38
losophy as reconciliation; for seeing that the conditions of a social world at least allow for that possibility affects our view of the world itself and our attitude toward it. No longer need it seem hopelessly hostile, a world in which the will to dominate and oppressive cruelties, abetted by prejudice and folly, must inevitably prevail. None of this may ease our loss, situated as we may be in a corrupt society. But we may reflect that the world is not in itself inhospitable to political justice and its good. Our social world might have been different and diere is hope for those at another time and place.
I.  FUNDAMENTAL IDEAS
PART II
Principles of Justice
11
sc:
§12. Three Basic Points
12.1. In Part II we discuss the content of the two principles of justice diat apply to the basic structure, as well as various grounds in favor of diem and replies to a number of objections. A more formal and organized argument for these principles is presented in Part III, where we discuss the reasoning that moves the parties in the original position. In that argument die original position serves to keep track of all our assumptions and to bring out their combined force by uniting them into one framework so that we can more easily see their implications.
1 begin with three basic points which review some matters discussed in Part I and introduce others we are about to examine. Recall first that justice as fairness is framed for a democratic society. Its principles are meant to answer the question: once we view a democratic society as a fair system of social cooperation between citizens regarded as free and equal, what principles are most appropriate to it? Alternatively: wliich principles are most appropriate for a democratic society that not only professes but wants to take seriously the idea that citizens are free and equal, and tries to realize that idea in its main institutions? The question of whether a constitutional regime is to be preferred to majoritarian democracy, we postpone until later
Ii;
(Part IV, §44)-
40 II.  PRINCIPLES  Of JUSTlCi:
political and social institutions and how they fit together into one unified system of cooperation (§4). We suppose that citizens are born into society and will normally spend their whole lives within its basic institutions. The nature and role of the basic structure importantly influence social and economic inequalities and enter into deterrmning the appropriate principles of justice.
In particular, let us suppose that the fundamental social and economic inequalities are the differences in citizens' life-prospects (their prospects over a complete life) as these are affected by such things as their social class of origin, their native endowments, their opportunities for education, and their good or ill fortune over the course of life (§16). We ask: by what principles are differences of diat kind—differences in life-prospects—made legitimate and consistent with the idea of free and equal citizenship in society seen as a fair system of cooperation?
12.3. The third point is that justice as fairness is a form of political liberalism: it tries to articulate a family of highly significant (moral) values that characteristically apply to the political and social institutions of die basic structure. It gives an account of these values in the light of certain special features of the political relationship as distinct from other relationships, associations, familial, and personal.
(a) It is a relationship of persons within the basic structure of society, a structure we enter only by birth and exit only by death (or so we may assume for the moment). Political society is closed, as it were; and we do not, and indeed cannot, enter or leave it voluntarily.
(b) Political power is always coercive power applied by the state and its apparatus of enforcement; but in a constitutional regime political power is at the same time the power of free and equal citizens as a coEective body. Thus political power is citizens' power, which they impose on themselves and one another as free and equal.
The idea of political liberalism arises as follows. We start from two facts: first, from the fact of reasonable pluralism, the fact that a diversity of reasonable comprehensive doctrines is a permanent feature of a democratic society; and second, from the fact that in a democratic regime political power is regarded as the power of free and equal citizens as a collective body. These two points give rise to a problem of political legitimacy. For if the fact of reasonable pluralism always characterizes democratic societies and if
12. Three Basic Points A\
T" .. "     = ■=
political power is indeed the power of free and equal citizeiis, in the light of : what reasons and values—of what kind of a conception of justice—can citi-
i zens legitimately exercise that coercive power over one another?
Political liberalism answers that the conception of justice must be a polit-1 ical conception, as defined in §9.1. Such a conception when satisfied allows us to say: political power is legitimate only when it is exercised in accordance with a constitution (written or unwritten) the essentials of which all citizens, as reasonable and rational, can endorse in the light of their common human reason. This is die liberal principle of legitimacy. It is a further desideratum that all legislative questions that concern or border on these essentials, or are highly divisive, should also be settled, so far as possible, by guidelines and values that can be similarly endorsed.
In matters of constitutional essentials, as well as on questions of basic justice, we try to appeal only to principles and values each citizen can endorse. A political conception of justice hopes to formulate these values: its shared principles and values make reason public, while freedom of speech and thought in a constitutional regime make it free. In providing a public basis of justification, a political conception of justice provides the frame-
l        work for the liberal idea of political legitimacy. As noted in §9.4, however, and discussed further in §26, we do not say that a political conception for-
;        mulates political values that can settle all legislative questions. This is nei-1   ther possible nor desirable. There are many questions legislatures must
'■ 1     consider that can only be settled by voting that is properly influenced by nonpolitical values. Yet at least on constitutional essentials and matters of
1        basic justice we do try for an agreed basis; so long as there is at least rough agreement here, fair social cooperation among citizens can, we hope, be
1 maintained.1
i 12.4. Given these three points, our question is: viewing society as a fair
system of cooperation between citizens regarded as free and equal, what principles of justice are most appropriate to specify basic rights and liber-i        ties, and to regulate social and economic inequalities in citizens' prospects t , 1     over a complete life? These inequalities are our primary concern.
To find a principle to regulate these inequalities, we look to our firmest considered convictions about equal basic rights and liberties, the fair value
1. It is not always clear whether a question involves a constitutional essential, as will be mentioned in due course. If there is doubt about this and the question is highly divisive, t,        then citizens have a duty of civility to try to articulate their claims 011 one another by refer-'.'  ;     ence to political values, if that is possible.
42 II. PRINCIPLES OF  JUSTICE J'l
"
of the political liberties as well as fair equality of opportunity. We look outside the sphere of distributive justice more narrowly construed to see '' whether an appropriate distributive principle is singled out by those firmest :; <||| convictions once their essential elements are represented in the original position as a device of representation (§6). This device is to assist us in working out which principle, or principles, the representatives of free and equal citizens would select to regulate social and economic inequalities in these prospects over a complete life when they assume that the equal basic liberties arid fair opportunities are already secured.
The idea here is to use our firmest considered convictions about the nature of a democratic society as a fair system of cooperation between free and equal citizens—as modeled in the original position—to see whether the combined assertion of those convictions so expressed will help us to identify an appropriate distributive principle for die basic structure with its economic and social inequalities in citizens' life-prospects. Our convictions about principles regulating those inequalities are much less firm and assured; so we look to our firmest convictions for guidance where assurance is lacking'and guidance is needed (Theory, §§4,20).
§13. Two Principles of Justice
13.1. To try to answer our question, let us turn to a revised statement of the two principles of justice discussed in Theory, §§11-14. They should now read:2
(a) Each person has the same indefeasible claim to a fully adequate scheme of equal basic liberties, which scheme is compatible with the same scheme of liberties for all; and
(b) Social and economic inequalities are to satisfy two conditions: first, they are to be attached to offices and positions open to all under conditions of fair equality of opportunity; and second, they are to
2. This section summarizes some points from "The Basic Liberties and Their Priority" Tanner Lectures on Human Values, vol. 3, ed. Sterling McMurrin (Salt Lake City: University of Utah Press, 198a), §1, reprinted in Political Liberalism. In that essay I try to reply to what I believe are two of the more serious objections to my account ofliberty in Theory raised by H. L. A. Hart in his splendid critical review essay, "Rawls on Liberty and Its Priority," University of Chicago Law Review 40 (Spring 1973); 551-555, reprinted in his Essays in Juris- , prudence and Philosophy (Oxford: Oxford University Press, 1983). No changes made injustice as fairness in this restatement are more significant than those forced by Hart's review.
fflljl^   13. Two Principles ofJustice : 43
11 'I' '
'in"'!
be to the greatest benefit of the least-advantaged members of society (the difference principle).3
Hi:
As I explain below, the first principle is prior to the second; also, in the second principle fair equality of opportunity is prior to die difference prin-I      ciple. This priority means that in applying a principle (or checking it ^      against test cases) we assume that the prior principles are fully satisfied. We '11      seek a principle of distribution (in the narrower sense) that holds within the 1 ! 1    setting of background institutions that secure the basic equal liberties (in-'1       eluding the fair value of the political liberties)4 as well as fair equality of op-;    portunify. How far that principle holds outside that setting is a separate question we shall not consider.5
,' 13.2. The revisions in the second principle are merely stylistic. But be-
fore noting the revisions in the first principle, which are significant, we should attend to the meaning of fair equality of opportunity. This is a difficult and not altogether clear idea; its role is perhaps best gathered from , 1      why it is introduced: namely, to correct the defects of formal equality of opportunity—careers open to talents—in the system of natural liberty, so-'; ,i   called (Theory, §18: 62ff.; §14). To this end, fair equality of opportunity is said to require not merely that public offices and social positions be open ,1       in the formal sense, but that all should have a fair chance to attain them. To
3. Instead of "the difference principle," many writers prefer the term, "the maxhnin principle," or simply "maximin justice," or some such locution. See, for example, Joshua Cohen's very full and accurate account of the difference principle in "Democratic Equality,"
', St!"a 99 (July 1989): 727-751- But I still use the term "difference principle" to emphasize first, that this principle and die maximin rale for decision under uncertainty (§28.1) are two very distinct things; and second, that in arguing for the difference principle over other dis-
\ tributive principles (say a restricted principle of (average) utility, which includes a social minimum), there is no appeal at all to the maxirniii rale for decision under uncertainty. The widespread idea that the argument for the difference principle depends on extreme aversion !u t0 uncertainty is a mistake, although a mistake unhappily encouraged by the faults of exposition in Theory, faults to be corrected in Part III of this restatement.
4. See Theory, §36:197-199.
5. Some have found this land of restriction objectionable; diey think a political conception should be framed to cover all logically possible cases, or all conceivable cases, and not
1 restricted to cases that can arise only within a specified institutional context. See for exam-
''     pie Brian Barry, The Liberal, Tlieory of Justice (Oxford: Oxford University Press, 1973), p. 113. In contrast, we seek a principle to govern social and economic inequalities in demo-j 1      cratic regimes as we know them, and so we are concerned with inequalities in citizens' life-111      prospects that may actually arise, given our understanding of how certain institutions work.
44
II.  PRINCIPLES  OF JUSTICE
Two Principles of Justice
45
specify the idea of a fair chance we say: supposing that there is a distribution of native endowments, those who have the same level of talent and ability and the same willingness to use these gifts should have the same prospects of success regardless of their social class of origin, the class into which diey are born and develop until die age of reason. In all parts of society diere are to be roughly the same prospects of culture and achievement for those similarly motivated and endowed.
Fair equality of opportunity here means liberal equality. To accomplish its aims, certain requirements must be imposed on the basic structure beyond those of the system of natural liberty. A free market system must be set. within a framework of political and legal institutions that adjust the long^: run trend of economic forces so as to prevent excessive concentrations of property and wealth, especially those likely to lead to political domination. Society must also establish, among other tilings, equal opportunities of education for all regardless of family income (§15).6
13.3. Consider now the reasons for revising the first principle.7 One is that the equal basic liberties in this principle are specified by a list as fol- , lo« s: freedom of thought and liberty of conscience; political liberties (for example, the right to vote and to participate in politics) and freedom of association, as well as the rights and liberties specified by the liberty and integrity (physical and psychological) of the person; and finally, the rights and liberties covered by the rule of law. That the basic liberties are specified by a list is quite clear from Theory, §11: 61 (1st ed.); but the use of the singular term "basic liberty" in the statement of the principle on Theory, §11: 60 (1st ed.), obscures this important feature of these liberties.
This revision brings out that no priority is assigned to liberty as such, as if the exercise of something called "liberty" had a preeminent value and were the main, if not the sole, end of political and social justice. While there is a general presumption against imposing legal and other restrictions 011 conduct without a sufficient reason, this presumption creates no special priority lor any particular liberty. Throughout the history of democratic
6. These remarks are the merest sketch of a difficult idea. We come back to it from time to time.
7. This principle may be preceded by a lexically prior principle requiring that basic needs be met, as least insofar as their being met is a necessary condition for citizens to understand and to be able fruitfully to exercise the basic rights and liberties. For a statement of such a principle with farther discussion, see R. G. Peffer, Marxism, Morality, and Social Justice (Princeton: Princeton University Press, 1990), p. 14.
thought the focus has been on achieving certain specific rights and liberties as well as specific constitutional guarantees, as found, for example, in various bills of rights and declarations of die rights of man. Justice as fairness ' follows this traditional view.
13.4. A list of basic liberties can be drawn up in two ways. One is historical: we survey various democratic regimes and assemble a list of rights and liberties that seem basic and are securely protected in what seem to be historically the more successful regimes. Of course, the veil of ignorance means that this kind of particular information is not available to the parties in the original position, but it is available to you and me in setting up justice as fairness.8 We are perfecdy free to use it to specify the principles of justice we make available to the parties.
A second way of drawing up a list of basic rights and liberties is analytical: we consider what liberties provide the political and social conditions essential for the adequate development and full exercise of the two moral powers of free and equal persons (§7.1). Following this we say: first, that the equal political liberties and freedom of thought enable citizens to develop and to exercise these powers in judging the justice of the basic structure of society and its social policies; and second, that liberty of conscience and freedom of association enable citizens to develop and exercise their moral powers in forming and revising and in rationally pursuing (individually or, more often, in association with others) their conceptions of the good.
Those basic rights and liberties protect and secure the scope required for the exercise of the two moral powers in die two fundamental cases just mentioned: that is to say, the first fundamental case is die exercise of those powers in judging the justice of basic institutions and social policies; while the second fundamental case is die exercise of those powers in pursuing our conception of the good. To exercise our powers in these ways is essential to us as free and equal citizens.
8. Here I should mention that there are three points of view injustice as fairness that it is essential to distinguish: the point of view of the parties in the original position, the point of view of citizens in a well-ordered society, and the point of view of you and me who are setting up justice as fairness as a political conception and trying to use it to organize into one coherent view our considered judgments at all levels of generality. Keep in rnind dial die parties are, as it were, artificial persons who are part of a procedure of construction that we frame for our philosophical purposes. We may know many things that we keep from them. For these three points of view, see Political Liberalism, p. 28.
46
II.  PRINCIPLES  OF JUSTICE
13.5. Observe that the first principle of justice applies not only to the basic structure (both principles do this) but more specifically to what we think of as the constitution, whether written or unwritten. Observe also that some of these liberties, especially the equal political liberties and freedom of thought and association, are to be guaranteed by a constitution (Theory, chap. IV). What we may call "constituent power," as opposed to "ordinary power,"9 is to be suitably institutionalized in the form of a regime: in the right to vote and to hold office, and in so-called bills of rights, as well as in the procedures for amending the constitution, for example.
These matters belong to the so-called constitutional essentials, these essentials being those crucial matters about which, given the fact of pluralism, working political agreement is most urgent (§9.4). In view of the fundamental nature of the basic rights and liberties, explained in part by the fundamental interests they protect, and given that the power of the people to constitute the form of government is a superior power (distinct from the ordinary power exercised routinely by officers of a regime), the first principle is assigned priority.
This priority means (as we have said) that the second principle (which includes the difference principle as one part) is always to be applied within a setting of background institutions that satisfy the requirements of the first principle (including the requirement of securing the fair value of die political liberties), as by definition they will in a well-ordered society.10 The fair value of the political liberties ensures that citizens similarly gifted and motivated have roughly an equal chance of influencing die government's policy and of attaining positions of authority irrespective of their economic and social class.11 To explain die priority of the first principle over the second:
9. This distinction is derived from Locke, who speaks of the people's power to constitute the legislative as the first and fundamental law of all commonwealths. John Locke, Second Treatise of Government, §§134,141,149.
10. It is sometimes objected to the difference principle as a principle of distributive justice that it contains no restrictions on the overall nature of permissible distributions. It is concerned, the objection runs, solely with the least advantaged. But this objection is incorrect: it overlooks the fact that the parts of the two principles of justice are designed to work in tandem and apply as a unit. The requirements of the prior principles have important distributive effects. Consider the effects of fair equality of opportunity as applied to education, say, or die distributive effects of the fair value of the political liberties. We cannot possibly take the difference principle seriously so long as we think of it by itself, apart from its setting within prior principles.
11. [See Political Liberalism, p. 358.]
13. Two Principles of Justice
47
this priority rules out exchanges ("trade-offs," as economists say) between the basic rights and liberties covered by the first principle and the social and economic advantages regulated by the difference principle. For example, the equal political liberties cannot be denied to certain groups on the grounds that their having these liberties may enable them to block policies needed for economic growth and efficiency.
Nor can we justify a selective service act that grants educational deferments or exemptions to some on the grounds that doing this is a socially efficient way both to maintain the armed forces and to provide incentives to those otherwise subject to conscription to acquire valuable skills by continuing their education. Since conscription is a drastic interference with the basic liberties of equal citizenship, it cannot be justified by any needs less compelling than those of the defense of these equal liberties themselves (Theory, §58: 333!.).
A further point about priority: in asserting the priority of the basic rights and liberties, we suppose reasonably favorable conditions to obtain. That is, we suppose historical, economic and social conditions to be such that, provided the political will exists, effective political institutions can be established to give adequate scope for the exercise of those freedoms. These conditions mean that the barriers to constitutional government (if such there are) spring largely from the political culture and existing effective interests, and not from, for instance, a lack of economic means, or education, or the many skills needed to run a democratic regime.12
13.6. It is important to note a distinction between die first and second principles of justice. The first principle, as explained by its interpretation, covers the constitutional essentials. The second principle requires fair equality of opportunity and that social and economic inequalities be governed by the difference principle, which we discuss in §§17-19. While some principle of opportunity is a constitutional essential—for example, a principle requiring an open society, one with careers open to talents (to use the eighteenth-century phrase)—fair equality of opportunity requires more than that, and is not counted a constitutional essential. Similarly, although a
13. The priority (or die primacy) of the basic equal liberties does not, contrary to much opinion, presuppose a high level of wealth and income. See Amartya Sen and Jean Dreze, Hunger and Public Action (Oxford: Oxford University Press, 1989), chap. 13; and Partha Dasgupta, An Inquiry into Weil-Being and Destitution (Oxford: Oxford University Press, 1993), chaps. 1-2, 5 and passim.
■
48 II.  PRINCIPLES  OF JUSTICE
social minimum providing for the basic needs of all citizens is also a constitutional essential (§38.3-4; §49.5), the difference principle is more demanding and is not so regarded.
The basis for the distinction between the two principles is not that the first expresses political values while die second does not. Both principles express political values. Rather, we see die basic structure of society as having two coordinate roles, the first principle applying to one, the second principle to the other (Theory, §11: 53). In one role the basic structure specifies and secures citizens' equal basic liberties (mcluding the fair value of the political liberties (§45)) and establishes a just constitutional regime. In the other role it provides the background insdtutions of social and economic justice in the form most appropriate to citizens seen as free and equal. The questions involved in the first role concern the acquisition and the exercise of political power. To fulfill'the liberal principle of legitimacy (§12.3), we hope to settle at least these questions by appeal to the political values that constitute the basis of free public reason (§36).
The principles of justice are adopted and applied in a four-stage sequence.13 In the first stage, the parries adopt the principles of justice behind a veil of ignorance. Limitations on knowledge available to die parties are progressively relaxed in the next three stages: the stage of die constitutional convention, the legislative stage in which laws are enacted as the constitution allows and as the principles of justice require and permit, and the final stage in which the rules are applied by administrators and followed by citizens generally and die constitution and laws are interpreted by members of the judiciary. At this last stage, everyone has complete access to all the facts. The first principle applies at the stage of the constitutional convention, and whether the constitutional essentials are assured is more or less visible on the face of the constitution and in its political arrangements and the way these work in practice. By contrast the second principle applies at the legislative stage and it bears on all kinds of social and economic legislation, and on the many kinds of issues arising at this point (Theory, §31: 172-176). Whether the aims of the second principle are realized is far more difficult to ascertain. To some degree these matters are always open to reasonable differences of opinion; they depend on inference and judgment in assessing complex social and economic information. Also, we can expect more agreement on constitutional essentials than on issues of distributive justice in the narrower sense.
13. [See Theory. §31:172-176, and Political Liberalism, pp. 397-398.]
|j)   '    23. lino Principles of Justice 49
I|' In
....
Thus the grounds for distinguishing the constitutional essentials covered '   ',  by die first principle and the institutions of distributive justice covered by the second are not diat the first principle expresses political values and the second does not. Rather, the grounds of the distinction are four:
(a) The two principles apply to different stages in the application of ! principles and identify two distinct roles of die basic structure;
(b) It is more urgent to setde the constitutional essentials;
(c) It is far easier to tell whether those essentials are realized; and
(d) It seems possible to gain agreement on what those essentials should 1 be, not in every detail, of course, but in the main outlines.
i      13.7. One way to see the point of the idea of constitutional essentials is to connect it with the idea of loyal opposition, itself an essential idea of a con-
:: stitutional regime. The government and its loyal opposition agree on these h constitutional essentials. Their so agreeing makes the government legitimate in intention and the opposition loyal in its opposition. Where the loyalty of both is firm and their agreement mutually recognized, a constitutional regime is secure. Differences about the most appropriate principles of distributive justice in the narrower sense, and the ideals that underlie them, can be adjudicated, though not always properly, within the existing political framework
While the difference principle does not fall under the constitutional es-
•  sentials, it is nevertheless important to try to identify the idea of equality most appropriate to citizens viewed as free and equal, and as normally and
1 fully cooperating members of society over a complete life. I believe this idea
i   involves reciprocity14 at the deepest level and thus democratic equality ■:■ ■ properly understood requires something like die difference principle. (I say "something like," for there may be various nearby possibilities.) The re-
14. [As understood injustice as fairness, reciprocity is a relation between citizens expressed by principles of justice that regulate a social world in which all who are engaged in cooperation and do their part as die rules and procedures require are to benefit in an appro-;   priate way as assessed by a suitable benchmark of comparison. The two principles of justice, !;| ■'   including the difference principle with its implicit reference to equal division as a bench-!i  mark, formulate an idea of reciprocity between citizens. For a fuller discussion of the idea of reciprocity, see Political Liberalism, pp. 16-17, 311(1 the introduction to the paperback edition, pp. xliv, xlvi, li. The idea of reciprocity also plays an important part hi "The Idea of : Public Reason Revisited " University of Chicago Law Review, 64 (Summer 1997): 765-807, ;    reprinted in The Law of Peoples (Cambridge, Mass.: Harvard University Press, 1999) and ii.   Collected Papers.]
50 II.  PRINCIPLES  OV J II ST H i
maining sections of this part (§§14-22) try to clarify the content of this prin- i ciple and to clear up a number of difficulties. , 1 %
§14. The Problem of Distributive Justice
14.1. - The problem of distributive justice injustice as fairness is always this: how are the institutions of the basic structure to be regulated as one unified scheme of institutions so that a fair, efficient, and productive system of social cooperation can be maintained over time, from one generation to the next? Contrast this with the very different problem of how a given bundle of commodities is to be distributed, or allocated, among various individuals whose particular needs, desires, and preferences are known to us, and who have not cooperated in any way to produce those commodities. This second problem is that of aflocative justice (Theory, §11: 56; §14: 77).
To illustrate: accepting the assumptions implied by interpersonal cardinal comparisons of well-being, we might, for example, allocate die bundle of commodities so as to achieve the greatest satisfaction summed over these individuals from die present into the future. As a political conception of justice, the classical principle of utility (as found in Bentham and Sidgvvick) can be seen as adapting die idea of allocative justice so as to be a single 1., principle for the basic structure over time.
14.2. We reject the idea of allocative justice as incompatible with the fundamental idea by which justice as fairness is organized: the idea of society as a fair system of social cooperation over time. Citizens are seen as cooperating to produce the social resources on which their claims are made. In a well-ordered society, in which both the equal basic liberties (with their fair value) and fair equality of opportunity are secured, the distribution of income and wealth illustrates what we may call pure background procedural justice. The basic structure is arranged so that when everyone fallows the publicly recognized rules of cooperation, and honors the claims the rules specify, the particular distributions of goods that result are acceptable as just (or at least as not unjust) whatever these distributions turn out to be.
To elaborate: within the framework of background justice set up by die basic structure, individuals and associations may do as they wish insofar as the rules of institutions permit. Observe that particular distributions cannot be judged at all apart from the claims (entidements) of individuals earned by their efforts within die fair system of cooperation from which those distributions result. In contrast to utilitarianism, the concept of allocative jiii-
1 |i, ,m
14. Distributive Justice 51
"1 lice has no application. There is no criterion for a just distribution apart
I " j from background institutions and the entidements that arise from actually
j working through the procedure.15 It is background institutions that provide
1 the setting for fair cooperation widiin which entidements arise.
! 14.3. These points can be made clearer as follows. The word "back-
i ground" in the phrase "background procedural justice" above is intended
to indicate that certain rules must be included in the basic structure as a j n system of social cooperation so that this system remains fair over time, from f one generation to the next.10
Consider an example. The draft rule in a professional sport such as basketball ranks teams in the opposite order from their standing in the league at the end of the season: championship teams go last in the draft of new players. This rule provides for regular and periodic changes in the roster of teams and is designed to ensure that teams in the league are more or less evenly matched from year to year, so that in any given season each team can j   give any other a decent game. These changes of players are necessary to 1    achieve the aims and attractions of die sport and are not foreign to its pur-!      i pose.
The required background rules are specified by what is necessary' to : fulfill die two principles of justice. Later on we survey some of these as found in a property-owning democracy (Part IV).1' For example, back-;    ground institutions must work to keep property and wealth evenly enough j shared over time to preserve the fair value of the political liberties and fair
I equality of opportunity over generations. They do this by laws regulating
bequest and inheritance of property, and other devices such as taxes, to prevent excessive concentrations of private power (Theory, §43: 245f£).
14.4. Since the difference principle applies to institutions as public systems of rules, their requirements are foreseeable. They do not involve any more continuous or regular interference with individuals' plans and actions than do, say, familiar forms of taxation. Since the effects of those rules are foreseen, they are taken into account when citizens draw up their plans in
lg. See Theory, §14: 74-77, and note the distinction made there between the three kinds of procedural justice.
16. The term "background" is introduced here and is not used in Theory. - 17. Property-owning democracy is discussed in Tliemy, chap. V, hut unfortunately the
contrast between it and welfare-state capitalism is not made clear enough. This defect I aim to correct in Part IV.
52 II.  PRINCIPLES  OF JUSTICE
the first place. Citizens understand that when they take part in social cooperation^ their property and wealth, and their share of what they help to produce, are subject to the taxes, say, which background institutions are known to impose. Moreover, the difference principle (as well as the first principle and the first part of the second principle) respects legitimate expectations based on the publicly recognized rules and the entidements earned by individuals (Theory, §§47-48).ls ■
The rules of background institutions required by the two principles of justice (including the difference principle) are designed to achieve the aims and purposes of fair social cooperation over time. They are essential to preserve background justice, such as the fair value of the political liberties and fair equality of opportunity, as well as to make it likely that economic and social inequalities contribute in an effective way to the general good or, more exacdy, to the benefit of the least-advantaged members of society. Like the draft ride in professional sports, the arrangements required by the difference principle are part of, and not foreign to, the conception of fair social cooperation in justice as fairness. Even with these rules of background justice, distributive justice may still be understood as a case of pure procedural justice.
§15. "^Im: Basic Structure as Subject: First Kind of Reason
i.j. 1. jffefeharacteristic feature of justice as fairness as a political conception is that it-takes the basic structure as its primary subject. I note two broad kinds of reasons for this: the first notes haw^social institutions work and the nature of the principles required t6 regulate them over time to maintain background justice. ..............
Consider an important criticism of Locke. Suppose we begin, as it seems he does, with the attractive idea diat persons' social circumstances and dieir relations with one another should develop over time in accordance with fair agreements fairly arrived at. Much as with Locke's conception of ideal history, we might use certain principles to specify various rights and duties of persons, as well as their rights to acquire and transfer property. Now suppose we start with a just initial state in which everyone's possessions are justly held. We then say that when everyone respects persons' rights and
18. The remarks in this paragraph reply to the kind of objection Nozick raises to the difference principle in Anarchy, State, and Utopia. His description of the Wilt Ghamberlin example, dhap. 7, pp. 160-164, suggests that to apply that principle to government must involve continual interference widl particular individual transactions.
ijl'.' J, t 15. Basic Structure as Subject: First 53
'i '    duties, as well as the principles for acquiring and transferring property, the
( 1 1 succeeding states are also just, no matter how distant in time. Call diis an n 1    ideal historical process view.19
To work out this idea we need an account not only of the just initial state and of fair agreements, but also of just social conditions under which fair ,    agreements are to be reached. Even though the initial state may have been just, and subsequent social conditions may also have been just for some 1     j    time, the accumulated results of many separate and seemingly fair agreements entered into by individuals and associations are likely over an ex-5    tended period to undermine the background conditions required for free v   and fair agreements. Very considerable wealth and property may accumu-I!1  late in a few hands, and these concentrations are likely to undermine fair equality of opportunity, the fair value of the political liberties, and so on. 1     1 1  The kind of limits and provisos that in Locke's view apply direcdy to the separate transactions of individuals and associations in the state of nature are not stringent enough to ensure that fair background conditions are maintained.20
15.3. To preserve these conditions is the task of the rules of pure procedural background justice. Unless the basicjtructure is regulated over time, earlierjust distributions of assets of all kinds do not ensure die justice of ; : later distributions, however free and fair particular transactions between individuals and associations may look when viewed locally and apart from • <■ background institutions. For the outcome of diese transactions taken together is affected by all kinds of contingencies and unforeseeable consequences. It is necessary to regulate, by laws governing inheritance and be-f quest, how people acquire property so as to make its distribution more
equal, to provide fair equality of opportunity in education, and much else. That such rules of background justice are in force over time does not de-i '.v. tract from but rather makes possible the important values expressed by free
:•■1 : and fair agreements reached by individuals and associations within the ba-\      :    sic structure. This is because principles applying to these agreements direcdy (for example, the law of contract) do not alone suffice to preserve I background justice.
What is needed, then, is a division of labor between two kinds of princi-
l lrj. Nozick's Anarchy, State, and Utopia is an example of this kind of view.
20. For example, in Locke's case they fail to guarantee the equal political liberties, as we
Ican tell from Second Treatise, §158. See Joshua Cohen, "Structure, Choice, and Legitimacy: 1   Locke's Theory of the State," Philosophy and Public Affairs 15 (Fall ig86): 301-324.
54
II.  PRINCIPLES OF JUSTICE
ples, each kind suitably specified: first, those that regulate the basic structure over time and are designed to preserve background justice from one generation to die next; and second, those diat apply direcdy to the separate and free transactions between individuals and associations. Defects in either kind of principle can result in a serious failure of the conception of justice as a whole.
15.3. Once the many transactions and agreements of individuals and associations are framed within a just basic structure, we have an ideal social process view, of which justice as fairness is an example. The contrast with Locke's ideal historical process view is in part this: while both views use the concept of pure_proLcedural justice, they specify this concept in different ways. The historical process view'focuses on the transactions of individuals and associations as these are constrained by the principles and provisos applying direcdy to the patties iiiparticular transactions.
By contrast, as a social process view^' justice as fairness focuses first on the basic structure and ôii thé~ŕégulations required to maintain background justice over time for all persons equally, whatever their generation or social position. Since a public conception of justice needs clear, simple, and intelligible rules, we rely onan mstitutiqnaldivision of labor betweenprinciples required to preserve background justice and principles that apply direcdy to particular transactions between individuals and associations. Once diis division of labor is set up, individuals and associations are then left free to advance their (permissible) ends within the framework of die basic structure, secure in the knowledge that elsewhere in the social system the regulations necessary to preserve background justice are in force.
Taking'the basic structure as the primary,subject enables us to regard distributive justice as a case of pure background procedural justice: when everyone follows the publicly recognized-rules of-coóperation, the particular distribution that results is acceptable as just whatever that distribution turns out to be (§14.2). This allows us to abstract from the enormous complexities of the innumerable transactions of daily life and frees us from having to keep track of the changing relative positions of particular individuals (Theory, §14: 76Ĺ). Society is an ongoing scheme of fair cooperation over time widiout any specified beginning or end relevant for political justice. The principles of justice specify the form of background justice apart from all particular historical conditions. What counts is the workings of social institutions now, and a benchmark of the state of nature—the level of weE-be-
iffi Basic Structure as Subject: Second
55
hig (however specified) of individuals in that state—plays no role. It is a historical surd, unknowable, but even if it could be known, of no significance.21
§ih'. The Basic Structure as Subject: Second Kind of Reason
16.1. The second kind of reason for taking jhe basic structure as the primary subject derives from its profound and pervasive influence on the persons who live under its institutionsrReeall that'ih explaining the emphasis put on die basic structure as subject, we said that we view citizens as born into society: it is there diat they will lead a complete life. They enter that social world only by birth, leave it only by death. And since any modern society, even a well-ordered one, must rely on some inequalities to be well designed and effectively organized, we ask what kinds of inequalities a well-ordered society would allow or be particularly concerned to avoid.
Justice as fairness focuses on inequalities in citizens' life-prospects—-their prospects over a complete life (as specified by an appropriate "index of primary goods)—as these prospects are affected by three kinds of contingen-
(a) dieir social class of origin: the class into which they are bora and develop before the age of reason;
(b) their native endowments (as opposed to their realized endowments); and their opportunities to develop these endowments as affected by their social class of origin:
(c) their good or ill fortune, or good or bad luck, over the course of life (how they are affected by illness and accident; and, say, by periods of involuntary unemployment and regional economic decline).
Even in a well-ordered society, then, our prospects over life are deeply affected by social, natural, and fortuitous contingencies, and by the way the basic structure, by setting up inequalities, uses those contingencies to meet certain social purposes.
Pointing out diese three kinds of contingencies is not enough, of course, to show conclusively diat die basic structure is the appropriate subject of political justice. No such decisive arguments are available, as everything de-
21. See T7i£ory, §12: 69, though the point is not stated sufficiently sharply. ,
56
II. PRINCIPLES Ol JUSTICE
pends on how the conception ofjustice as fairness hangs together as a whole. Yet if we ignore die inequalities in people's prospects in life arising from these contingencies and let those inequalities work themselves out while failing to institute the regulations necessary to preserve background justice, we would not be taking seriously the idea of society as a fair system of cooperation between citizens as free and equal. This reminds us that what we are asking is precisely: what principles of background justice are presupposed in taking seriously that idea of society (§12.1)?
16.2. If citizens of a well-ordered socieJy-are.,to recognize one another as free and equal, basic institutions must ieducafe them to this conception of themselves, as well as publicly exhibit and encourage this ideal of political justice. This task of educatirarbelongs triwhat-we-may call the wide role of a political conception.22 In this role such a conception is part of the public political culture: its first principles are embodied in the institutions of the basic structure and appealed to in their interpretation. Acquaintance with and participation in that public culture is one way citizens learn to conceive of themselves as free and equal, a conception which, if left to their own reflections, diey would most likely never form, much less accept and desire to realize.
Consider further how the three contingencies (noted above) affect the content of people's final ends and purposes, as well as the vigor and confidence with which they pursue them. We assess our prospects in life according to our place in society and we form our ends and purposes in the light of the means and opportunities we can realistically expect. So whether we are hopeful and optimistic about our future, or resigned and apathetic, depends both on the inequalities associated with our social position and on the public principles of justice that society not merely professes but more or less effectively uses to regulate the institutions of background justice. Hence the basic structure as a social and economic regime is not only an arrangement that satisfies given desires and aspirations but also an arrangement tEktiuniBErfllftt£T^e8iJres and aspirations in thejfutnre. This it docs by the expectations and ambitions it encourages in the present, and indeed over a complete life.
Moreover, native endowments of various kinds (say, native intelhgence
22. By contrast^ the narrow role might be something like that of specifying the basic principles and most essential rules that must be followed for political society to be enduring and stable. H. L. A. Hart's idea of the irnnirnum content of natural law, which draws on Hume, is an example. See The Concept of Law (Oxford: Oxford University Press, 1961), pp. 189-195.
ill 111 I
< I1/ I '   17. Who Are the Least Advantaged?
57
I,
Uli
Vik:
and natural ability) are not fixed natural assets with a constant capacity. They are merely potential and cannot come to fruition apart from social conditions; and when realized they can take but one or a few of many possible forms. Educated and trained abilities are always a selection, and a small selection at diat, from a wide range of possibilities that might have been .. fulfilled. Among what affects their realization are social attitudes of encouragement and support, and institutions concerned with their early discipline and use. Not only our conception of ourselves, and our aims and ambitions, but also our realized abilities and talents, reflect our personal history, opportunities and social position, and the influence of good and ill fortune.
16.3. To sum up: for the two kinds of reasons noted in this and the preceding section, we take the basic structure as the primary subject. This structure comprises social institutions within which human beings may de-i velop their moral powers and become fully cooperating members of a society of free and equal citizens. And as a framework that preserves background justice over time from one generation to the next it realizes the idea (central to justice as fairness) of pure background procedural justice as an ideal social process (as explained under the first kind of reason). It also answers to the public role of educating citizens to a conception of themselves as free and equal; and, when properly regulated, it encourages in diem attitudes of optimism and confidence in their future, and a sense of being treated fairly in view of the public principles which are seen as effectively regulating economic and social inequalities (as explained under the second kind of reason).
So far, then, starting with die basic structure seems to accord with the other ideas of justice as fairness, something we could not have foreseen at the outset. A sharp definition of diat structure might have gotten in the way of fitting it into these other ideas, just as a sharp definition of them would have gotten in the way of fitting them to it. (Recall the remarks in §4.3.)
§17. Who Are the Least Advantaged?
17.1. We have referred to the least advantaged, but who are diey and how are they singled out? To answer these questions we introduce the idea of primary goods. These are various social conditions and all-purpose means that are generally necessary to enable citizens adequately to develop and fully exercise their two moral powers, and to pursue their determinate conceptions of die good. Here we look to the social requirements and die nor-
58
II. PRINCIPLES OF JUSTICIÍ
lilií,
vj. Wh.o Are the Least Advantaged?
59
mal circumstances of human life in a democratic society. Primary goods are things needed and required by persons seen in the light of die political con- ; ception of persons, as citizens who are fully cooperating members of society, and not merely as human beings apart from any normative conception. These goods :are tilings citizens need as free and equal persons living a complete life; they are not things it is simply rational to want or desire, or to prefer or even to crave. We use the political conception, and not a comprehensive moral doctrine, in specifying those needs and requirements.
What count as primary goods depends, of course, on various general facts about human needs and abilities, their normal phases and requirements of nurture, relations of social interdependence, and much else. We need at least a rough idea of rational plans of life showing why these plans usually have a certain structure and depend on certain primary goods for their formation, revision, and successful execution.23 But, as stressed above, the account of primary goods does not rest solely on psychological, social, or historical facts. While the list of primary goods rests in part on the general facts and requirements of social life, it does so only together with a political conception of the person as free and equal, endowed with the moral powers, and capable of being a fully cooperating member of society. This normative conception is necessary to identify the appropriate list of primary goods.24
17.2. We distinguish five kinds of such goods:
(i) The basic rights and liberties: freedom of thought and liberty of conscience, and die rest (§13). These rights and liberties are essential institutional conditions required for the adequate development and full and informed exercise of the two moral powers (in the two fundamental cases (§134)).
(ii) Freedom of movement and free choice of occupation against a background of diverse opportunities, which opportunities allow the pursuit of a variety of ends and give effect to decisions to revise and alter diem.
(hi) Powers and prerogatives of offices and positions of authority and responsibility.
(iv) Income and wealth, understood as all-purpose means (having an ex-
23. A sketch of the features of rational plans is given in TJiwry, chap. VTI.
24. Unhappily Theory is at best ambiguous on this matter. I am indebted to a number of people, especially to Joshua Cohen and Joshua Rabinowitz; and to Allen Buchanan, T. M. Scanlon, and Samuel Schemer for valuable discussion; and to Michael Teitelman, who first raised the difficulty.
change value)25 generally needed to achieve a wide range of ends whatever they may be.
■1    (v) The social bases of self-respect, understood as those aspects of basic ► institutions normally essential if citizens are to have a lively sense of their worth as persons and to be able to advance their ends with self-confidence.
17.3. The two principles of justice assess the basic structure according to how it regulates citizens' shares of primary goods, these shares being specified in terms of an appropriate index. Note that primary goods are given by reference to objective features of citizens' social circumstances, features open to piiblic view: their secured institutional rights and liberties, dieir available fair opportunities, their (reasonable) expectations of income and wealth seen from their social position, and so on. As we have said, the inequalities to which the difference principle applies are differences in citizens' (reasonable) expectations of primary goods over a complete life. These expectations are their life-prospects. In a well-ordered society where all citizens' equal basic rights and liberties and fair opportunities are secure, the least advantaged are diose belonging to the income class with the lowest expectations.2fi To say that inequalities in income and wealth are to be ar-; ranged for the greatest benefit of the least advantaged simply means that we are to compare schemes of cooperation by seeing how well off the least advantaged are under each scheme, and then to select the scheme under
25. For this idea of exchange value, see Rawls, "Fairness to Goodness," Philosophical Review 84 (October 1975): §111. Also in Collected Papers.
26. Note here that in the simplest form of the difference principle the individuals who belong to the least advantaged group are not identifiable apart from, or independendy of, their income and wealth. The least advantaged are never identifiable as men or women, say, or as whites or blacks, or Indians or British. They are not individuals identified by natural or other features (race, gender, nationality, and the like) diat enable us to compare their situation under all the various schemes of social cooperation it is feasible to consider. Taking these cooperative schemes as possible social worlds (let's say) over which the names of individuals refer to (rigidly designate) the same individuals in each possible (social) world, the term "the least advantaged" is not a rigid designator (to use Saul Kripke's term, see Naming and Necessity (Cambridge, Mass.: Harvard University Press, 197a)). Rather, the worst off under any scheme of cooperation are simply the individuals who are worst off under that particular scheme. They may not be those worst off in another. Even supposing, for example, that it turns out, as commonsense political sociology might suggest, that the least advantaged, identified by income and wealth, include many individuals born into the least-favored social class of origin, and many of the least (naturally) endowed and many who experience more bad luck and misfortune (§16), nevertiieless diose attributes do not define the least advantaged. Rather, it happens that there may be a tendency for such features to characterize many who belong to that group.
60 II.  PRINCIPLES  OF JUST! I )
which the least advantaged are better off than they are under any otiier i1[ scheme. 'jl'^'ll
To highlight die objective character of primary goods, note diat it is not selfrrespect as an attitude toward oneself but die social bases of self-respect that count as a primary good.27 These social bases are things like the insti- ; l tutional fact that citizens have equal basic rights, and the public recognition of that fact and diat everyone endorses the difference principle, itself a form of reciprocity. The objective character of primary goods also is shown in the fact diat in applying the principles of justice we do not consider estimates of citizens' overall happiness as given, say, by the fulfillment of their (rational) preferences, or their desires (as in a utilitarian view). Nor do we consider their good in the light of some moral or associational or personal ideal. We also do not need a measure of citizens' moral powers and oil] i abilities, or of how far citizens have actually realized them, so long as their powers and abilities suffice for them to be normal cooperating members of society.28 1!
Citizens' appropriate shares of primary goods are not regarded as approximating to their good as specified by any particular comprehensive re- 1 'i |' ligious, philosophical, or moral doctrine; or even as specified by what several such doctrines may hold in common concerning our (comprehensive) good. The account of primary goods belongs, then, wholly within justice as fairness as a political conception of justice. The reason for remaining widiin die political conception is by now familiar: it is in order to hold open the possibility of finding a public basis of justification supported by an overlapping consensus. .
Primary goods, then, are what free and equal persons (as specified by the political conception) need as citizens. These goods belong to a partial con- j ception of the good that citizens, who affirm a plurality of conflicting comprehensive doctrines, can agree upon for the purpose of making the inter- f personal comparisons required for workable political principles. While pluralism means that there can be no agreement on a complete conception of the good grounded in a comprehensive doctrine, some conceptions of the good are indispensable for any account of justice, political or other; and they can be freely used in justice as fairness so long as they fit within it as political conceptions (§43)- The partial conception of the good set out by the account of primarv goods is one of them.
.......
27. Theory is ambiguous on this point. It fails to distinguish between self-respect as an attitude, the preserving of which is a fundamental interest, and the social bases that help to support that attitude.
28. On this point, see the discussion of the provision of health care in Part IV, §51.
I,    18. The Difference Principle 61
17.4. A final comment: there are at least two ways to proceed in specifying a list of primary goods. One is to look at the various comprehensive doctrines actually found in society and specify an index of such goods as a kind of average of what those who affirm the opposing doctrines would
! need by way of institutional protections and all-purpose means. Doing this might seem the best way to achieve an overlapping consensus.
But this is not how justice as fairness proceeds. Instead, it works up a political conception from the fundamental idea of society as a fair system of social cooperation. The hope is that this conception with its account of primary goods can win the support of an overlapping consensus. We leave
: aside the comprehensive doctrines that now exist, or have existed, or might exist. Our thought is not that primary goods are fair to conceptions of the good associated with comprehensive doctrines by striking a fair balance between them. Rather, primary goods are fair to free and equal citizens: these goods enable them to advance their permissible conceptions of die good (those the pursuit of which are compatible with justice).
§18. The Difference Principle: Its Meaning
' 18.1. We now turn to the difference principle as a principle of distributive justice in the narrow sense. Recall that it is subordinate to both the first principle of justice (guaranteeing the equal basic liberties) and the principle of fair equality of opportunity (§13.1). It works in tandem with these two prior principles and it is always to be applied within background institutions in which diose principles are satisfied.29
Social cooperation, we assume, is always productive, and without cooperation there would be nothing produced and so nothing to distribute. This assumption is not emphasized sufficiendy in Theory, §§12-13. Figure 1 implies that there is production: MAG and LAG (xa and xa in the figure in Theory) are now representative individuals of the more and the less advantaged groups respectively, the two groups engaged in productive cooperation.30 The OP curve (P for production) runs northeast from the origin until it bends downward to the southeast.31
2g. The explanation of the principle of (Pareto) efficiency for institutions is found in Theory, §12: 58-62.
30. This figure is similar to figure 6 in Theory, §13: 66.
31. For example, the figures in'Theory, §12: 5gf. assume that there is an already given bundle of goods to be shared between the two persons X! and xa. This is shown in the fact that the efficiency frontier runs northwest and southeast. And no mention is made of these
:   persons as being engaged in cooperation to produce diose goods.
62
II.  PRINCIPLES  OF JUSTICE
x — MAG
In this figure the distances along the two axes are measured in terms of an index of primary goods, with the x-axis the more advantaged group (MAG), the y-axis the less advantaged (LAG). The line JJ parallel to the x-axis is the highest equal-justice line touchedby the OP curve at its maximum at D. Note that D is the efficient point nearest to equality, represented by the 45-dcgree line. N is the Nash point, where the product of utilities is maximized (if we assume utilities to be linear in indexes of primary goods), and B is the Benthara. point, where the sum of individual utilities is maximized (again with the same assumption). The set of efficient points goes from D to the feudal point F, at which the OP curve becomes vertical.
We imagine the whole space southeast of the 45-degree line to be filled with parallel equal-justice lines. Thus from each point on the 45-degree line mnning northeast from the origin, there is an equal-justice line. The line JJ is simply the highest such line diat can be reached when we are constrained to move along the OP curve. Society aims, other things equal, to reach the highest equal-justice line measured by the distance from O along the 45-degree line. To do this it moves as far northeast as possible along the OP curve and stops when this curve bends to the southeast,
Observe that the parallel lines are equal-justice lines and not indifference fines of the familiar kind that represent evaluations of individual or social welfare. Equal-justice lines represent how claims to goods cooperatively produced are to be shared among diose who produced them, and they reflect an idea of reciprocity. They are equal-justice fines in tlie sense that any point on a line is equally acceptable provided it is reached by an OP curve of a scheme of cooperation satisfying the principles of justice prior to the difference principle. The fact that they are parallel means that a greater index of primary goods (here understood as a person's prospects of income and wealth over a complete life) for one group (the MAG) is justified only insofar as it adds to the index of the other group (the LAG). When this is no longer true, even uSough the index increases for the more advantaged group, as it does beyond D, then
111
18. The Difference Principle 63
the reciprocity implicit in the difference principle no longer obtains. This is shown by the fact that N and B are on lower equal-justice lines than D. A utilitarian equal-justice line through B would be a smooth curve convex to the origin running from northwest to southeast, showing the MAG may gain more even though die LAG receive less. Contrary to reciprocity, trade-offs are permitted.
Finally, note that as indicated in §17, the MAG and the LAG are specified by reference to their shares in the output and not as particular individuals identifiable independently of the scheme of cooperation. When we represent the index of the MAG on the x-axis, the OP curve lies everywhere southeast of the 45-degree line.
18.2. A scheme of cooperation is given in large part by how its pubhc rules organize productive activity, specify the division of labor, assign various roles to those engaged in it, and so on. These schemes include schedules of wages and salaries to be paid out of output. By varying wages and salaries, more may be produced. This is because over time the greater returns to the more advantaged serve, among other tilings, to cover the costs of training and education, to mark positions of responsibility and encourage persons to fill them, and to act as incentives. A given OP curve is paired with a particular scheme of cooperation: it indicates the returns to the two groups when only wages and salaries are changed. The origin of the OP curve represents the equal division point: bodi groups receive die same remuneration.
To explain: take any point on the OP curve: if the wages to the more advantaged is the corresponding point on the x-axis, then the wages to the less advantaged is the corresponding point on the y-axis. Thus there are, in general, different OP curves for different schemes of cooperation; and some schemes are more effectively designed than others. One scheme is more effective than another if its OP curve always gives a greater return to the less advantaged for any given return to the more advantaged.32 Other things being equal, die difference principle directs society to aim at the highest point on die OP curve of the most effectively designed scheme of cooperation.
18.3. A further feature of the difference principle is that it does not require continual economic growdi over generations to maximize upward indefinitely the expectations of the least advantaged (assessed in terms of
33. When these curves criss-cross, the one tangent to the highest JJ line is best; if they touch the same JJ line, the one whose tangent is to the left of the oilier is best.
iífMíj
64
II. PRINCIPLES OF JUSTICE
income and wealth). That would not be a reasonable conception of justice. We should not rule out Mill's idea.of a society in a just stationary state: i where (real) capital accumulation may cease.33 A well-ordered society is specified so as to allow for this possibility. What the difference principle, does require is that during an appropriate interval of time the differences in income and wealth earned in producing the social product be such that if the legitimate expectations of the more advantaged were less;, those of the less advantaged would also be less. Society would always be on the upward-rising part or at the top of the OP curve.34 Permissible inequalities (thus defined) satisfy that condition and are compatible with a social product of a steady-state equilibrium in which a just basic structure is supported and re- ; produced over time.
Another aspect of the same point is this: the difference principle requires that however great the inequalities in wealth and income may be, and however willing people are to work to earn their greater shares of output, existing inequalities must contribute effectively to the benefit of the least advantaged. Otherwise the inequalities are not permissible. The general level of wealth in society, mcluding the well-being of the least advantaged, depends on people's decisions as to how to lead their lives. The priority of liberty means that we cannot, be forced to engage in work that is higldy productive in terms of material goods. What kind of work people do, and how hard they do it, is up to them to decide in light of the various incentives society offers. What the difference principle requires, then, is that however great the general level of wealth—whether high or low—the existing inequalities; are to fulfill the condition of benefiting others as well as ourselves. This condition brings out that even if it uses the idea of maximizing the expectations of the least advantaged, die difference principle is essentially a principle of reciprocity.
18.4. We have seen that the two principles of justice apply to citizens as identified by their indexes of primary goods. It is natural to ask: Why are distinctions of race and gender not explicitiy included among the three contingencies noted earlier (§16)? How can one ignore such historical facts as slavery (in the antebellum South) and the inequalities between men and women resulting from the absence of provisions to make good women's ex-
33. See his Principles of Political Economy, bk. IV, ch. VI.
34. See the distincaon at Theory, §13: 68 between perfectly just schemes and those just diroughout.
18. The Difference Principle
65
tra burden in the bearing, raising, and educating of children so as to secure their fair equality of opportunity?
The answer is that we are mainly concerned with ideal theory: the account of the well-ordered society of justice as fairness. Within that account we need to distinguish two questions: first, what contingencies tend to generate troubling inequalities even in a well-ordered society and thus prompt us, along with other considerations, to take the basic structure as the primary subject of justice; and second, how within ideal theory should the least advantaged be specified?
While there is some tendency for individuals most adversely affected by the three contingencies (§16.1) to be among die least advantaged, this group is denned not by reference to those contingencies but by an index of primary goods (§17, n. 26). Taking the simplest form of the difference principle, the least advantaged are those who share with other citizens die basic equal liberties and fair opportunities but have the least income and wealth. We use income and wealth to specify this group; and the particular individuals who belong to it may change from one arrangement of the basic structure to another.
18.5. In ideal theory, as stated in Theory, §16, the two principles of justice are to be applied to the basic structure by assessing it from certain standard points of view: namely, those of the representative equal citizen (whose basic equal liberties and fair opportunities are secure) and of the representatives of various levels of income and wealth. Nevertheless, sometimes other positions must be taken into account. Suppose, for example, that certain fixed natural characteristics are used as grounds for assigning unequal basic rights, or aUowing some persons only lesser opportunities; then such inequalities will single out relevant positions. Those characteristics cannot be changed, and so the positions they specify are points of view from which the basic structure must be judged.
Distinctions based on gender and race are of this kind. Thus if men, say, have greater basic rights or greater opportunities than women, these inequalities can be justified only if they are to the advantage of women and acceptable from their point of view. Similarly for unequal basic rights and opportunities founded on race (Theory, §16: 85). It appears that historically these inequalities have arisen from inequalities in political power and control of economic resources. They are not now, and it would seem never have been, to the advantage of women or less favored races. To be sure, so sweeping a historical judgment may occasionally be uncertain. However, in
66
II. PRINCIPLES OF JUSTICE
a well-ordered society in the present age no such uncertainty obtains, so justice as fairness supposes that the standard relevant positions specified by die primary goods should suffice.
18.6. To conclude: when used in a certain way, distinctions of gender and race give rise to further relevant positions to which a special form of the difference principle applies {Theory, §16: 85). We hope that in a well-ordered society under favorable conditions, with the equal basic liberties and fair equality of opportunity secured, gender and race would not specify relevant points of view. Theory takes up only two questions of partial compliance (or nonideal) dieory, civil disobedience and conscientious refusal to serve in an unjust war. The serious problems arising from existing discrimination and distinctions based on gender and race are not on its agenda, which is to present certain principles of justice and then to check them against only a few of the classical problems of political justice as these would be setded within ideal theory.
This is indeed an omission in Theory; but an omission is not as such a fault, either in that work's agenda or in its conception of justice. Whether fault there be depends on how well that conception articulates the political values necessary to deal with these questions. Justice as fairness, and other liberal conceptions like it, would certainly be seriously defective should they lack the resources to articulate the political values essential to justify the legal and social institutions needed to secure the equality of women and minorities. In Part TV, §50, there is a brief discussion of die nature of the family and the equality of women.
§19. Objections via Counterexamples
19.1. Part of the idea of reflective equilibrium is to test the soundness of first principles by seeing whether we can endorse on reflection the judgments to which they lead in cases sometimes framed for this purpose: counterexamples so-called. To be a proper counterexample a case must satisfy all the relevant assumptions made in applying or in arguing for the principles of justice; otherwise it misses the mark. Let us look at three objections via counterexamples to illustrate this.
Consider first two related objections: (a) suppose the most effective OP curve rises very slowly to its maximum; then the share of the more advantaged is much greater than the share to the less advantaged. (In Figure 1 (§18.1), imagine D moved far to the right along the hue JJ.) This may seem unjust to the less advantaged. On the other hand: (b) suppose that the most
jg. Objections via Counterexamples
67
effective OP curve falls very slowly after its maximum; then the more advantaged do not receive a much greater share even diough their receiving this share would only slighdy reduce the share of the less advantaged. (In the figure, imagine the arc from D through N and B and beyond stretched far to the right). This may seem unjust to die more advantaged, i In bodi cases, the troubling feature is die rather flat slope of die OP curve, in one case before, in the odier after, the maximum. This means diat large potential gains (or losses) to one group are paired with small potential losses (or gains) for the other group. In such cases we are tempted to think some adjustments should be made to achieve a greater overall gain. The reply is that, given the required background institutions securing both the equal basic liberties and fair equality of opportunity, and the many possibilities of social organization, the most effective OP curve is very unlikely to have the flat slopes described above. Thus:
(i) In reply to (a): if citizens have fair and equal opportunities to develop their native endowments and to acquire socially productive skills, and if the scheme of cooperation is effectively designed, then the OP curve should rise quickly enough to its maximum so that the ratio of shares in favor of the more advantaged is not likely to strike us as unjust. The idea is that given the equal basic liberties and fair equality of opportunity, the open competition between the greater numbers of the well-trained and better ed-
; ucated reduces the ratio of shares until it lies within an acceptable range. Notice here how, in meeting the objection, we rely on the way the difference principle works in tandem with the prior principles. Widi background institutions of fair equality of opportunity and workable competition required by the prior principles of justice, the more advantaged cannot unite as a group and then exploit their market power to force increases in their income.35 This has been mentioned before; here we see it illustrated.
(ii) In reply to (b): given the same assumptions as in (i), there surely ex-
35. For example, background institutions prevent doctors from forming an association to push up the cost of medical care and thus to raise the income of doctors, say by restricting entry into the medical profession, or by agreeing to charge higher fees. High earnings, however, are not sufficient proof of collusion. The income of opera singers seems largely determined by free demand and supply; the demand is high, the supply is low, and in the short run nearly fixed, but not forever fixed as is the supply of the paintings by old masters. The number of opera singers is small enough so that their earnings are not a serious worry in any case; besides they work hard and spread joy. Whereas doctors are a large group, and should competitive background institutions with fair opportunity not work properly in their case, or in similar cases, we would have to examine die causes for the failure of competitive arrangements and try to fix them if that can be done consistent with die prior principles. Wc might also have to reconsider the soundness of the difference principle.