public identity as free persons is not affected by changes over time in
their conception of the good. For example, when citizens convert from
one religion to another, or no longer affirm an established religious faith,
they do not cease to be, for questions of political justice, the same persons
they were before. There is no loss of what we may call their public identity,
their identity as a matter of basic law. In general, they still have the same
basic rights and duties; they own the same property and can make the
same claims as before, except insofar as these claims were connected
with their previous religious affiliation. We can imagine a society (indeed,
history offers numerous examples) in which basic rights and recognized
claims depend on religious affiliation, social class, and so on. Such a
society has a different political conception of the person. It may not have
a conception of citizenship at all; for this conception, as we are using it,
goes with the conception of society as a fair system of cooperation for
mutual advantage between free and equal persons.
It is essential to stress that citizens in their personal affairs, or in the
internal life of associations to which they belong, may regard their final
ends and attachments in a way very different from the way the political
conception involves. Citizens may have, and normally do have at any
given time, affections, devotions, and loyalties that they believe they
would not, and indeed could and should not, stand apart from and objectively evaluate from the point of view of their purely rational good.
They may regard it as simply unthinkable to view themselves apart from
certain religious, philosophical, and moral convictions, or from certain
enduring attachments and loyalties. These convictions and attachments
are part of what we may call their “nonpublic identity.” These convictions
and attachments help to organize and give shape to a person’s way of
life, what one sees oneself as doing and trying to accomplish in one’s
social world. We think that if we were suddenly without these particular
convictions and attachments we would be disoriented and unable to carry
on. In fact, there would be, we might think, no point in carrying on. But
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
242 Philosophy & Public Affairs
our conceptions of the good may and often do change over time, usually
slowly but sometimes rather suddenly. When these changes are sudden,
we are particularly likely to say that we are no longer the same person.
We know what this means: we refer to a profound and pervasive shift,
or reversal, in our final ends and character; we refer to our different
nonpublic, and possibly moral or religious, identity. On the road to Damascus Saul of Tarsus becomes Paul the Apostle. There is no change in
our public or political identity, nor in our personal identity as this concept
is understood by some writers in the philosophy of mind.24
The second respect in which citizens view themselves as free is that
they regard themselves as self-originating sources of valid claims. They
think their claims have weight apart from being derived from duties or
obligations specified by the political conception of justice, for example,
from duties and obligations owed to society. Claims that citizens regard
as founded on duties and obligations based on their conception of the
good and the moral doctrine they affirm in their own life are also, for our
purposes here, to be counted as self-originating. Doing this is reasonable
in a political conception of julstice for a constitutional democracy; for
provided the conceptions of the good and the moral doctrines citizens
affirm are compatible with the public conception of justice, these duties
and obligations are self-originating from the political point of view.
When we describe a way in which citizens regard themselves as free,
24. Here I assume that an answer to the problem of personal identity tries to specify the
various criteria (for example, psychological continuity of memories and physical continuity
of body, or some part thereof) in accordance with which two different psychological states,
or actions (or whatever), which occur at two different times may be said to be states or
actions of the same person who endures over time; and it also tries to specify how this
enduring person is to be conceived, whether as a Cartesian or a Leibnizian substance, or
as a Kantian transcendental ego, or as a continuant of some other kind, for example, bodily
or physical. See the collection of essays edited by John Perry, Personal Identity (Berkeley,
CA: University of California Press, 1975), especially Perry’s introduction, pp. 3-30; and
Sydney Shoemaker’s essay in Personal Identity (Oxford: Basil Blackwell, I984), both of
which consider a number of views. Sometimes in discussions of this problem, continuity
of fundamental aims and aspirations is largely ignored, for example, in views like H. P.
Grice’s (included in Perry’s collection) which emphasizes continuity of memory. Of course,
once continuity of fundamental aims and aspirations is brought in, as in Derek Parfit’s
Reasons and Persons (Oxford: Clarendon Press, I984), Pt. III, there is no sharp distinction
between the problem of persons’ nonpublic or moral identity and the problem of their
personal identity. This latter problem raises profound questions on which past and current
philosophical views widely differ, and surely will continue to differ. For this reason it is
important to try to develop a political conception of justice which avoids this problem as
far as possible.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
243 Justice as Fairness
we are describing how citizens actually think of themselves in a democratic society should questions of justice arise. In our conception of a
constitutional regime, this is an aspect of how citizens regard themselves.
That this aspect of their freedom belongs to a particular political conception is clear from the contrast with a different political conception in
which the members of society are not viewed as self-originating sources
of valid claims. Rather, their claims have no weight except insofar as
they can be derived from their duties and obligations owed to society, or
from their ascribed roles in the social hierarchy justified by religious or
aristocratic values. Or to take an extreme case, slaves are human beings
who are not counted as sources of claims, not even claims based on social
duties or obligations, for slaves are not counted as capable of having
duties or obligations. Laws that prohibit the abuse and maltreatment of
slaves are not founded on claims made by slaves on their own behalf, but
on claims originating either from slaveholders, or from the general interests of society (which does not include the interests of slaves). Slaves
are, so to speak, socially dead: they are not publicly recognized as persons
at all.25 Thus, the contrast with a political conception which allows slavery
makes clear why conceiving of citizens as free persons in virtue of their
moral powers and their having a conception of the good, goes with a
particular political conception of the person. This conception of persons
fits into a political conception of justice founded on the idea of society
as a system of cooperation between its members conceived as free and
equal.
The third respect in which citizens are regarded as free is that they
are regarded as capable of taking responsibility for their ends and this
affects how their various claims are assessed.26 Very roughly, the idea is
that, given just background institutions and given for each person a fair
index of primary goods (as required by the principles of justice), citizens
are thought to be capable of adjusting their aims and aspirations in the
light of what they can reasonably expect to provide for. Moreover, they
are regarded as capable of restricting their claims in matters of justice
25. For the idea of social death, see Orlando Patterson, Slavery and Social Death (Cambridge, MA: Harvard University Press, i982), esp. pp. 5-9, 38-45, 337. This idea is interestingly developed in this book and has a central place in the author’s comparative study
of slavery.
26. See “Social Unity and Primary Goods,” in Utilitarianism and Beyond, eds. Amartya
Sen and Bernard Williams (Cambridge: Cambridge University Press, i982), Sec. IV, pp.
I 67-70.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
244 Philosophy & Public Affairs
to the kinds of things the principles of justice allow. Thus, citizens are
to recognize that the weight of their claims is not given by the strength
and psychological intensity of their wants and desires (as opposed to their
needs and requirements as citizens), even when their wants and desires
are rational from their point of view. I cannot pursue these matters here.
But the procedure is the same as before: we start with the basic intuitive
idea of society as a system of social cooperation. When this idea is developed into a conception of political justice, it implies that, viewing
ourselves as persons who can engage in social cooperation over a complete
life, we can also take responsibility for our ends, that is, that we can
adjust our ends so that they can be pursued by the means we can reasonably expect to acquire given our prospects and situation in society.
The idea of responsibility for ends is implicit in the public political culture
and discernible in its practices. A political conception of the person articulates this idea and fits it into the idea of society as a system of social
cooperation over a complete life.
To sum up, I recapitulate three main points of this and the preceding
two sections:
First, in Section III persons were regarded as free and equal in virtue
of their possessing to the requisite degree the two powers of moral personality (and the powers of reason, thought, and judgment connected
with these powers), namely, the capacity for a sense of justice and the
capacity for a conception of the good. These powers we associated with
two main elements of the idea of cooperation, the idea of fair terms of
cooperation and the idea of each participant’s rational advantage, or good.
Second, in this section (Section V), we have briefly surveyed three
respects in which persons are regarded as free, and we have noted that
in the public political culture of a constitutional democratic regime citizens conceive of themselves as free in these respects.
Third, since the question of which conception of political justice is
most appropriate for realizing in basic institutions the values of liberty
and equality has long been deeply controversial within the very democratic tradition in which citizens are regarded as free and equal persons,
the aim of justice as fairness is to try to resolve this question by starting
from the basic intuitive idea of society as a fair system of social cooperation
in which the fair terms of cooperation are agreed upon by citizens themselves so conceived. In Section IV, we saw why this approach leads to
the idea of the original position as a device of representation.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
245 Justice as Fairness
VI
I now take up a point essential to thinking of justice as fairness as a
liberal view. Although this conception is a moral conception, it is not, as
I have said, intended as a comprehensive moral doctrine. The conception
of the citizen as a free and equal person is not a moral ideal to govern
all of life, but is rather an ideal belonging to a conception of political
justice which is to apply to the basic structure. I emphasize this point
because to think otherwise would be incompatible with liberalism as a
political doctrine. Recall that as such a doctrine, liberalism assumes that
in a constitutional democratic state under modern conditions there are
bound to exist conflicting and incommensurable conceptions of the good.
This feature characterizes modern culture since the Reformation. Any
viable political conception of justice that is not to rely on the autocratic
use of state power must recognize this fundamental social fact. This does
not mean, of course, that such a conception cannot impose constraints
on individuals and associations, but that when it does so, these constraints
are accounted for, directly or indirectly, by the requirements of political
justice for the basic structure.27
Given this fact, we adopt a conception of the person framed as part of,
and restricted to, an explicitly political conception of justice. In this sense,
the conception of the person is a political one. As I stressed in the previous
section, persons can accept this conception of themselves as citizens and
use it when discussing questions of political justice without being committed in other parts of their life to comprehensive moral ideals often
associated with liberalism, for example, the ideals of autonomy and individuality. The absence of commitment to these ideals, and indeed to
any particular comprehensive ideal, is essential to liberalism as a political
doctrine. The reason is that any such ideal, when pursued as a comprehensive ideal, is incompatible with other conceptions of the good, with
forms of personal, moral, and religious life consistent with justice and
which, therefore, have a proper place in a democratic society. As com27. For example, churches are constrained by the principle of equal liberty of conscience
and must conform to the principle of toleration, universities by what may be required to
maintain fair equality of opportunity, and the rights of parents by what is necessary to
maintain their childrens’ physical well-being and to assure the adequate development of
their intellectual and moral powers. Because churches, universities, and parents exercise
their authority within the basic structure, they are to recognize the requirements this
structure imposes to maintain background justice.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
246 Philosophy & Public Affairs
prehensive moral ideals, autonomy and individuality are unsuited for a
political conception of justice. As found in Kant and J. S. Mill, these
comprehensive ideals, despite their very great importance in liberal
thought, are extended too far when presented as the only appropriate
foundation for a constitutional regime.28 So understood, liberalism becomes but another sectarian doctrine.
This conclusion requires comment: it does not mean, of course, that
the liberalisms of Kant and Mill are not appropriate moral conceptions
from which we can be led to affirm democratic institutions. But they are
only two such conceptions among others, and so but two of the philosophical doctrines likely to persist and gain adherents in a reasonably
just democratic regime. In such a regime the comprehensive moral views
which support its basic institutions may include the liberalisms of individuality and autonomy; and possibly these liberalisms are among the
more prominent doctrines in an overlapping consensus, that is, in a consensus in which, as noted earlier, different and even conflicting doctrines
affirm the publicly shared basis of political arrangements. The liberalisms
of Kant and Mill have a certain historical preeminence as among the first
and most important philosophical views to espouse modem constitutional
democracy and to develop its underlying ideas in an influential way; and
it may even turn out that societies in which the ideals of autonomy and
individuality are widely accepted are among the most well-governed and
harmonious.29
By contrast with liberalism as a comprehensive moral doctrine, justice
as fairness tries to present a conception of political justice rooted in the
basic intuitive ideas found in the public culture of a constitutional democracy. We conjecture that these ideas are likely to be affirmed by each
of the opposing comprehensive moral doctrines influential in a reasonably
just democratic society. Thus justice as fairness seeks to identify the
kernel of an overlapping consensus, that is, the shared intuitive ideas
which when worked up into a political conception of justice turn out to
28. For Kant, see The Foundations of the Metaphysics of Morals and The Critique of
Practical Reason. For Mill, see On Liberty, particularly Ch. 3 where the ideal of individuality
is most fully discussed.
29. This point has been made with respect to the liberalisms of Kant and Mill, but for
American culture one should mention the important conceptions of democratic individuality
expressed in the works of Emerson, Thoreau, and Whitman. These are instructively discussed by George Kateb in his “Democratic Individuality and the Claims of Politics,” Political Theory 12 (August I984).
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
247 Justice as Fairness
be sufficient to underwrite a just constitutional regime. This is the most
we can expect, nor do we need more.3? We must note, however, that
when justice as fairness is fully realized in a well-ordered society, the
value of full autonomy is likewise realized. In this way justice as fairness
is indeed similar to the liberalisms of Kant and Mill; but in contrast with
them, the value of full autonomy is here specified by a political conception
of justice, and not by a comprehensive moral doctrine.
It may appear that, so understood, the public acceptance of justice as
fairness is no more than prudential; that is, that those who affirm this
conception do so simply as a modus vivendi which allows the groups in
the overlapping consensus to pursue their own good subject to certain
constraints which each thinks to be for its advantage given existing circumstances. The idea of an overlapping consensus may seem essentially
Hobbesian. But against this, two remarks: first, justice as fairness is a
moral conception: it has conceptions of person and society, and concepts
of right and fairness, as well as principles of justice with their complement
of the virtues through which those principles are embodied in human
character and regulate political and social life. This conception of justice
provides an account of the cooperative virtues suitable for a political
doctrine in view of the conditions and requirements of a constitutional
regime. It is no less a moral conception because it is restricted to the
basic structure of society, since this restriction is what enables it to serve
as a political conception of justice given our present circumstances. Thus,
in an overlapping consensus (as understood here), the conception of
justice as fairness is not regarded merely as a modus vivendi.
Second, in such a consensus each of the comprehensive philosophical,
religious, and moral doctrines accepts justice as fairness in its own way;
that is, each comprehensive doctrine, from within its own point of view,
is led to accept the public reasons of justice specified by justice as fairness.
We might say that they recognize its concepts, principles, and virtues as
theorems, as it were, at which their several views coincide. But this does
not make these points of coincidence any less moral or reduce them to
mere means. For, in general, these concepts, principles, and virtues are
accepted by each as belonging to a more comprehensive philosophical,
religious, or moral doctrine. Some may even affirm justice as fairness as
30. For the idea of the kernel of an overlapping consensus (mentioned above), see Theory,
last par. of Sec. 35, pp. 22of. For the idea of full autonomy, see “Kantian Constructivism,”
pP. 528ff.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
248 Philosophy & Public Affairs
a natural moral conception that can stand on its own feet. They accept
this conception of justice as a reasonable basis for political and social
cooperation, and hold that it is as natural and fundamental as the concepts
and principles of honesty and mutual trust, and the virtues of cooperation
in everyday life. The doctrines in an overlapping consensus differ in how
far they maintain a further foundation is necessary and on what that
further foundation should be. These differences, however, are compatible
with a consensus on justice as fairness as a political conception of justice.
VI
I shall conclude by considering the way in which social unity and stability
may be understood by liberalism as a political doctrine (as opposed to a
comprehensive moral conception).3′
One of the deepest distinctions between political conceptions of justice
is between those that allow for a plurality of opposing and even incommensurable conceptions of the good and those that hold that there is but
one conception of the good which is to be recognized by all persons, so
far as they are fully rational. Conceptions of justice which fall on opposite
sides of this divide are distinct in many fundamental ways. Plato and
Aristotle, and the Christian tradition as represented by Augustine and
Aquinas, fall on the side of the one rational good. Such views tend to be
teleological and to hold that institutions are just to the extent that they
effectively promote this good. Indeed, since classical times the dominant
tradition seems to have been that there is but one rational conception of
the good, and that the aim of moral philosophy, together with theology
and metaphysics, is to determine its nature. Classical utilitarianism belongs to this dominant tradition. By contrast, liberalism as a political
doctrine supposes that there are many conflicting and incommensurable
conceptions of the good, each compatible with the full rationality of human persons, so far as we can ascertain within a workable political conception of justice. As a consequence of this supposition, liberalism assumes that it is a characteristic feature of a free democratic culture that
a plurality of conflicting and incommensurable conceptions of the good
are affirmed by its citizens. Liberalism as a political doctrine holds that
31. This account of social unity is found in “Social Unity and Primary Goods,” referred
to in footnote 27 above. See esp. pp. i6of., 170-73, I83f.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
249 Justice as Fairness
the question the dominant tradition has tried to answer has no practicable
answer; that is, it has no answer suitable for a political conception of
justice for a democratic society. In such a society a teleological political
conception is out of the question: public agreement on the requisite
conception of the good cannot be obtained.
As I have remarked, the historical origin of this liberal supposition is
the Reformation and its consequences. Until the Wars of Religion in the
sixteenth and seventeenth centuries, the fair terms of social cooperation
were narrowly drawn: social cooperation on the basis of mutual respect
was regarded as impossible with persons of a different faith; or (in the
terminology I have used) with persons who affirm a fundamentally different conception of the good. Thus one of the historical roots of liberalism
was the development of various doctrines urging religious toleration. One
theme in justice as fairness is to recognize the social conditions that give
rise to these doctrines as among the so-called subjective circumstances
of justice and then to spell out the implications of the principle of toleration.32 As liberalism is stated by Constant, de Tocqueville, and Mill in
the nineteenth century, it accepts the plurality of incommensurable conceptions of the good as a fact of modern democratic culture, provided, of
course, these conceptions respect the limits specified by the appropriate
principles of justice. One task of liberalism as a political doctrine is to
answer the question: how is social unity to be understood, given that
there can be no public agreement on the one rational good, and a plurality
of opposing and incommensurable conceptions must be taken as given?
And granted that social unity is conceivable in some definite way, under
what conditions is it actually possible?
In justice as fairness, social unity is understood by starting with the
conception of society as a system of cooperation between free and equal
persons. Social unity and the allegiance of citizens to their common
institutions are not founded on their all affirming the same conception
of the good, but on their publicly accepting a political conception of justice
to regulate the basic structure of society. The concept of justice is independent from and prior to the concept of goodness in the sense that
its principles limit the conceptions of the good which are permissible. A
just basic structure and its background institutions establish a framework
32. The distinction between the objective and the subjective circumstances of justice is
made in Theory, pp. 126ff. The importance of the role of the subjective circumstances is
emphasized in “Kantian Constructivism,” pp. 540-42.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
250 Philosophy & Public Affairs
within which permissible conceptions can be advanced. Elsewhere I have
called this relation between a conception of justice and conceptions of
the good the priority of right (since the just falls under the right). I believe
this priority is characteristic of liberalism as a political doctrine and something like it seems essential to any conception of justice reasonable for
a democratic state. Thus to understand how social unity is possible given
the historical conditions of a democratic society, we start with our basic
intuitive idea of social cooperation, an idea present in the public culture
of a democratic society, and proceed from there to a public conception
of justice as the basis of social unity in the way I have sketched.
As for the question of whether this unity is stable, this importantly
depends on the content of the religious, philosophical, and moral doctrines available to constitute an overlapping consensus. For example,
assuming the public political conception to be justice as fairness, imagine
citizens to affirm one of three views: the first view affirms justice as
fairness because its religious beliefs and understanding of faith lead to
a principle of toleration and underwrite the fundamental idea of society
as a scheme of social cooperation between free and equal persons; the
second view affirms it as a consequence of a comprehensive liberal moral
conception such as those of Kant and Mill; while the third affirms justice
as fairness not as a consequence of any wider doctrine but as in itself
sufficient to express values that normally outweigh whatever other values
might oppose them, at least under reasonably favorable conditions. This
overlapping consensus appears far more stable than one founded on views
that express skepticism and indifference to religious, philosophical, and
moral values, or that regard the acceptance of the principles of justice
simply as a prudent modus vivendi given the existing balance of social
forces. Of course, there are many other possibilities.
The strength of a conception like justice as fairness may prove to be
that the more comprehensive doctrines that persist and gain adherents
in a democratic society regulated by its principles are likely to cohere
together into a more or less stable overlapping consensus. But obviously
all this is highly speculative and raises questions which are little understood, since doctrines which persist and gain adherents depend in part
on social conditions, and in particular, on these conditions when regulated
by the public conception of justice. Thus we are forced to consider at
some point the effects of the social conditions required by a conception
of political justice on the acceptance of that conception itself. Other things
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions
251 Justice as Fairness
equal, a conception will be more or less stable depending on how far the
conditions to which it leads support comprehensive religious, philosophical, and moral doctrines which can constitute a stable overlapping consensus. These questions of stability I cannot discuss here.33 It suffices
to remark that in a society marked by deep divisions between opposing
and incommensurable conceptions of the good, justice as fairness enables
us at least to conceive how social unity can be both possible and stable.
33. Part III of Theory has mainly three aims: first, to give an account of goodness as
rationality (Ch. 7) which is to provide the basis for identifying primary goods, those goods
which, given the conception of persons, the parties are to assume are needed by the persons
they represent (pp. 397, 433f.); second, to give an account of the stability of a conception
of justice (Chs. 8-9), and of justice as fairness in particular, and to show that this conception
is more stable than other traditional conceptions with which it is compared, as well as
stable enough; and third, to give an account of the good of a well-ordered society, that is,
of a just society in which justice as fairness is the publicly affirmed and effectively realized
political conception of justice (Chs. 8-9 and culminating in Sec. 86). Among the faults of
Part III, I now think, are these. The account of goodness as rationality often reads as an
account of the complete good for a comprehensive moral conception; all it need do is to
explain the list of primary goods and the basis of the various natural goods recognized by
common sense and in particular, the fundamental significance of self-respect and selfesteem (which, as David Sachs and Laurence Thomas have pointed out to me, are not
properly distinguished), and so of the social bases of self-respect as a primary good. Also,
the account of the stability of justice as fairness was not extended, as it should have been,
to the important case of overlapping consensus, as sketched in the text; instead, this account
was limited to the simplest case where the public conception of justice is affirmed as in
itself sufficient to express values that normally outweigh, given the political context of a
constitutional regime, whatever values might oppose them (see the third view in the overlapping consensus indicated in the text). In view of the discussion in Secs. 32-35 of Ch.
4 of liberty of conscience, the extension to the case of overlapping consensus is essential.
Finally, the relevance of the idea of a well-ordered society as a social union of social unions
to giving an account of the good of ajust society was not explained fully enough. Throughout
Part III too many connections are left for the reader to make, so that one may be left in
doubt as to the point of much of Chs. 8 and 9.
This content downloaded from 68.232.23.94 on Thu, 20 Feb 2014 16:19:56 PM
All use subject to JSTOR Terms and Conditions