JOSHUACOHEN
zyxwvu
zy
Taking People as They Are?
zyxw
zyxw
My purpose is to consider if, in political society,
there can be any legitimate and sure principle of
government,taking men as they are and laws as
they might be.
-Jean-Jacques Rousseau, The Social Contract
Following Rousseau’s opening thought in The
Social Contract. . . ,I shall assume that his phrase
“men as they are” refers to persons’ moral and
psychological natures and how that nature works
within the framework of political and social
institutions.
-John Rawls, The Law ofpeoples
INTRODUCTION
According to John Rawls’s “difference principle,” inequalities are fully
just if and only if they, roughly, maximize the advantage of the least advantaged group. Although the difference principle imposes no direct
limits on the range of the income distribution (difference between top
and bottom), or on any other measure of income dispersion, it is nevertheless an egalitarian principle in requiring that income and wealth inequalities be justified by reference to their contribution to the common
advantage, and, more particularly, by reference to the benefits they
confer on the least advantaged. In a series of papers on “incentive inequalities,” however, G. A. Cohen has shown that the content of the difference principle is less determinate-perhaps less egalitarian-than it
may seem.’ The resolution of this indeterminateness has, I believe, im-
zyxwvu
zyxwv
I presented earlier versions of this article at Yale University for a conference in celebration of JerryCohen’s sixtieth birthday, and at the Harvard Government Department’sPolitical Theory Colloquium. For comments and for discussion of related issues, I am grateful
to Michael Otsuka, Paula Casal, Roxanne Fay, Joshua Flaherty, Celia Kerstenetzky, Andrew
Williams, Andrea hgiovanni, Thad Williamson, and the editors of this journal. I am also
grateful to Roxanne Fay for research assistance. My debt to Jerry Cohen is large. His work
and life demonstrate the possibility of combining philosophical depth, moral conviction,
and a sweet soul.
I. References to Cohen’s writings are included in the text: IIC: “Incentives,Inequality,
and Community,” in Grethe B. Peterson, ed., Tanner Lectures on Human Values, vol. 13
2002 by Princeton University Press.
Philosophy G Public Affairs 30, no. 4
364
zyxw
zyx
zyx
zyxwvutsr
Philosophy G Public Affairs
portant consequences both for our understanding of what justice requires and for a family of fundamental methodological issues in political philosophy-including the plausibility of political liberalism, the
role of substantive claims about social order and human nature in normative political thought, and the appropriate way to justify principles
of justice.’
Much of Cohen’s writing about incentive inequalities has been presented as a critique of John Rawls’s theory of justice, and for the purposes of this article, I will largely preserve this polemical setting. In section 11, I explain what I mean by “incentive inequalities,” and introduce a
range of cases in which incentive inequalities countenanced by what I
will call an “ultralax”interpretation of the difference principle are objectionable. While a strict version of the difference principle, which condemns all incentive inequalities, suffers from unreasonably rigorous expectations about our willingness to serve the good of others, only an
unacceptably ultralax version would permit incentive inequalities in all
the cases I consider. In section 111, I show that at least in some circumstances a Rawlsian conception of justice will require, contrary to Cohen’s
assertions, changes in the social ethos because that is how institutional
changes lead to a more just distribution. With these observations about
institutions and ethos as background, I then consider, in section rV,how
a Rawlsian can handle the cases of intuitively objectionable incentive
inequalities sketched in section 11. The Rawlsian response underscores
the essential role in Rawls’s formulation of egalitarian liberalism of substantive claims about social order and human motivations: in particular,
claims about the pervasive influence of institutions on politicaleconomic outcomes and on culture and identity. In section V, I suggest
that if we reject this strong “institutional determinism,” we seem to face
a choice between preserving political liberalism’s institutional focus
while backing off from egalitarianism, or preserving egalitarianism and
extending the scope of our account of justice so that the social ethos is
zyxwv
zyx
zy
zyxw
zy
(Salt Lake City: University of Utah Press, isg2); WAI: “Where the Action Is: On the Site of
Distributive Justice,” Philosophy& Public Affairs26, no. I (1997):3-30; PAI: “The Pareto Argument for Inequality,” Social Philosophy and Policy 12 (1995): 160-85; IYE: If You’re An
Egalitarian,How Come You’reSo Rich? (Cambridge, MA: Harvard University Press, 2 0 0 0 ) .
2. Cohen’s most recent work focuses on these methodological questions. See his “Rescuing Justice From Constructivism” (manuscript, April 2001). My presentation here indicates one way that Quinean holism influences Rawls’s argument for justice as fairness.
365
zyxw
zyx
zyxwvuts
Taking People as They Are?
zyx
zyxw
brought explicitly within its scope. Reflection on the justifiability of incentive inequalities, then, sharpens our sense of the available options in
political philosophy, puts pressure on Rawls’s proposed marriage of political liberalism and egalitarianism, and highlights the role of ideas
about social order and human motivations in sustaining that marriage?
LAXITY
Cohen’s criticisms of the difference principle have shown, I think, that
the content of that principle is importantly indeterminate. This indeterminateness owes to the uncertain place of a social ethos in an account of
distributive justice. By the “social ethos” I mean-and I take Cohen to
mean-socially widespread preferences and attitudes about the kinds of
rewards it is acceptable to insist on, and, associated with those preferences and attitudes, a sense about the ways of life that are attractive, exciting, good, and worthy of pursuit. To use some familiar jargon, a social
ethos is a matter of ethics at least as much as it is a matter of moralityas much an encompassing view about how to live as about what we owe
to one another. But it provides important background for moral views
(including views in political morality) because it helps to shape our
judgments about when we are making reasonable demands and when
we are being asked to carry an unacceptably large burden-to give up
3. Standard responses to Cohen’s criticisms (in the literature and in informal discussion) are of two kinds. One claim is that Rawls is right to restrict the application of the difference principle to institutions and policies and not to extend it to the social ethos or to
individual preferences and attitudes (to the shape of the labor supply function), either because public institutions are the proper object of collectively chosen principles or because
of the excessive informational demands of an unrestricted difference principle or because
of the intrusive informational demands of such a principle. See Andrew Williams, “Incentives, Inequality, and Publicity,” Philosophy €+ Public Affuirs 27, no. 3 (1998): 225-47;
Jonathan Wolff, “Fairness,Respect, and the Egalitarian Ethos,”Philosophy C Public Afluirs
27, no. a (1998):97--122. A second claim is that Cohen exaggerates the distance of his view
from Rawls’s,since Cohen acknowledges that at least some incentive demands are reasonable, and is committed to allowing that an even wider range is reasonable. See David
Estlund, “Liberalism,Equality, and Fraternity in Cohen’s Critique of Rawls,” TheJournalof
Political Philosophy 6 , no. 1 (1998):99-112. Although I agree with the second claim, and
think there is some force to the first, both lines of argument may leave the unfortunate
suggestion that justice as fairnesshas no way, and perhaps needs no way, to respond to the
kinds of intuitively objectionable incentive inequalities that I explore in section 11. The
same point applies to Thomas W. Pogge, “Onthe Site of Distributive Justice:Reflections on
Cohen and Murphy,” Philosophy GPublicAfSairs29. no. a (aooo):137-69.
zyxwv
zyx
zy
366
zyxw
zyx
zyxwvuts
Philosophy G Public Affairs
too much in the way of the requisites of a fully flourishing life. The issue,
then, is whether an account of distributive justice should treat the social
ethos as part of the fixed background in judgments of justice or as one of
the proper objects of such judgments.
Difference Principle
To see how concerns about the social ethos raise questions about the
content of Rawls’s difference principle, consider the intuitive idea behind
that principle: that the way we live together-the social arrangements in
a democratic society, of free and equal people-should leave no one less
well-off than anyone needs to be. If social arrangements leave no one
less well-off than anyone needs to be, then no one has good reason to
complain: each person can reasonably think of herself as treated with
respect as an equal, one whose life matters as much as any one'^.^
The difference principle states a forceful ideal of distributive justice
for a democratic society, but also a very incomplete one. To fill out its
content, we need in particular to interpret the expression “needs to be”
in the phrase “no one is less well-off than anyone needs to be.” Thus,
zyxwvu
4. The argument for the difference principle sketched in this paragraph focuses on the
idea that persons can reasonably expect to be treated by others with the respect owed to
equals. Rawls claims that being treated with such respect fosters self-respect (it provides
the “socialbasis” of self-respect),and that self-respectis a fundamental precondition for
living a good life. These two claims together imply that being treated by others with respect
is in a person’s rational interest-that being treated with respect as an equal contributes to
the person’s good. This conclusion, in turn, enables Rawls to import any ethical argument
about the treatment that persons can reasonably expect from others as elements of being
treated with respect (for example, the argument for the difference principle briefly
sketched in the text above) into the original position as an argument based on judgments
of rational advantage made under ignorance. Thus, suppose we think that all members of
a society are treated with respect as equals only if they are all treated in accordance with
principle P. Then it is also true that being treated in accordance with Pprovides social support to self-respect, and therefore helps to ensure that each person lives well. The ethical
argument is that adopting P is required to treat everyone with respect as an equal. The
original position argument, then, is that P would be chosen under ignorance by persons
concerned with their own good-and who are therefore concerned with their self-respect,
and therefore with being treated with respect by others (which by hypothesis depends on
adopting 0.So the original position argument does not really model the ethical argument
by presenting an independent argument about rational choice under ignorance; it assumes the force of the ethical argument, and uses it, while also making the udditional substantive (and controuersiaZ1 assumptions about the connections between respect from others, self-respect, and a person’s good required to present it as a step in an argument about
what is rationally advantageous.
zy
367
zyx
zyxwvuts
Taking People as TheyAre?
suppose that people with scarce productive talents demand incentives
for the use of those talents. To clarify: when I say that they “demand incentives,” I do not mean that because they possess scarce talents and a
range of employment options they are able to bargain for compensation
above the reservation wage for which they would be willing to work: this
is not demanding an incentive but taking what you can get. In the case
of genuine incentive demands, people would genuinely prefer not to deploy their scarce, productive talents unless they are compensated at a
level that turns out to be higher than the average compensation of others. Consider a doctor who is perfectly able to doctor for the median
wage of (say) $30,000, and would live at least as well as everyone else if
she did, but really values writing literary fiction and is therefore unwilling to doctor for less than $ioo,ooo. Contrast this case with the doctor
with scarce surgical talents who would fully willingly use them for
$30,000 if she thought she could not get more for doctoring, but knows
that others are willing to pay more than that, and bargains to capture
more than the reservation wage for which she would be entirely willing
to work. The latter case seems very important in the world, but the former presents the philosophical problem that occupies me here.
One additional clarification: in explaining the relevant sense of “incentive demand,” I said that “people would genuinely prefer not to deploy their scarce, productive talents unless they are compensated at a
level that turns out to be higher than the average compensation of others.”Why the italicized clause? Because when it comes to incentives, the
crucial point is that the compensation simply turns out to be greater
than the average compensation that others receive; the incentive demander is not someone who strives for relative advantage and insists on
being paid more than others, whatever their level of compensation happens to be. If I demand to be paid $250,000,then I demand to be paid
more than anyone else in my department, given what others are in fact
paid, but it is another matter if I insist that, whatever they might come to
be paid, I must be paid 30 percent more. There is an important question
about whether demands for relative advantage-say the demand to be
paid 30 percent more-are especially suspect, but the incentives issue
arises even without any concern for relative position.
Returning now to the difference principle: suppose that we permit the
incentives, Are the least well-off then less well-off than anyone needs to
be? Perhaps, After all, they are less well-off than they would be if the pos-
zyxw
368
zyxw
zyx
zyxwvutsr
Philosophy & Public Affairs
zyxw
zyxwvuts
sessors of the scarce talents willingly used the talents without getting the
incentives: willingly used them, for example, because the benefits conferred by that use on other people fulfill what they take to be a duty to
contribute according to their abilities.5And surely such willing contribution is typically something the possessors of scarce talents could do.
Or should we say instead that they are no less well-off than anyone
needs to be because “needs to be” should be interpreted holding people’s preferences and attitudes as fixed-taking those preferences and
attitudes as given, so to speak. This is one way to interpret Rousseau’s famous phrase “taking men as they are, and laws as they might be.””So if,
for example, we think we ought to take preferences and attitudes as
given because the correct account of justice says that they are private
matters, then the payment of incentives leaves no one less well-off than
anyone needs to be, consistent with our doing what we ought to do, in
other respects.
Lax I Strict
To capture this distinction, Cohen distinguishes two interpretations of
the difference principle. On the lax interpretation, we hold preferences
(Cohen says “intentions”) as given and maximize the advantage of the
least advantaged relative to existing preferences (intentions). On the
strict interpretation, we maximize the well-being of the least advantaged
without taking preferences as given but treating the preferences instead
as assessable by our norms of justice.’ Although it suffices for drawing
attention to the distinction in interpretations of the difference principle,
this lax / strict distinction is too crude because it fails to capture intuitively relevant differences among the preferences that might underlie
incentive demands. Correspondingly,some interpretations of the difference principle take some but not all preferences as given for the purposes of maximizing the advantage of the least advantaged, while other
interpretations take all preferences as given. I will call the latter-the allpreferences-as-given interpretation-the ultralax difference principle.
zyxwv
zyxw
5. For defense of such a duty on the grounds that the acknowledgment of it would benefit the least advantaged by eliminating the need for incentive inequalities, see Joseph H.
Carens, “Rightsand Duties in an Egalitarian Society,”Political Theoryis,no. 1 (1986):31-49.
6. Although probably not the right way, as the quotation from Rawls at the start of this
essay indicates.
7. ‘hssessable”does not mean “alterable,”much less that their alteration ought to be
made a matter of public policy.
369
zyx
zyxwvuts
Taking People as TheyAre?
zyxw
Thus, consider five cases.
1. Ascriptive Group Preferences: Consider a community with a betteroff white racist majority whose members refuse to take on jobs that will
benefit the worse-off nonwhite minority unless the racists are paid substantial incentives. Because they are racists, they attach negative weight
to any benefits to the nonwhite minority and so require substantial material rewards to compensate for what they regard as costs to them. According to the ultralax difference principle, justice requires (and certainly permits) the payment of these incentives: by paying them, the
worse off nonwhite minority is made better-off, and given the racist social ethos, that minority cannot otherwise be made better-off. We might
tell a similar story for the case of incentives to men who regard economic benefits to women as a cost, or Brahmins who regard economic
benefits to Dalits as a cost, and, in general, for incentives paid to individuals who despise members of another group and assign negative weight
to benefits that flow to those membemH
2 . Class Preferences: Consider next an ethnically homogeneous society
divided into a better-off land-owning class and a worse-off labor-owning class. Imagine, too, that the land-owning class aims to preserve its
privileges: The members want to maintain social distance between the
better- and worse-off and are prepared to act in socially constructive
ways (use their land for productive investment) only if a substantial gap
is preserved between themselves and those who are less well-off (say,
the ratio of the group medians must be at least 5 0 ~ ) So
. here the ultralax
difference principle supports incentives to members of the privileged
class because the circumstances of the less well-off will not be improved
without the incentives.
3. Greedy / Needy: In a third case, the Greedy I Needy case, the greedy
live in a society with a needy group. Although the greedy are fully aware
that others in the society are living pretty badly (needy), their willingness to supply labor that benefits the less advantaged, needy group is
unresponsive to changes in expected reward until the changes are very
large-not because such discontinuous responsiveness helps in bargaining, or because they despise the less well-off, but because they genuinely would prefer tending their gardens to being captains of industry
zyx
zyxwvu
zyx
8. Although I focus on cases of Ascriptive Group Preferences, what matters, as the last
sentence in the text indicates,is the assignmentof negative weight, not that the groups are
ascriptively, rather than, say, confessionally, defined.
370
zyxw
zyx
zyx
zyxwvuts
Philosophy & Public Affairs
and finance unless the rewards are sufficiently great. If these preferences
are held fixed, then maximizing the advantage to the least advantaged
will mean providing large gains to the greedy to ensure small gains to the
needy. This case is different from Class Preferences because the greedy
are not assumed to care about relative advantage or social distance but
only about their absolute position: it is not true of them that-to borrow
some lines from Rousseau that I will return to later-they “value the
things they enjoy only to the extent that the others are deprived of them,
and they would cease to be happy if, without any change in their own
state, the people ceased to be miserable.+’ But is also essential to this
case, as distinct from the next, that the less well-off group in the society
is reasonably badly off.
4. Large Dispersion: In a fourth case, we once more have a population with a greedy, advantaged group with discontinuous labor supply
functions, but this time in a community with no very needy members.
Paying the incentives that are necessary, given de facto preferences, to
maximize the advantage of the least advantaged results in extreme income dispersion: a dispersion that we are inclined to find objectionable or excessive, either because of the preferences that produce it or,
more likely, just because it strikes us as objectionable that some people live that much better than others.‘”Using a piece of Rawlsian terminology here, let’s say that a “contribution curve” is a curve that represents the magnitude of the advantages to the better-off that are
needed to generate different levels of advantage to the least well-off.
So in Large Dispersion, the contribution curve is relatively flat: it rises
slowly from the origin, and its peak, where advantage to the least advantaged is greatest, is at some considerable distance from the 45degree line of equality.”
5. Absolute Incentives:In the final case, possessors of scarce talents demand incentives, but they are not-as in Ascriptive or Class Prefer-
zyxwvut
zyxwvut
9. Jean-JacquesRousseau, Second Discourse, trans.Victor Gourevitch, in TheDiscourses
and Other Early Political Writings, ed. Victor Gourevitch (Cambridge: Cambridge University Press, 19971,p. 184.
10. Of course, large dispersions are often condemned because of their effects on political equality or equality of occupational opportunity. I put those possibilities aside because
those concerns are supposed to be covered in justice as fairness by the requirements of the
fair value of political liberty and fair equality of opportunity.
11. Erin Kelly, ed., Justice as Fairness: A Restatement (Cambridge, MA: Harvard University Press, 2001) ,p. 66.
371
zyxw
zyx
zyxwvutsr
Taking People as TheyAre?
zyxwvu
zyxwv
ences-concerned with relative advantage, nor are the less well-off
badly off, as in Greedy / Needy, nor does the dispersion seem intuitively
objectionable or excessive, as in Large Dispersion. Instead it is simply
the case that those who possess scarce productive talents are unwilling
to use those talents in ways that benefit others unless the possessors are
compensated at a level that makes them better-off than those who lack
the scarce talents.’*Some Absolute Incentive cases will reflect the agent’s
moral view that a person deserves to be rewarded for making a socially
beneficial contribution; other cases will reflect an agent’s assignment of
greater weight to his / her own good in making decisions about occupational choice or intensity of effort.
I will return later to these cases, focusing on the first four because it
seems uncontroversial that at least some Absolute Incentive cases are
unobjectionable.
Earlier I mentioned Cohen’s distinction between strict and lax readings of the difference principle. But that distinction is too simple. In light
of the variety of incentive inequality cases, we might describe a family of
“lax” difference principles, distinguished by reference to the types of
cases in which each member of the family permits incentive inequalities. Thus Cohen says that “in its lax reading, it [the difference principle]
countenances intention-relativenecessities” (IIC, p. 3n),meaning that it
countenances, as consistent with justice, inequalities that are necessary
given the intentions of talented producers and not simply those inequalities that are necessary apart from people’s intentions. But he does
not say whether it permits all or some intention-relative necessities.
What seems fully clear is that he thinks-and here I certainly agreethat the ultralax difference principle is not at all plausible as a requirement of justice. But that leaves a large field open.
At the same time, I do not think that Rawlsian justice condones the ultralax difference principle. I find it inconceivable that Rawls believes
that the inequalities that would emerge if we were, for example, to pay
large incentives to the greedy to confer small benefits on the needy are
an acceptable part of a just society. Certainly he thinks that the needy
have a reasonable objection to such inequalities:
12. In this case, as in the others, the preferences of people who are paid incentives may
be completely typical of the society. Indeed, preferences in the society may be completely
uniform.What distinguishes the circumstancesof those who get the incentives from those
who do not may simply be that the former possess a scarce resource or talent.
372
zyxw
zyx
zyxwvutsr
Philosophy G Public Affairs
zyxw
zyxwv
Sometimes the circumstances evoking envy are so compelling that
given human beings as they are, no one can reasonably be asked to
overcome his rancorous feelings. A person’s lesser position as measured by his index of objective primary goods may be so great as to
wound his self-respect; and given his situation, we may sympathize
with his sense of loss. Indeed, we can resent being made envious, for
society may permit such large disparities in these goods that under existing social conditions these differences cannot help but cause a loss
of self-esteem. . . . When envy is a reaction to the loss of selfrespect in circumstances where it would be unreasonable to expect
someone to feel differently, I shall say that it is excusable. Since selfrespect is the main primary good, the parties would not agree, I shall
assume, to count this sort of subjective loss as irre1e~ant.I~
This passage condemns Large Dispersion, and surely if there are reasonable objections to very large differences, then there are even stronger
such objections to very large differences when the less advantaged are
badly off.
But nothing in the formulation of the principles of justice as fairness
directly condemns the inequalities generated by such incentive payments as unjust. And this leads me to conclude-I will argue for the conclusion later on-that Rawls must,have some alternative way to exclude
at least some of the incentive inequalities permitted under the ultralax
difference principle-that is, alternative to introducing qualifications or
limiting conditions in the statement of the principles themselves. More
particularly, Rawls never says what we should do if, for example, Large
Dispersion or Greedy / Needy inequalities arise within institutions that
meet the principles of justice as fairness. And that leads to the thought
that the alternative strategy is to argue in effect that the objectionable
incentives inequalities will not arise and therefore do not command
separate treatment in an account of justice.
zyxwv
Rigorism
Before addressing these issues, I want to take note of two points in
Cohen’s view that help to locate more precisely the challenge presented
by these cases. First, Cohen rejects the difference principle in its strict
13. John Rawls, Theory of Justice, revised edition (Cambridge,MA: Harvard University
Press, iggg), p. 468, emphases added.
373
zyx
zyxwvuts
Taking People as They Are?
reading, as too “rigoristic” (IIC, 303-4; IYE, 206 na4, 2131136). Thus, it is
not unreasonable for self-interests to play a role in deciding how to
spend one’s time-that is, in making decisions about occupational
choice-and so not unreasonable for individuals to demand some incentives for performing work that brings large benefits to the less welloff but that they would otherwise prefer not to do. Moreover, Andrew
Williams rightly observes it seems hard to say precisely how large an incentive a person might reasonably demand, and David Estlund observes
that it is difficult, too, to say that reasonable incentive-demanding, inequality-generating behavior is confined to occupational choice motivated by self-interest, as distinct, for example, from concerns about the
well-being of one’s family or social group or about moral requirements
other than the requirement to benefit the least advantaged.I4And Cohen
appears to acknowledge these additional prerogatives. So Cohen accepts the justice of at least some incentive inequalities that emerge in
absolute incentives cases, and surely must agree that it is difficult precisely to circumscribe the acceptable cases. For this reason, I put these
cases to the side.
Second, Cohen urges that the extent to which people may reasonably
pursue their self-interestdepends on how badly off the least advantaged
are. If the worst off are not badly off, then the better-off may reasonably
give more weight to their own interests: intuitively, improvements in the
conditions of the least advantaged expand the scope of acceptable incentive-demanding behavior. Thus Cohen notes that he sometimes
refers to the least well-off as the badly off, or says that they are in conditions of “want.”And he acknowledgesthat this language might seem unacceptably tendentious because the least well-off might not be too badly
off. So it might seem that he is illegitimately drawing fuel for a general
criticism of incentives from the special Greedy / Needy cases, in which
the worst-off are badly off. Thus Cohen says, “It might be objected that
in characterizing the position of the less well off as one of deprivation or
want, 1 am unfairly tilting the balance against the incentive argument;”
incentive demanders generally are shown in a bad light because they are
seen as insisting on large rewards in circumstances in which others are
very badly off. In response, he says: “Where the worst off are not too
badly off, it looks more fanatical to assign absolute priority to their
zyx
z
14. Williams, “Incentives,”and Estlund, “Liberalism.”
374
zyxw
zyx
zyxwvuts
Philosophy G Public Affairs
zyxw
claims. But the stronger the case for ameliorating the situation of the
badly off is, the more discreditable (if I am right) the incentives argument is on the lips of the rich” (ICI, 303). So incentive demands are especially objectionable in Greedy / Needy cases, certainly more so than in
Absolute Incentive or Large Dispersion case^.'^
Reality?
To conclude this sketch of the issues, I want to register a point on which
Cohen and I may disagree, although the point is about the world, not
about philosophy. Cohen suggests-and it is a suggestion, not an argument-that the repulsive inequalities we now observe in the United
States and other advanced economies are incentive inequalities of the
kind under examination here. He suggests, that is, that the inequalities
exist because the possessors of scarce talents have high reservation
wages as a result of the social ethos that dominates these societies. In
the case of high reservation wages, again, a person is genuinely unwilling to perform a task unless she is highly compensated, that is, compensated in ways that make her better off than others. Cohen calls this the
standard case (PAT, i72-73), and he suggests, more particularly, that differences of social ethos provide a plausible explanation of current differences in income dispersion between Germany and the United States,
and of the relatively limited dispersion in England in the early 1950s
(WAI, 27). I have some doubts, and wish to register them because Cohen’s suggestion may lead to the conclusion that repulsion at current
levels of inequality requires endorsement of his view.
Consider,then, an alternative explanation of high dispersion. It could
be that possessors of scarce talents are typically perfectly willing to work
at the same occupation and with equal intensity for much less than the
wage they command in the market but are able to secure large advantages in part because of a combination of high demands for their talents
(which present them with a range of opportunities), relatively low supplies of comparable skill, citizens’s unwillingness or lack of desire to tax
high levels of market compensation, and weak capacity for collective action by less skilled workers. If this is the case, then the problems of inequality are less matters of a social ethos than a combination of political
15. I suspect that people will generally agree, too, that incentive demands in Greedy /
Needy cases are worse than such demands in cases of Class Preferences. I am grateful to
Paula Casal for correcting an earlier draft on this point.
z
375
zyx
zyxwvuts
Taking People as TheyAre?
failures: a failure to impose tax rates that would not actually diminish
the supply of labor or otherwise distort the operations of labor markets,
a failure in the systems of education and training to increase the supply
of talents, and a failure of labor market institutions (unions and minimum wage laws) to reduce wage dispersion.
We find some suggestive evidence for this non-ethos-based explanation in the very low elasticity of labor supply. Thus the number of hours
that people work seems to be only very mildly responsive to wage
changes: a 1 percent wage change is associated with a 0.05 percent increase in hours worked.16More generally, employment decisions seem to
be pretty insensitive to changes in compensation, which is not what we
would expect if people were already working at their reservation wages,
which happened to be very high.
zyxwv
INSTITUTIONS
AND ETHOS
Ethology, it is fair to say, is not a well-developed discipline. We do not
know much about either the sources or consequences of a social ethos.
But one outlook places considerable weight in accounting for a society’s
ethos on the institutions of the society.Thus, Rousseau said “that man is
naturally good, and that it is solely by [our] institutions that men become wicked.””He describes the ethos of a society founded on inequality this way: “If one sees a handful of powerful and rich men at the pinnacle of greatness and fortune while the masses grovel in obscurity and
misery, it is because the former value the things they enjoy only to the extent that others are deprived of them, and they would cease to be happy
if, without any change in their own state, the people ceased to be miserable.”’8Rousseau here makes a strong assertion about the social effects
of an ethos: it is because of desires for relative advantage-desires that
reflect the valuations of the wealthy-that disparities are as large
as they are. But he also thought inegalitarian institutions generate the
dispersion-enhancing attention to relative advantage.
zyxwv
16. For a summary of the major empirical studies on labor supply elasticity for men, see
John Pencavel, “LaborSupply of Men,” in Orly Ashenfelter and Richard Layard, eds., The
Handbook ofLubor Economics, vol. 1 (Amsterdam:North Holland, 1986), 3-102.
17. My translation from Rousseau’s Lettres a Malesherbes, in J.J. Rousseau, Oueures
completes, vol. 1, ed. Bernard Gagnebin and Marcel Raymond (Paris: PlCiade, 19591, p. 1136.
18. Rousseau, Discourse on Inequulity, p. 184, emphasis added.
376
zyxw
zyxw
zyx
zyxwvutsr
Philosophy G Public Affairs
Now Rawls formulates his theory of justice so that it applies to basic
institutions and policies, from which Cohen concludes that if changes in
preferences and attitudes-in the social ethos-are necessary to improve the conditions of the least advantaged, then the Rawlsian view
would not require such changes. I think this conclusion is misleading,
and this section of the article will present the most straightforward reasons for thinking so: namely, that Rawls shares with Rousseau the thesis
that institutions make a large difference to ethos. Once we see why
Rawlsian justice may require some changes in the social ethos, then we
will be in a better position to understand the general Rawlsian strategy
for addressing the four objectionable cases mentioned earlier.
Thus, suppose contribution curve A lies higher than curve B at every
point from the origin so for any specified level of advantage to the least
advantaged, A (the higher curve) requires less inequality than B (a
smaller range), and for any specified level of advantage to the better off,
A gives greater advantaged to the least advantaged than B. We might
think of each curve as representing an ethos. So Cohen claims that a failure to raise the contribution curve from B to A would not, in Rawls’s
view, amount to a failure to make the society more just because principles of justice apply to institutions. They operate relative to a background of preferences, attitudes, and sensibilities; they do not assess
preferences, attitudes, and sensibilities themselves. So if the framework
of social and political institutions maximizes the advantage of the least
advantaged, given the regnant ethos-expressed by curve B-then the
distribution is ips0 facto just. “There is, Rawls would of course readily
agree, scope, within a just structure, for distribution-affecting meanness
and generosity, but generosity, though it would alter the distribution,
and might make it more equal than it would otherwise be, could not
make it more just than it would otherwise be, for it would then be doing
the impossible, to wit, enhancing the justice of what is already established as a perfectly just distribution by virtue merely of the just structure in conformitywith which it is produced” (WAI, 13).
In a similar vein, Cohen comments on a proposal by Ronald Dworkin
“that a Rawlsian government might be charged with a duty, under the
difference principle, of promoting . . . an [equality-promoting] ethos”
(WAI, 13). To which Cohen responds that while there might be such a
duty, it could not be imposed on grounds of justice, because the distribution is just if the institutions are just, and since the ethos is assumed
not to be part of the institutional structure, an equality-enhancing
zyx
zyxw
377
zyx
zyxw
zyxwvuts
TakingPeople as They Are?
change of ethos does not affect the justice of the institutions, and therefore not the justice of the distribution.
But this seems misleading, for a straightforward reason: It might be
that changes in institutions and policies would change the distribution
by shifting the contribution curve from B to A and produce that shift by
changing the preferences, attitudes, and sensibilities that constitute the
social ethos.''' Surely it could not be that principles of justice that require
us to adopt the institutions and policies that make the greatest contribution to the least advantaged instruct us not to make the changes when
the effects on the least advantaged come from changes in the social
ethos that result from institutional changes. One way to raise the contribution curve is to foster a wider distribution of marketable skills, thus
limiting the market power of possessors of scarce talents. But if, without
infringing basic liberties or limiting fair equality of opportunity, we can
make institutional changes that foster a more solidaristic direction and
thus raise the contribution curve, then why shouldn't such changes be
required by justice?To be sure, the justice of the distribution reflects the
justness of the basic structure within which it emerges. But the basic
structure is just in part because it produces the ethos that raises the contribution curve.
Thus, consider two not entirely implausible examples, each of which
suggests that institutional changes might lift the contribution curvethus reducing the extent of inequality needed for improving the wellbeing of the least advantaged-by changing the social ethos. Arend
Lijphart distinguishes two types of democracy: consensual and majoritarian (or Westminster-style).In more consensual forms of democracy,
we have, inter alia, multiparty arrangements (typically founded on proportional representation) with executive power sharing among the parties, cooperation between legislative and executive, and coordinated
representation of interests; in addition, consensual systems have strong
federalism, judicial review, relatively rigid constitutions, strong bicameralism, and independent central banks.20Majoritarian democracies differ on each of these dimensions. Intuitively, the distinction is that con-
zy
zyxw
z
19. In a n excellent (and as-yet) unpublished paper, Roxanne Fay argues this point with
considerable force. I have benefited greatly from reading her account and discussing it
with her.
20. See Arend Lijphart, Patterns of Democracy: Gouernment Forms and Performance in
Thirty-SixCountries (New Haven:Yale University Press, iggg), esp. chapter 16, on the differences that the different forms of democracy make. I am treating the two kinds of democracy as polar types, although of course they come in degrees.
378
zyx
zyxwvuts
Philosophy& Public Affairs
sensual democracies demand broader political coalitions than majoritarian democracies: policies can only be adopted with broad support
across parties, groups, and regions. According to Lijphart, these differences between consensual and majoritarian forms of democracy are
consequential: the institutional differences make a difference to outcomes. In particular, consensual democracies show stronger political
representation for women, greater economic equality, and stronger welfare states.
I do not wish here to assess the evidence for these findings or whether
Lijphart’s outcomes are the right ones for evaluating forms of democracy within justice as fairness. But the findings are at least suggestive.
Assuming, with Lijphart, that consensual and majoritarian democracies
are both capable of satisfying the first principle of justice, we might argue for a more consensual form of democracy because it appears to be
more suited to advancing the requirements of fair equality and the difference principle that fall under the second principle.
Of particular relevance is the not wildly implausible (though certainly
hard to demonstrate) claim that consensual institutions are associated
with a strong welfare state and greater economic equality because those
institutions,which they demand broad political coalitions, foster a more
solidaristic and inclusive ethos.z’So if we can improve the conditions of
the least advantaged by shifting to a consensual form of democracy,
then why is the change not required by justice? How could the fact that
the institutional change lifts the contribution curve by changing the
ethos undercut this conclusion? And if the conclusion remains intact,
then this seems to be a case in which justice requires a change in the
ethos, and the sharp distinction between focusing on institutions or focusing on the ethos may be misleading.
A similar thought is suggested by the role of labor market institutions-in particular, union density and level of the minimum wage-in
accounting for cross-national differences in wage dispersion and longitudinal changes in such dispersion. In particular, differences in labormarket institutions arguably account for much of the disparity between
the relatively limited growth of German and French wage dispersion in
the 1980s and the much greater growth of such dispersion in AngloAmerican economies. Once more, the mechanism is not clear, and ar21.
Of course, a solidaristic ethos may explain both consensualismand greater equality.
z
z
379
zyx
zyx
zyxwvuts
Taking People as They Are?
guably “underlying social and political attitudes play a major role in
shaping labor-marketbehavior”*’-that is, arguably, the causation moves
from ethos to institutions to performance. But it also seems as arguable
that causation goes from the institutions to the ethos: that labor market
institutions reduce wage disparity by fostering (or at least reinforcing) a
sense of social solidarity and social inclusion. And if that is right, then
we need to change the ethos.
Suppose, then, that we can improve the well-being of the least advantaged by changing either the form of democracy or the shape of labor
market institutions. And suppose, too, that these institutional changes
work their effects on the contribution curve by changing the social
ethos-in particular, by reducing the incentives demanded by the better
off. If so, then justice requires the institutional changes, unless we are
prepared to say that the existing institutions are just because they “make
the most” of the existing preferences and attitudes. Cohen seems to disagree. He says, for example, that the difference principle was better realized in Germany in 1988 than in the United States because (roughly)
wage dispersion was smaller in Germany. But because the smaller dispersion was, Cohen assumes, a result of ethos not law, Rawls cannot
agree with this proposition about the relative justice of Germany and the
United States. Of course he clearly could say this if the difference in
ethos were the result of institutional differences.
With.one qualification: Suppose we reject, as too demanding, a blanket condemnation of incentive demands, and thus acknowledge that inequalities generated by reasonable self- or group-interests and associated incentive demands are fully compatible with justice. Whatever the
permissible level of incentive demands, any smaller demands would
also be permissible. So if some incentive demands are acceptable, then
it follows that there is an acceptable range of incentive demands. And if
we have an acceptable range, then we could have two systems; the incentive demands are smaller in one than in the other, but both are
within the acceptable range. So suppose (improbably enough) that the
German ethos and the American ethos both give rise to incentive de-
zyx
zyxw
22. Philippe Aghion, “Inequalityand Economic Growth,” in Aghion and Jeff’reyWilliamson,
Growth, Inequality, and Globalization (Cambridge: Cambridge University Press, 1998),p. 73.
For discussion of the issues in this paragraph, see Aghion, “Inequality,”pp, 68-73, and Richard
Freeman and Lawrence E Katz, ed., Differencesand Changes in Wage Structures (Chicago and
London: University of Chicago Press, i995),esp. the introduction by Freeman and Katz.
380
zyxw
zyxw
zyx
zyxwvu
Philosophy G Public Affairs
mands that lie within the acceptable range. Then even if a shift in institutions led to a more solidaristic ethos which in turn reduced the inequalities required to improve the circumstances of the least advantaged, that institutional change might not be required by justice. But
that’s because the change of ethos would not produce a more just distribution on any but an unacceptably rigorist view.
But this leaves a large question. What happens if the ethos-if attitudes, preferences, and sensibilities-are relatively unresponsive to institutions? I have said that if changing institutions would improve the
minimum at a lower level of inequality, then justice requires the
change, even if the effects are mediated by a change in ethos: even if the
principles apply to the institutions, justice would require promoting
the ethos that reduces incentives demands. But how general is this
case? Can we generally assume that objectionable forms of incentivedemanding behavior, and the inequalities that issue from them, are responsive to institutions?
THEFOURBADCASES
zy
To address this question, I propose to focus on the four bad cases I
sketched earlier: Ascriptive Preferences, Class Preferences, Greed /
Needy, and Large Dispersion. I believe (and will indicate along the way
some reasons for believing) that Rawls thinks (rightly) that incentives
paid in response to such preferences are objectionable: the inequalities
that flow from acceding to such incentive demands are intuitively objectionable, so we would not be in reflective equilibrium if our conception of justice Fermitted them. But it does not follow that they must be
condemned directly by a principle of justice. The important point is
that inequalities of the objectionable kinds do not emerge in a just society. And if they don’t, then principles formulated for institutions will
suffice. Thus, some inequalities will be condemned directly by a principle of justice, while others will not emerge in a just society because of
the operations of just institutions, at least on plausible assumptions
about human beings and social order-assumptions of the kind that
are in any case required if a just society is to take hold in human
thought and motivation.
So there are two related strategies for responding to the objectionable
cases, corresponding to two reasons for thinking that institutions pro-
381
zyx
zyx
zyxwvuts
Taking People as TheyAre?
zyxw
vide the fundamental focus of principles of justice.23A first reason is that
institutions play a large role in shaping economic and political outcomes, given preferences: of special relevance in the current context is
the thought that arrangements associated with achieving fair equality of
opportunity-social investment in human capital, wide dispersion of
capital-will address some of the troubling cases. The second reason is
that institutions play a large role in shaping a society’s culture and the
identity of members: “But what about the character and interests of individuals themselves? These are not fixed or given. A theory of justice
must take into account how the aims and aspirations of people are
formed; and doing this belongs to the wider framework of thought in the
light of which a conception of justice is to be explained.” Rawls continues that “an economic regime . . . is not only an institutional scheme for
satisfying existing desires and aspirations but a way of fashioning desires and aspirations in the future. More generally, the basic structure
shapes the way the social system produces and reproduces over time a
certain form of culture shared by persons with certain conceptions of
their good.”“
Consider how these two considerations-these two aspects of the importance of social arrangements-bear on the four objectionable cases.
Take first the cases of preferences for relative advantage, whether Ascriptive or Class. Here, the expectation is that two factors will limits the
presence of such preferences in a society in which the institutions satisfy
the principles. First, members of a just society, with equal rights, will acquire the conception of one another as equals, and this will limit the role
of specifically ascriptive preferences, which are founded on views about
the lesser significance of people who belong to certain groups and a corresponding sense of entitlement to advantage over those groups.
Moreover, second, preferences for relative advantage more generally-even if not of the ascriptive kind-should play a limited role.
Rawls, recall, derives principles of justice on the assumption that people
zyx
23. In the “Basic Structure,”Rawls discusses these two kinds of reasons in sections IV
and V, respectively. The second kind of reason, having to do with the effects of social
arrangements on identity and culture, is not mentioned in Theory of justice in the initial
discussion of why the focus of the theory is the basic structure (see p. 7 ) , although it is introduced later, at p. 229,in the account of justice and political economy. For discussion of
the two distinctreasons,see Rawls,Justice as Fairness:A Restatement, pp. 52-57.
24. Rawls, “Basic Structure,” in Political Liberalism (New York: Columbia University
Press, 1996), p. 269.
382
zyx
zyx
zyxwvuts
Philosophy6 Public Aflairs
care about their absolute score, not their relative score in the distribution of resources: parties in the original position do not “try to gain relative to each other.”2sThus, individuals are assumed, for example, not to
be given to envy (if they are less well-off) or to grudgingess (if they are
better off). But to show that this assumption, made for the purposes of
choosing principles, is reasonable, we need to show that people in a society with just institutions will not be so prone to concern about their
relative situation. Here the idea is that equal standing as citizens provides support for self-respect independent of relative standing in the
distribution of income, provided the income dispersion is not too great.
The thought is that concerns about relative standing arise from concerns about respect, about how one is valued by others; in particular,
they reflect a lack of assured respect from others, which in turn may undercut the fundamental good of self-respect. But the presence of basic
liberties and fair equality provides that respect: “Self-respect is secured
by the public affirmation of the status of equal citizenship for all.”21i
And
this, if true, will substantially diminish attention to relative shares, and
perhaps generate diminishing concerns about absolute scores as well.
What about Large Dispersion? Large dispersion cases-objectionable
in part because the sheer magnitude of differences may be taken as indicating lack of respect (see above, p. 372)-are supposed to be taken care
of by the condition of fair equality of opportunity (if equality in public
status generates diminishing concerns about absolute scores, then dispersion will also be less of a concern). “[Bloth the absolute and the relative differences allowed in a well-ordered society are probably less than
those that have often prevailed. Although in theory the difference principle permits indefinitely large inequalities in return for small gains to
the less favored, the spread of income and wealth should not be excessive in practice, given requisite background institution~.”~~
Intuitively,
the idea is that large investments in education and training-investments in human capital-will substantially increase the supply of skills,
thus limiting both the bargaining power of the possessors of scarce skills
who are prepared to take more than their reservation wage, and the capacity of persons with scarce skills and high reservation wages to command those wages: “The idea is that given the equal basic liberties and
zyxwv
z
zyxwvu
Theory ofJustice, p. 125.
Ibid., p. 478.
27. Ibid., p. 470.
25.
26.
383
zyxw
zyxw
zyx
zyxwvuts
Taking People as They Are?
zyx
zyxw
fair equality of opportunity, the open competition between the greater
numbers of the well-trained and better educated reduces the ratio of
shares until it lies within the acceptable range.. ..With background institutions of fair equality, the open competition between the greater numbers of the well-trained and the better educated reduces the ration of
shares until it lies in the acceptable range.”28The idea is that the “advantaged cannot unite as a group and then exploit their market power to
force increases in their income.”29And if this hopeful assumption is
wrong? Then you “might also have to reconsider the soundness of the
difference principle,” especially if the advantaged group is large.3n
As to the Greedy / Needy cases, surely they are objectionable if Large
Dispersion cases are objectionable. A large dispersion in a population
that includes people in conditions of “want”is plainly worse than a large
dispersion in a population all of whose members are at least moderately
well-off. Those who are less well-off can certainly reasonably object to
the demandingness of the advantaged and to the resulting inequalities.
After all, if a large absolute difference suggests lack of respect, then so,
afortiori, does such a difference when the less well-off person is badly
off. Here, again, the expectation is that the objectionable inequalities
will be taken care of in part by generating a greater dispersion of skills,
thus increasing the supply of skills and reducing the capacity of the advantaged to enforce high wage demands. In addition, the human capital
investments designed to ensure fair equality should improve the capacity of the less well-off to succeed in the labor market.
I do not intend here to assess the plausibility of these views. Instead I
offer them to show that it is entirely consistent with the Rawlsian version
of egalitarianism to affirm that the social ethos matters to distributive
justice. Rather than denying the relevance of the social ethos, the claim
is that preferences and attitudes that would, if they were in place, lead to
objectionable distributions are neither likely to arise within just institutions because of the educative effects of those institutions, nor will have
their objectionable effects once the other principles of justice are in
place, because of the broad dispersion of human and nonhuman capital. To be sure, Rawls does not consider what we ought to do-what conclusions about principles of justice we should draw-in the face of an
28.
zyxwv
z
zyxwvu
Restatement, p. 67.
29. Ibid.
30. Ibid, note 35.
384
zyxw
zyx
zyx
zyxwvutsr
Philosophy & Public Affairs
ethos both that generates unacceptable distributions and is difficult to
change through changes in laws and policies. But that is not because an
account of justice is to take preferences as given, and to live with their
effects, whatever they are. Instead, the claim is that such cases are unlikely to arise in part because preferences and attitudes are so influenced by institutions-that the social ethos is important but so pervasively shaped by institutions and the political culture that it does not
require independent treatment by principles of justice.
zyx
CONCLUSION
My principal thesis might be stated as a biconditional: We can be egalitarians and endorse a political conception of justice whose principles
focus on the basic structure and its institutions if but (arguably) only if
we are prepared to endorse a set of substantive assumptions about how
social arrangements work, assumptions about the pervasive influence
of social institutions on political-economic outcomes and on culture
and identity. If we do not think that we can substantiallyreduce income
dispersion by investing lots in education and training, and by keeping
the ownership of other forms of capital widely dispersed, then we will
have to face Large Dispersion and Greedy / Needy cases. And if we do
not think, with Rousseau and Rawls, that ensuring a decent standard of
living and equality of basic liberties leads people to downplay concerns
about their relative position-to invest less of their identity in their
standing relative to others, and more in their success in pursuing a plan
of life-then we will need to face Ascriptive Group and Class Preference
cases. In saying this I do not mean to be saying anything that Rawls
would deny or find surprising; indeed, a central part of his view is that
an account of justice belongs to a “wider framework of thought in the
light of which a conception of justice is to be e~plained.”~’
That framework of thought includes substantive, arguably deep, but contingent
truths about human beings and social cooperation-perhaps hopeful
assumptions, articles of reasonable faith that are consistent with the rest
of our empirical knowledge. In any case, those truths, as Rawls presents
them, assign a large role to social institutions in shaping outcomes,
identities, and aspirations.
zyxw
31. Rawls, Political Liberalism,p. 269.
385
zyxw
zyxw
zyxwvuts
Taking People as They Are?
If we endorse those views, then we do not need to have principles of
justice that address the four cases because they simply can be expected
not to arise. We need not have anything definite to say about what we
would do if the satisfaction of the difference principle led to Wide Dispersion or Greedy I Needy problems, despite the full realization of fair
equality of opportunity. All we can say in advance is that if they did arise,
we would not be in reflective equilibrium and something would have to
give.
On the other hand, if we reject the substantive assumptions that
Rawls endorses, then we have, I think, two principal options. We can
preserve political liberalism, with its focus on basic institutions, but
then we will need to give up on any strong formulation of egalitarianism
because we will locate important aspects of the social ethos outside the
scope of our conception of justice: Taking preferences and attitudes as
given, we will likely be led-assuming we endorse the difference principle-to accept Large Dispersion and Greedy / Needy inequalities.
Or we can preserve our egalitarianism, and extend the scope of our
views of justice outside the arena of institutions to include the social
ethos. That outlook may seem worrisome-to invite political regulation
of private social attitudes. But such worries are misguided, for two reasons. First, whether attitudes count as private and inappropriate as an
object of public concern is not something we can decide in advance of
deciding on a view about what justice requires. “The spheres of the political and the public, of the nonpublic and the private, fall out from the
content and application of the conception of justice and its principles. If
the so-called private sphere is alleged to be a space exempt from justice,
then there is no such thing.”3’The thesis that a decision is private, and
should not therefore be regulated except for especially compelling reasons, is not best understood as a premise in normative political argument-because the decision is private, we should not regulate it-but as
a conclusion of such argument.
But more fundamentally, it is entirely open to the egalitarian to deny
that it would be either good or even possible to produce the relevant
changes in social ethos by authoritative political means. Instead, what is
needed may be a direct assault on the ruling social ethos through cul-
zyxw
z
zyxwvutsrq
zyxw
32. “The Idea of Public Reason Revisited,” Ilniversity of Chicago Law Review, 64 (1997):
765-807.
386
zyxw
zyx
zyx
zyxwvuts
Philosophy C Public Affairs
zyx
tural criticism and other tools available in what Habermas calls the “informal public sphere”-including expressions of disgust for “the handful
of powerful and rich men [wholive] at the pinnacle of greatness and fortune while the mass crawls in obscurity and misery,” open contempt for
the publicists and intellectuals who celebrate a culture of greed in a
world of pain, unbridled hostility for craven silence in the face of organized degradation and humiliation, along with constructive reminders of
the possibility of living well while living a minimally decent life.