Elements of Democratic Justice: Ian Shapiro Yale University

ELEMENTS OF DEMOCRATIC JUSTICE
IAN SHAPIRO
Yale University
DEMOCRATIC MOVEMENTS DERIVE much of their moral authority

from the hope they hold out of displacing unjust social arrangements. This
reflects the fact that the promise of democracy and that of social justice are
intimately linked in the modern political imagination. The great democratic
movements of the nineteenth century were less concerned to implement an
abstract democratic ideal than, as John Dewey observed, "to remedy evils
experienced in consequence of prior political institutions."' One only has to
think of the ways in which the lack of democracy and the presence of socia
injustice were fused in the ideologies of opposition to communism in the
Soviet bloc and the apartheid in South Africa to see that, in the twentieth
century no less than in the nineteenth, many people blame social injustice on
the lack of democracy and assume that democracy is an important weapon in
replacing unjust social relations with just ones.
Yet this popular expectation is at variance with much academic orthodoxy,
which recognizes that achieving political democracy guarantees nothing
about the attainment of social justice.2 In countries where the basic demo-
cratic institutions of popularly elected governments based on universal
franchise prevail, wealth may or may not be redistributed, minorities may or
may not be respected, opportunities may or may not be open to all, and
religious dissent may or may not be tolerated. Far from necessarily promoting
it, then, democracy can actually undermine whatever social justice might be
thought to require, and it is partly for this reason that bills of rights and other
constitutional restraints on democratic politics are argued to be worthwhile.
They limit the possibility of social injustice by constraining what can be done
by those who wield state power in democracy's name.
Once it is held that democracy should be constrained by the requirements
of social justice, a difficulty arises, however. There are many competing
theories of social justice and no evident way to choose among them. Else-
where I have argued that the standard attempts to resolve this difficulty fail,
POLITICAL THEORY, Vol. 24 No. 4, November 1996 579-619

? 1996 Sage Publications, Inc.
579
This content downloaded from 131.175.12.9 on Sat, 02 May 2020 03:57:50 UTC
All use subject to https://about.jstor.org/terms
580 POLITICAL THEORY / November 1996
and that the apparent tensions between democracy and justice need to be
rethought along lines that have more in common with popular expectations
than with academic orthodoxy. On my view, although democracy is not
sufficient for social justice, arguments about democracy and social justice are
more deeply entwined with one another than the conventional opposition
suggests. The mutual dependence of these two ideals is signaled by the fact
that, on the one hand, most arguments for democracy rest at bottom on
intuitions about what is just, and, on the other, if we dig deeply enough into
arguments about social justice we frequently discover that they rest on
appeals to democratic moral intuitions. This is not to say that commitments
to democracy and to social justice entail one another; it is to say, however,
that no account of either that undermines one's moral intuitions about the
other is likely to be judged satisfactory.3
Drawing out the implications of this observation, I have sought to develop
a view of social justice in which democratic considerations play a three-part
foundational role: in the definition of social goods, in the determination of
principles by which conflicts over goods should be resolved, and in the
appropriate stance toward implementing principles of justice in the actual
world of day-to-day politics.4 These three ways to be a democrat set the basic
terms of the view that I characterize as democratic justice, in contrast to
liberal, socialist, conservative, and communitarian views that are in wide
academic currency today. Although my view is conceived of as an alternative
to these, I mean to make it attractive to many of their proponents because it
depends less on alternative sets of assumptions about social justice and more
on making explicit the implications of democratic moral intuitions to which
many of them, on reflection, will find themselves committed. My earlier
defenses of democratic justice have been ground-clearing and programmatic,
amounting to a downpayment, at best, on a positive argument for democratic
justice. My aim here is to deliver on the first installment of the construc-
tive account.
Three preliminary points: First, democracy as I defend it is a subordinate
good. By this I mean that although democracy is necessary for ordering social
relations justly, we should resist every suggestion that it is sufficient, that it
is the highest human good, that it is the only human good, or that it should
dominate the activities in which we engage. Democracy operates best when
it sets the terms for our civil interactions without thereby determining their
course. Our lives require much else as well to be satisfactory, and it is wrong-
headed to expect democracy to deliver those other things. This conception
follows from the thought that because power relations form part-but not
all-of most collective activities, democracy appropriately conditions those
activities but it does not appropriately displace them. Although we should
Shapiro / ELEMENTS OF DEMOCRATIC JUSTICE 581
aspire to get on with our collective lives in democratic ways on my account,

we should nonetheless aspire to get on with them. The creative political
challenge is to devise mechanisms of institutional governance that can make
this possible.
Second, my argument for democratic justice is semicontextual in that it
varies partly, but only partly, with time and circumstance. Aspects of what
democratic justice might reasonably be thought to require may change over
time and vary both across the domains of civil society and from culture to
culture. This means that a satisfying elaboration of the argument can only
be developed as its injunctions are explored through a variety of contexts.
Shouldering that burden is the central task of the larger project on which I
am engaged. In the present article my attention is confined to exploring aspects
of the argument for democratic justice that can be defended in general terms.
Third, in making the case for democratic justice my central focus is on the
procedural and institutional level of analysis, not on matters concerning
higher order human interests and questions of ultimate justification. In this
regard my aprroach is similar to Rawls's in his "political, not metaphysical"
mode, although Rawls seems to me to press implausibly far the claim that a
political conception of justice can be developed independently of controver-
sial philosophical commitments.5 The account developed here rests on skep-
ticism toward the absolutist epistemologies and ontologies that a Platonist or
a classical Marxist might embrace, but this skepticism is political rather than
metaphysical; I take no position on whether or not the accounts are true, only
on whether or not it is wise to allow our lives to be governed by their
injunctions. This is not the same thing as philosophical neutrality, however,
because partisans of such absolutist views are likely to find their political
aspirations frustrated by the politics I am advocating in ways that many
philosophical fallabilists, pragmatists, empiricists, realists, and philosophical
antifoundationalists will not. My claim is that given the impossibility of
neutrality among ultimate philosophical commitments, the democratic con-
ception of social justice that I describe is the most appropriate foundational
political commitment.
It might be said that, having conceded that neutrality about questions of
higher order interests and ultimate justification is not possible, one is not free
to turn to the institutional and procedural level of analysis without first
defending the higher order assumptions on which a given analysis rests. This
conclusion seems to me to be unwarranted for three related reasons. It is true,
first, that every political theory rests on higher order assumptions, but it is
also true that all such assumptions are controversial. Consequently, if we put
off the questions of institutional design until the higher order questions are
settled, we may get to them at the time of Godot's arrival. In the meantime,
however, life goes on and we need grounds for preferring some institutio
arrangements over others. Second, although it is common to think that we
should start with general matters because people are more likely to agree
them and then move to more specific and divisive matters-to the details
wherein the devil is thought to lurk-exactly the opposite is often true, as
Cass Sunstein has usefully noted. A faculty may be able to reach agreement
that a particular person should be granted tenure even though its members
could never agree on the reasons why. By extension, in arguing about the
merits of different political arrangements, it is often wise to avoid-or at least
minimize-attention to controversial questions of higher order interests and
ultimate justification.6 That is the assumption behind the present discussion.
Last, because no one can do everything, it behooves practitioners in different
disciplines to reflect on where they are most likely to make a useful contri-
bution. In my view normative political theorists best devote attention to
analyzing how to structure the power-dimensions of human interaction,
leaving to psychologists, moral philosophers, and metaphysicians full analy-
sis of higher order interests and issues of ultimate justification. Although
every intellectual division of labor will be unsatisfactory from some defen-
sible point of view, choices inevitably have to be made in this regard. The
reader must judge whether I have made the right ones here.
TWO DIMENSIONS OF DEMOCRATIC JUSTICE
Democrats are committed to rule by the people. They insist that no arist
crat, monarch, bureaucrat, expert, or religious leader has the right, in virt
of such status, to force people to accept a particular conception of their pro
.common life. People should decide for themselves, via appropriate proce-
dures of collective decision, what their collective business should be. The
may reasonably be required to consult and take account of one another, a
of others affected by their actions, but beyond this no one may legitimate
tell them what to do. The people are sovereign; in all matters of collectiv
life they rule over themselves.
Although this is less often commented on in the academic literature,
democracy is as much about opposition to the arbitrary exercise of power
it is about collective self-government. In this connection Barrington Mo
remarks that historically democracy has been a weapon "of the poor and
many against the few and the well-to-do." Those who have actively soug
it in organized political movements "have wanted it as a device to increas
their share in political rule and weaken the power and authority of those w
actually rule."7 In the modern world at least, democratic movements have

derived much of their energy and purpose from opposition to socioeconom
legal, and political hierarchies that seemed capricious from a democratic
point of view. Rooted in the remnants of feudal and absolutist regimes an
shaped by the vicissitudes of conquest and chance, the political orders of
eighteenth- and nineteenth-century Europe and North America seemed to th
dispossessed to personify arbitrary hierarchy and domination. This reality, a
much as anything else, motivated working-class and other democratic mov
ments. The English philosophic radicals, the French and American revolu
tionaries, the nineteenth-century Chartists, and the anticolonial movemen
in the third world after the Second World War all wanted to free themselve
from hierarchical orders for which they could see no rationale or justificatio
It was to this oppositional dimension of the democratic ideal that Nelson
Mandela appealed at his sentencing for treason by a South African court in
1961. Conceding that he had disobeyed the law by inciting resistance to the
government, he nonetheless wondered whether "the responsibility does not
lie on the shoulders of the government which promulgated that law, knowing
that my people, who constitute the majority of the population of this country,
were opposed to that law, and knowing further that every legal means of
demonstrating that opposition had been closed to them by prior legislation,
and by government administrative action."8
Mandela's formulation might be taken to embody the conventional view
that democracy is primarily about collective self-government and only sec-
ondarily about opposition. Part of his claim, after all, is that he should not be
bound by "a law which neither I nor any of my people had any say in
preparing."9 But he insists also that the law lacks legitimacy because every
avenue of legal opposition to it has been sealed off. In a world of ideal political
institutions a derivative view of the place of opposition in democratic politics
might be sustainable. But in the actual world, where social orders come to be
what they are in morally arbitrary ways, and where all procedures of govern-
ment turn out on close inspection to be flawed, opposition must enjoy a more
independent and exalted status in a persuasive account of just democratic
politics. Or so I will argue; but first let us attend to the governance side of
the equation.
Collective Self-Government
If democracy is understood to require that the people be sovereign over

their collective goals, it exhibits considerable overlap with liberalism as a
political ideology. Both are rooted in antivanguardist conceptions of the
good; their proponents resist the idea that values should be imposed on peopl
against their wishes in the name of some greater social good. The reasons for
affirming this antivanguardist stance vary: they can range from commitmen
to variants of philosophical skepticism, pragmatism, and antifoundational-
ism, to beliefs in the psychological value of critical reflection and contested
authority, to the conviction that a degree of pluralism about values is socio-
logically or politically desirable. Liberals and democrats do not divide predict-
ably over these foundational matters, but for most of both some combination
of them issues in a principled resistance to moral vanguardism.'0
Liberals and democrats do divide predictably, however, over the institu-
tional implications they draw from their moral antivanguardism. Liberals,
who typically regard individual freedom as the greatest good, characteristically
focus on devices to protect the individual from the realm of collective action.
Democrats, by contrast, try to structure collective action appropriately to
embody the preferences of the governed. Liberals characteristically resist
this logic on the grounds that no procedure can fairly embody the preferences
of all the governed. For liberals, democratic decision rules all too readily
become devices by which phantom majorities-sometimes even manipulat-
ive minorities-tyrannize over individuals."
Although there is merit to the liberal argument, it rests on flawed assump-
tions about the nature of politics and about the limits of collective action.
Concerning the first, the characteristic liberal mistake is to focus on the forms
of tyranny performed by and through government as the only-certainly the
principal-kind of tyranny that should worry political theorists. Liberal
commitments to negative freedom, conventional constructions of public/pri-
vate dichotomies, and arguments for limited government are all shaped by
this governmentalist view of politics. Governmental power is one potential
site of domination, but there are many others that permeate the different domains
of "private" life. Government can be an instrument for mitigating domination
as well as a source of its generation. As a result, the choices and tradeoffs that
can minimize domination throughout society will likely defy such simplify-
ing formulae as "the government that governs least governs best."
The liberal view is flawed also because its proponents tend to think that
whether or not our lives should be governed by collective institutions is an
intelligible question about politics. Hence Robert Nozick's remark that the
fundamental question of political theory "is whether there should be any state
at all."''2 This view is misleading because the institutions of private property,
contract, and public monopoly of coercive force that its proponents charac-
teristically favor were created and are sustained by the state, partly financed
by implicit taxes on those who would prefer an alternative system. In the
modern world, Nozick's assertion makes as much sense as would a claim that
the fundamental question of astronomy is whether or not there ought to be

planets. A characteristic liberal sleight of hand involves trying to naturalize
or otherwise obscure liberal institutional arrangements in order to disguise
this reality. Such subterfuges have received more attention than they deserve
in the recent history of political theory; they cannot any longer detain us.'3
This is not to say that the liberal fear of majority rule is groundless. It is
to say that we need a different response to it than the conventional liberal one.
We can begin to develop this by noting, first, that there is no reason to think
that there is one best rule of collective decision. Different rules will be
appropriate in different domains of social life, depending on the nature of the
domain in question, the importance of the decision to participants, the
potential costs of decisions to third parties, and related contingent factors.
Such a plural attitude about decision rules flows naturally out of the view that
civil society is made up of domains of social action that differ qualitatively
from one another.14
Few liberals would deny this last claim, but they usually regard unanimity
rule as the best default option, the decision rule most likely to protect indi-
viduals against violations of their rights. This is at least partly why liberals
so often find markets attractive. Markets embody unanimity rule in that every
transaction requires the consent of both parties. On the liberal view, classi-
cally advocated in The Calculus of Consent, it is always departures from
unanimity that stand in presumptive need of justification, whether on effi-
ciency or other grounds. James Buchanan and Gordon Tullock argue in that
work that, under conditions of a hypothetical social contract, rational indi-
viduals concerned to safeguard their interests would insist on a hierarchy of
decision rules, starting with unanimity rule for constitutional matters. Next
come "those possible collective or public decisions which modify or restrict
the structure of individual human or property rights after these have once
been defined and generally accepted by the community." Foreseeing that
collective action may "impose very severe costs on him," the individual will
tend "to place a high value on the attainment of his consent, and he may be
quite willing to undergo substantial decision-making costs in order to insure
that he will, in fact, be reasonably protected against confiscation." He will
thus require a decision rule approaching unanimity. Last is the class of
collective actions characteristically undertaken by governments. For these
"the individual will recognize that private organization will impose some
interdependence costs on him, perhaps in significant amount, and he will, by
hypothesis, have supported a shift of such activities to the public sector."
Examples include provision of public education, enforcement of building and
fire codes, and maintenance of adequate police forces. For such "general
legislation" an individual at the constitutional stage will support less inclu-
sive decision rules, though not necessarily simple majority rule, and indeed,
within this class, different majorities might be agreed on as optimal for
different purposes. "The number of categories, and the number of decision-
making rules chosen, will depend on the situation which the individual
expects to prevail and the 'returns to scale' expected to result from using the
same rule over many activities." 15
This story is intuitively plausible only if we take the contractualist meta-
phor on which it rests seriously, assuming a prepolitical status quo where
there is no collective action and then a series of consensual moves that lead
to the creation of what we know of as political society. But, as Brian Barry,
Douglas Rae, and others have pointed out, once this assumption is jettisoned
there is no particular reason to regard unanimity rule as the most appropriate
default decision rule.'6 In the real world of ongoing politics, if I assume that
I am as likely to oppose a given policy as to support it regardless of whether
it is the status quo, then majority rule or something close to it is the logical
rule to prefer. Once we move from majority toward unanimity rule, we begin
to privilege the status quo. This will rightly seem arbitrary in a world that has
not evolved cooperatively from a precollective condition. In short, other
things being equal tyranny of the majority is something that people should
rationally fear, but not as much as they should fear tyranny of the minority."7
The preceding discussion reinforces the suggestion that there is no single
best decision rule for democratic governance. In domains of social life where
relations really do tend to approximate the contractualist story-in that they
are both created ex nihilo by the participants and are basically cooperative in
character-a presumptive commitment to unanimity rule is defensible. One
might think of marriage in contemporary America as a paradigm case. It is
created consensually, usually with the expectation that in important matters
day-to-day governance will also be consensual. (Indeed, with the advent of
no-fault divorce since the 1970s, we see an unusually strong form of the
unanimity requirement at work. In most American states either spouse can
insist-subject to a brief waiting period-on a divorce unilaterally: the
marriage continues only so long as both parties agree. Far from privileging
the status quo, this variant of unanimity rule makes it perpetually vulnerable,
since the rule is not defined by reference to the status quo but by in effect
recreating the conditions antecedent to it at the wish of either party.)'8
Many social relationships do not approximate the contractualist ideal; they
are not created ex nihilo in the sense that contemporary American marriages
usually are, and they are to a high degree structured by forces other than the
wills of the participants. Even childless marriages involve the generation of
reliances and externalities that can undermine their exclusively consensual
character. These are questions of degree, however. Social relations are often
not contractualist to anything like the extent that marriage is, even when su
reliances are taken into account; most obviously, think of parent-child rel
tions. Constitutional political arrangements are often pointed to as presum
tively contractualist because of their foundational character and their pl
in the social contract tradition. Such arrangements might once have been
consented to by the relevant parties, although even in the American foundin
a narrowly circumscribed class agreed in fact-and then not unanimously.
Generations later, whatever contractualist element these arrangements on
exhibited has receded into the mists of time. In such circumstances (and n
doubt there are others) there is no evident reason to regard unanimity rule
best on the grounds that it embodies the consent of the governed.
Nor are there good reasons to think that some alternative decision rule
should appropriately govern all relations where a contractualist element is
either missing or overdetermined by other factors. As the examples just
mentioned indicate, this is a heterogeneous class. In some domains, the sort
Rae evidently has in mind, majority rule is prima facie the best decision rule.
These include relations typically characterized by arms-length transactions,
where substantial aspects of the collective action in question are competitive
rather than cooperative, and where there are no obvious reasons to counte-
nance paternalistic decision making. They are also often circumstances in
which people are either born into structural relations that cannot easily be
escaped, or, if there is a contractualist element to their participation, it is
accompanied by a good deal of what Marxists like to think of as "structural
coercion." Whatever the surface appearances, the relations in question are
not substantially voluntary. Arguments for workplace democracy in which
majority rule plays a substantial role generally appeal to some combination
of these characteristics in justifying their appeal; Rae's logic supplies us with
reasons for accepting them."9
Not every noncontractualist or minimally contractualist form of associa-
tion should be governed by majority rule, however. Both Buchanan and
Tullock's and Rae's reasoning take it for granted that, ceteris paribus, deci-
sion-making costs should be minimized, for which they have sometimes been
criticized by participatory democrats.20 Rather than follow the participatory
democrat's reasoning (which creates difficulties of its own),2' the argument
here is that participation must itself be thought about in a context-sensitive
way. In some circumstances participation is no more than a cost to be
minimized, subject to achieving or preventing a particular outcome. Anyone
who has sat through enough faculty meetings will know what at least one of
those circumstances is. In other situations, institutions may reasonably be
structured to maximize participation. Juries are an obvious example. Una-
nimity is generally required just because it forces discussion and joint
deliberation, which, in turn, are believed most likely to lead to discovery of

the truth in trial courts, and that is the point of the exercise. Parent-child
relations are also noncontractualist relations (since the child does not ask to
be born, let alone to be born to the parent in question) that do not lend
themselves to governance by majority rule, at least not on many questions.
In these relations more flexibility is necessary in delineating the appropriate
scope for participation by different parties, because they include the total
dependence of young children on their parents, relations among more-or-less
equal adults, and relations between adults and their aging parents. And since
human beings are developmental creatures for whom decision making has to
be learned over time, there has to be space for regimes of domestic govern-
ance to adapt to peoples' changing capacities and dependencies. To be
appropriate, the decision rules governing domestic relations must be able to
respond to this complex reality.
Taking note of such complexities lends credence to the suggestion that
when they can be discovered and made to work, local solutions to local
problems are to be preferred. The kinds of knowledge that are pertinent to
democratizing an activity will often be disproportionately available to insid-
ers because of their hands-on experience and their participants' understanding
of the activity in question. In this respect the argument of democratic justice
is compatible with the aspirations of many who think of themselves as
communitarians. Notice, however, that there will be circumstances in which
no local decision rule can be made to work effectively from the standpoint
of democratic self-government, the most obvious being when the obstacles
to exit are insuperable for some yet easily overcome for others. The American
history of white flight from inner-city school districts since the 1960s stands
as eloquent testimony to that fact. Whether the substantially white middle-
class population opts out of the public school system, moves out of the inner
city (or both, to avoid both using and paying for the inner-city schools), its
ability to leave undermines democracy in educational provision. Local ma-
jority rule promotes white flight, but local unanimity rule gives disaffected
individuals veto power that enables them to avoid contributing to the provi-
sion of public education. In this type of circumstance, the presence of
collective action problems suggests that constraints other than choosing one
local decision rule over another should come into play.
The decision rules appropriate to different walks of life vary, then, with
the activity in question and the purposes around which it is organized. Yet to
say this is to solve one problem by raising another, because these activities
and purposes are never fixed and there is usually, perhaps endemically,
disagreement about them.22 How can we say that the nature of the activity in
question makes one decision rule more appropriate than another, having
conceded that those purposes and activities are inevitably in contention?

Whereas most liberals would say that all social relations should be redesigned
to approximate the contractualist ideal as much as possible (regardless of how
they are currently organized), the subordinate character of the democratic
commitment in the argument for democratic justice precludes my defending
an analogous claim. Instead it recommends a more pragmatic approach that
is antivanguardist in method as well as substance, because we should neither
accept things as they have evolved nor aspire to redesign them, tabula rasa.
Rather, the goal should be to take social relations as we find them and discover
ways to democratize them as we reproduce them. Democratic justice thus has
a Burkean dimension, but this is tempered by the aspiration to create a more
democratic world over time. Prevailing ways of doing things reasonably
make a partial claim on our allegiance, but this claim is conditional and
always subject to revision in democratic ways; the inertial legitimacy of
existing modes of governance can never achieve a status greater than that of
a rebuttable presumption. The creative challenge is to devise methods of
governance that both condition existing ways of doing things democratically
and open the way to their reevaluation over time.
Although there is no best decision rule for the governance of different
domains of civil society, a general constraint for thinking about decision rules
follows from what has been said so far: everyone affected by the operation
of a particular domain of civil society should be presumed to have a say in
its governance. This follows from the root democratic idea that the people
appropriately rule over themselves. To require that everyone affected should
have a say is not to require that this presumption be conclusive or that every
say should necessarily be of equal weight. There are often-but not always-
good reasons for granting outsiders to a domain (who may be subject to its
external effects) less of a say than insiders concerning its governance, and
even within a domain there may be compelling reasons to distribute govern-
ing authority unequally, and perhaps even to disenfranchise some participants
in some circumstances. What these circumstances are cannot be specified in
general, but we can say that we begin with a presumption of universal inclusion.
We can also say that proposals to undermine universal inclusion reason-
ably prompt suspicion, whatever their source. In the limiting case, if someone
sells herself into slavery, her agreement should be regarded as void ab initio.
Most incursions on inclusion are considerably less radical than selling oneself
into slavery is. As a consequence, evaluating policies and practices that limit
the nature and extent of the governed's participation in decisions that affect
them is more difficult (and controversial) from the standpoint of democratic
justice than is the case with slavery. In the ongoing world of everyday politics
there will often be circumstances in which inclusion is reasonably traded off
against other imperatives. But the general argument counsels suspicio

these tradeoffs; the burden of persuasion lies with those who advocate
Although the requirements of the principle of universal inclusion wi
with circumstances, it is possible to defend at least one constraint on it in
general terms that goes beyond the limiting case of slavery: Participation by
one individual or group in ways that render the participation of other
legitimate participants meaningless is unacceptable. For instance, it is esti-
mated that in the United States the health care and medical insurance
industries spent upward of $50 million in 1993-4 on advertising and lobbying
to kill the Clinton Administration's proposed health care reform legislation,
and that they would have spent whatever was necessary to achieve this result.23
Granting, arguendo, this account of the facts, we can say that an under-
standing of participation that permits such a result goes too far. It gives one
set of interests affected by the proposal the power to obliterate the participa-
tion of others, and to determine the result more or less unilaterally. To be sure,
this does not tell us which types of limitations on lobbying and political
speech are appropriate, but it sets an outer constraint on the debate.
The account of collective self-government defended here is causally
based. The right to participate comes from one's having an interest that can
be expected to be affected by the particular collective action in question. In
this respect the argument for democratic justice differs from liberal and com-
munitarian views, both of which tend to regard membership of the relevant
community as a trump (liberals by assumption, communitarians by express
argument).24 Once the contractualist way of thinking has been dethroned, it
is difficult to see any principled basis for regarding membership as primary.
On the view advanced here, the structure of decision rules should follow the
contours of power relations, not those of political memberships. Adopting
the causally based view has implications for a host of issues relating to
intergenerational justice and the handling of externalities. In a world in which
international military and environmental questions increasingly dominate
political agendas, whether or not one adopts it can be expected to be conse-
quential over an expanding portion of the political landscape.
It will be objected that serious difficulties arise in determining who is
affected by a particular decision and who is to determine whose claims about
being affected should be accepted. To provide a full defense of the causally
based view here would take us too far afield, but two points should be noted.
First, although who is affected by a decision is bound to be controversial, this
fact scarcely distinguishes causally based arguments from membership-
based arguments about social justice. Who is to decide, and by what authority,
who is to be a member is as fraught with conceptual and ideological baggage
as who is to decide, and by what authority, who is causally affected by a
Shapiro ! ELEMENTS OF DEMOCRATIC JUSTICE 591
particular collective decision. These difficulties should not therefore count

as decisive against the causally based view if the membership-based view is
seen as the alternative. Second, there is considerable experience with causally
based arguments in tort law. Although tort actions are often concerned with
the causal effects of individual rather than collective decisions, in dealing
with them courts have developed mechanisms for determining whose claims
should be heard, for sorting genuine claims from frivolous ones, and for
distinguishing weaker from stronger claims to have been adversely affected
by an action and shaping remedies accordingly. This is not an argument for
turning politics into tort law; the point of the comparison is rather to illustrate
that in other areas of social life institutional mechanisms have been developed
to assess and manage conflicting claims of being causally affected by actions
They may be imperfect mechanisms, but they should be evaluated by refer-
ence to the other imperfect mechanisms of collective decision making that
actually prevail in the world, not by comparison with an ideal that prevails
nowhere.25
Institutionalizing Opposition
Barrington Moore contends that "the existence of a legitimate and, to some

extent effective, opposition" is the defining criterion of democracy. One need
not go all the way with him to be persuaded that institutions fostering "loyal"
opposition are essential to democratic life.26 This is true for several related
reasons. First, opposition institutions perform the functional role of providing
sites for potential alternative leaderships to organize themselves, making possi-
ble periodic turnovers of power that are necessary-though not sufficient-
for democratic governance. Second, opposition institutions help legitimate
democracy by attracting social dissent toward antigovernment forces within
the regime, rather than directing it at the regime's foundations. Anger and
disaffection can thus be directed at particular power holders, without endan-
gering the democratic order's legitimacy. Third, opposition institutions serve
the public interest by ensuring that there are groups and individuals who have
incentives to ask awkward questions, shine light in dark corners, and expose
abuses of power. The importance of these considerations should not be
minimized, but there is a more basic reason why the possibility of effective
opposition is an essential requirement of democratic justice. Unless people
can challenge prevailing norms and rules with the realistic hope of altering
them, the requirement that the inherited past not bind us unalterably would
be empty. The Burkean dimension of democratic justice could not be tem-
pered in the ways that democratic justice requires.
Below I argue that the imperative to make effective opposition pos

leads to three conditioning constraints on the exercise of power. Procedu
it suggests that rules of governance should be deemed unacceptable if
render revision of the status quo impossible. Mechanisms should always
exist, therefore, through which opposition can be articulated. Procedural
guarantees need to be backed up by permissive freedoms of speech, petition,
and association if they are to be effective, but even then they may often be
insufficient to ensure meaningful opposition. Some have suggested that
"substantive" democracy is the appropriate remedy for this malady, and
although I resist this conclusion, I do defend a quasi-substantive constraint
on the exercise of power: that hierarchies should generally be presumed
suspect because of their propensity to atrophy into systems of domination.
The presumption against hierarchy should be rebutted in many circum-
stances, some of which I explore via a series of queries that democratic justice
bids direct at hierarchical arrangements. But the burden of justification
appropriately falls on defenders of particular hierarchies.
Since there are no perfect decision rules, the products of even the best
democratic procedures will leave some justifiably aggrieved. Libertarian
writers draw from this the implication that- it is better to have as little
collective action as possible, but in a world of ubiquitous power relations, I
argued earlier, that option is neither satisfying nor plausible. To point this out
is not to deny that democratic decision rules can lead to the imposition of
outcomes on one group by another, or that the justification for those imposi-
tions classical democratic theorists believed available is not.27 Rather, it leads
to the suggestion that procedures for expressing opposition should be thought
valuable no matter what the prevailing mechanisms of collective decision.
Recognizing that life goes on and decisions have to be made, we should seek
the most appropriate rules of democratic governance for every circumstance.
But people should nonetheless remain free to oppose what has been decided,
and try to change it.
To require that meaningful opposition be tolerated is frequently to require
more than dominant groups want to accept, since it can weaken their control
of collective values and purposes. Accordingly, they often seek to oppose
opposition or render it ineffective. Part of the challenge of democratic justice
is to institutionalize ways to stop them, and it seems safe to assume that, on
their own, procedural guarantees of the freedom to oppose will not secure
this goal. Democratic theorists who value effective opposition, such as
Huntington, have long recognized that unless procedural guarantees are
backed up by permissive civil and political freedoms of speech, press,
assembly, and organization, they are all too easily rendered meaningless.28
The history of sham democracies during the communist era illustrates what
can happen when permissive freedoms are not honored. Democrats would be
unwise ever to think them dispensable.
Although permissive freedoms are often essential to securing the space
for opposition and fostering it, it would be a mistake to conclude that in
general they will be sufficient. To see why this is so, notice that permissive
freedoms can actually undermine the possibility of challenging the status quo,
as my earlier discussion of the proposed Clinton health care reform indicated.
Inequalities in control over the resources needed to transform permissive
freedoms into the service of effective opposition can mean that strategically
powerful groups, when committed to prevailing arrangements, may be able
to block all attempts to alter them. Thus, although permissive freedoms are
reasonably deemed valuable for their propensity to permit and even foster
opposition, they are not a panacea. When those committed to the status quo
have unmatched access to information, wealth, and organizational resources,
they may actually be able to use permissive freedoms to cement their
advantages in place.
Awareness of the combined impact of imperfect decision rules and differ-
ential control over political resources has led some commentators to defend
"substantive" conceptions of democracy over "procedural" ones, usually by
appeal to some variant of Justice Stone's celebrated fourth footnote in United
States v. Carolene Products Co. Noting that well-functioning democratic
processes might lead to the domination of "discrete and insular minorities,"
Stone countenanced the possibility that their operation might reasonably be
limited when this occurs. Stone dealt with a circumscribed class of cases. But
his point is a general one, and others have employed his reasoning more
expansively.29 For instance, John Hart Ely defended much of the judicial
activism of the Warren Court by reference to Carolene Products reasoning.30
Likewise, Charles Beitz has pressed the same considerations into an argu-
ment that the quantitative fairness of equal voting power will never ensure
substantively democratic outcomes.3' In Beitz's view, a truly democratic
system of "qualitative fairness" requires a prior system of "just legislation,"
since mere equal voting power can never be relied upon to produce fair
outcomes.32
Substantively democratic views merely have to be stated for the difficulty
with them to be plain: How can Ely know what democratic processes ought
to have achieved had they not been corrupted by the Carolene problem?
Whence the theory of just legislation against which Beitz will evaluate the
results of voting procedures? Writers like Ely and Beitz have little to say to
those who are unpersuaded by their respective conceptions of "equal concern
and respect" and "qualitative fairness." If, as I maintain, there is no criterion
for justice that is anterior to what democracy generates, this should not be
surprising. To say this is not, however, to respond to the difficulty that

motivates Carolene-type reasoning: There are no perfect decision rules and
those who are better placed to translate permissive freedoms into political
power should be expected, ceteris paribus, to get their way. The problem is
real but the proffered solutions overreach, suggesting the desirability of
finding a middle ground. "More than process, less than substance" might be
an appropriate slogan.
My suggestion is that we stake out the middle ground with the proposition
that hierarchies should be presumed suspect. The reason is that although
hierarchies can exist for many legitimate purposes, by definition they contain
both power inequalities and truncated opportunities for opposition. Power,
as Lord Acton said, tends to corrupt. Even, and perhaps especially, when they
acquire it legitimately, power-holders all too easily convince themselves that
their authority should expand in space and time, that critics are ignorant or
irresponsible, and that subordinates lack the requisite ability to ascend from
inferior roles. The allure of power can thus divert power-holders within
hierarchies from their legitimate goals, leading to the reduction of hierarchies
to their power dimensions. The comparatively limited scope for opposition
within hierarchies makes it difficult to block or check their atrophy into
systems of domination; indeed as atrophy advances the possibilities for
opposition are likely to be increasingly constrained. It is for this reason that
democrats should keep a skeptical eye on all hierarchical arrangements,
placing the burden of justification on their defenders. Power need not be
abused but it often is, and it is wise for democrats to guard against that
possibility.
The preceding observations do not imply that all hierarchies should be
eliminated, even if this could be achieved.33 Rather, they suggest that there
are good grounds for prima facie suspicion of them, even when they result
from democratic collective decisions. Too often, avoidable hierarchies mas-
querade as unavoidable ones, involuntary subordination is shrouded in the
language of agreement, unnecessary hierarchies are held to be essential to the
pursuit of common goals, and fixed hierarchies are cloaked in myths about
their fluidity. Democratic justice suggests mistrust of prevailing hierarchies;
it invites us to look for institutional and other structuring devices to limit them
and to mitigate their unnecessary and corrosive effects. Such devices may be
thought of as contributing to the evolving frameworks of democratic con-
straints within which people should be free to negotiate and renegotiate the
terms of their cooperation and conflict.
To say that hierarchies are suspect is not to say anything about what is to
count as sufficient to rebut the presumption. Nor is it to say anything about
what kinds of constraints on hierarchies should be employed in different
circumstances, or about how these constraints should be enforced. By itself,

the general argument cannot answer these questions. But it does generate a
series of appropriate queries about hierarchies; ways of probing them in the
name of democratic justice.
The first concerns the degree to which a given hierarchy is inevitable.
Consider the differences between adult domestic relations and parent-child
relations. Both have taken a multiplicity of forms, even in the recent history
of the West, yet almost all of these have been explicitly hierarchical in
character. It is evident, however, that parent-child relations are inevitably
hierarchical in ways that adult domestic relations are not. If a relationship is
not inevitably hierarchical, the first question that arises is why should it be
hierarchical at all? There may bejustifiable reasons for a particular inessential
hierarchy (that it is comparatively efficient, that it has been chosen, that the
relevant people like it, or some other), but from the standpoint of democratic
justice the presumption is against hierarchy and the proponent of such reasons
should adopt the burden of persuasion.
When relations are inevitably hierarchical a different class of considera-
tions becomes relevant. We begin by asking: is it necessary that the relations
in question be maintained at all? Parent-child relations of some kind must
exist, but not all inevitably hierarchical relations are of this sort. To consider
a limiting case once again, history has shown that the institution of slavery
need not exist. If an inescapably hierarchical relationship is unnecessary,
it immediately becomes suspect-a kind of surplus hierarchy-from the
standpoint of democratic justice. Slavery thus fares badly from this
standpoint quite apart from its incompatibility with the presumption of uni-
versal inclusion.
A second class of appropriate inquiries about hierarchies concerns their
pertinence to the activity at hand: are the hierarchical relations that exist
appropriately hierarchical? Parent-child relations, for example, may be more
hierarchical than they need to be in many instances and they may include
unnecessary kinds of hierarchical authority. They may also be maintained for
a variety of reasons, ranging from the convenience of parents to desires to
dominate, that have nothing to do with the interests of their charges. Of
alterable hierarchies we should thus always inquire: in whose interests are
they sustained? Those who would sustain a hierarchy of a particular kind, or
sustain it for longer than necessary, take on the burden of establishing that
this operates in the interests of those who are subjected to the relevant
hierarchy. For this reason one would be unmoved by the argument advanced
by Amish parents in the Wisconsin v. Yoder litigation, namely, that they should
be free not to send their teenage children to school because experience had
taught them that this induced in the teenagers the desire to leave the Amish
community, which interfered with their (i.e., the parents') rights of free
religious exercise.34
Democratic justice also bids us to attend to the degree to which hierarchies
are ossified or fluid. We should distinguish self-liquidating hierarchies, as
when children become adults or students graduate, from non-self-liquidating
hierarchies such as caste systems and hierarchies constituted by hereditary
transmissions of wealth and power. We should also distinguish hierarchical
orders in which anyone can in principle ascend to the top from those where
that is not so. No woman can aspire to become Pope, a fact that makes the
Catholic religion less attractive than some others from the standpoint of
democratic justice on this score. In general the argument tells us to prefer
fluid hierarchies over ossified ones, other things being equal. Fluid hierar-
chies may not create permanently subordinated classes whereas ossified
hierarchies will. Of course other things seldom are equal; nonetheless the
requirement is a useful starting point. It tells us what the presumption is and
by whom the burden of persuasion should be carried.
Similarly, asymmetrical hierarchies are questionable whereas symmetri-
cal ones are not necessarily so. Polygamous marriages are generally asym-
metrical, for example, and as such suspect from the standpoint of democratic
justice: a husband can have many wives but a wife cannot have many
husbands.35 If these polygamous regimes were symmetrical, or had their
members practiced "complex marriage" as did the nineteenth-century Oneida
perfectionist community (in which any number of men could marry any
number of women), they would not be questionable by reference to this aspect
of the argument for democratic justice. Again, there may be other reasons
rooted in democratic justice for objecting to such arrangements, but their
symmetry would count in their favor.36
Closely related to questions about the relative fluidity and symmetry of
hierarchies are questions about the degree to which they are imposed. Did
the people who are subjected to them elect to be thus subjected? What were
their other realistic options at the time? Whether or not they chose to enter,
what degree of freedom to exit now exists? Generally nonimposed hierarchies
fare better than imposed ones, and less imposed hierarchies fare better than
more imposed ones. If someone elected to participate at the bottom of a hier-
archical relationship when she had alternative nonhierarchical (or less hier-
archical) options in front of her, the fact of her choice confers some presump-
tive legitimacy on the state of affairs. Analogously, if someone remains in a
hierarchical order when we are fairly confident that she has the resources to
leave, we have less reason to be troubled from the standpoint of democratic
justice than when this is not the case.
Finally, the general argument for democratic justice directs us to attend to

the relative insularity of hierarchies. To what extent do they consist of
self-contained groups of people minding their own business who want to be
left alone by outsiders? Withdrawing sects like the Old Order Amish, or
migrating groups like the Mormons who went to Utah in the nineteenth
century to escape persecution in the east, have at least prima facie valid claims
that they should be able to set the terms of their association unimpeded. Such
groups do not proselytize, or seek to shape the world outside their commu-
nities (as religious fundamentalists, e.g., often do). Hierarchical and undemo-
cratic as these groups might be in their internal organization, they are of little
consequence to the outside world. By contrast, a hierarchical established
church whose influence on outsiders could not be escaped without substantial
cost would not enjoy the same prima facie claim to be left alone. Relatively
insular groups may be objectionable from the standpoint of democratic
justice on some of the other grounds just discussed, but the fact of their
insularity diminishes any extemality-based claim by outsiders to restructure
or abolish them.
SOME DIFFICULTIES CONSIDERED
The preceding elaboration of the two central dimensions of demo

justice is a first installment; as such it raises many questions that it
answer. In the space that remains I will say something about what se
me to be the most important of these, having to do with the complex c
of democratic justice, tensions between it and other goods, and the ap
ate role for the state that is implied by the general argument.
Conflicts Internal to Democratic Justice
Any argument for an internally complex set of principles must co

the possibility that they cannot be satisfied simultaneously. The questio
arises: how are conflicts among the different injunctions to be resolv
response to this question is to come up with a system of meta-rul
resolving conflicts when they occur. For instance Rawls's theory of
consists of a number of principles that, he argues, should be lexically r
In the case of conflicts the principles that are higher in his lexical o
trump those that are lower.37
It is evident that democratic justice exhibits an analogous potent
internal conflict. What the general argument recommends as the app
system of governance in a domain may conflict with the presumptive

cion of hierarchy. People might choose to create a hierarchy by voluntary
action or majority rule. Likewise, the various injunctions against hierarchy
might produce contradictory prescriptions as far as a particular practice is
concerned. The insular character of withdrawing sects, such as Mormons and
the Amish, counsels leaving them alone, yet their internally hierarchical
practices prompt suspicion from the standpoint of democratic justice. How
should conflicts of this kind be resolved?
The two alternatives here are either to try to come up with a system of
meta-principles analogous to Rawls's lexical rules or to supply a principled
defense of a more underdetermined view. To try to come up with a complete
system of meta-principles that would resolve every possible tension that
could arise out of the complexities of the general argument seems to me to
be so demanding a task that it would almost certainly fail.38 The range of
circumstances that can arise is exceedingly large, if not infinite, and the
complexity of the social world is such that there will always be challenges to
the logic internal to democratic justice. This is less troubling than might at
first appear to be the case. For one thing, the lack of a complete system of
meta-principles does not silence democratic justice in every circumstance.
We can still say, for example, that a practice that runs up against a great many
of democratic justice's presumptions is correspondingly more suspect for that
reason. Slavery is an easy case for democratic justice just because this is so.
It violates basic principles of collective self-governance and it is on the wrong
side of every presumption about hierarchy that I discussed: it is unnecessary,
it is not usually entered into voluntarily, it is hard or impossible to escape, it
is both asymmetrical and non-self-liquidating, and it has external effects
that permeate through the social world. By the same token a practice that
turns out to be on the right side of every presumption will be equally easy to
deal with.
The more difficult and interesting cases are those that are less clear-cut.
In many of these instances it may be possible to find accommodations among
conflicting injunctions. For instance, in the case of the Amish one might take
the view that the withdrawing character of the group and the absence of a
threat posed by it to the rest of society counsels against any attempt to
interfere with its existence, but that the state should nonetheless insist that
Amish children be educated so that they have the capacities to function
outside the Amish community in the event that they decide to leave. Thus
while on this view one would not tolerate all Amish educational practices, in
other respects they would be left alone.39 Likewise, although the ossified and
non-self-liquidating hierarchies in the Catholic Church contravene some of
the presumptions of democratic justice, the history of domination that has
accompanied established churches might also counsel that there is wisdom

in an especially wide latitude of tolerance as far as religious matters are
concerned. A government guided by the principles of democratic justice
might nonetheless attach some costs to religions that contravene them, such
as denying tax-exempt status to those in which some offices are reserved for
men, persons of a particular race, or any other group that is defined in a
morally arbitrary way. The governing body of the religion in question would
then be free to decide whether to live with the sanctions in question or to
adjust its practices to avoid themn.'
These examples indicate that once we recognize that there is a range of
possible sanctions and of feasible responses to them, apparently conflicting
imperatives can be managed in a variety of ways that can, over time, be
expected to encourage civil institutions to evolve in comparatively demo-
cratic directions. To some even this approach will sound like opening the way
for dangerously radical interference with freedom of religious worship. But
reflection on our current laws concerning racially exclusionary organizations,
and on the distinctions we comfortably draw between religions and cults and
between education and brainwashing, should reveal that we routinely make
many judgments of this kind, however implicitly. These examples also
underline the fact that when we value more than one commitment we
sometimes have either to live with tensions among them or to come up with
creative solutions to the tensions. This is no less true of the world in which
we actually live than it would be in a world in which democratic justice
furnished the basic principles of governance. The imperatives that follow
from the constituent parts of the United States constitution and its amend-
ments generate many tensions, and just as courts and legislatures have to
order, rank, and accommodate them in particular contexts, so the same would
have to be done in a world governed by considerations derived from the
general argument for democratic justice. Admittedly, to say this is not to
resolve any specific tensions, but it does perhaps indicate the limits of what
it is reasonable to expect from a general statement of principles.
A different objection to the internally complex character of democratic
justice is that it is unnecessarily complex. My claim that freedom to oppose
collective outcomes is not derivative of rights to inclusive participation might
be granted, but, if this freedom is valued regardless of whether decisions were
made democratically, then why value democratic decision-making? Is not
democracy, on my account, reducible to opposition? My answer is that
although inclusive participation and freedom to oppose are valuable inde-
pendently of one another, the ways in which they are exercised are not without
mutual implications. In particular, I propose the following injunction: The
more democratically those who win in battles over collective decisions
conduct themselves in victory, the stronger is the obligation on the de

to ensure that their opposition is loyal rather than disloyal-and vice versa.
Processes of inclusive consultation, meaningful hearings, good-faith consid-
eration of how to mitigate external effects of decisions, and willingness to
consider alternatives all build legitimacy for democratic decision making,
and they should. No less appropriately, their opposites breed cynicism and
mistrust on the part of losers, which erode democracy's legitimacy in predict-
able ways. By linking the obligation to make opposition loyal to how
democratically those in power conduct themselves, protagonists on all sides
are reminded of the imperfection of the rules that give present winners
their victories and losers their losses. In addition, if the two are linked, both
winners and losers have incentives to search for mechanisms that can dimin-
ish the distances between them.
Tradeoffs between Democratic Justice and Other Goods
Additional sources of tension arise from my argument being premised on

the notion that democracy is a conditioning good-subordinate to the activi-
ties whose pursuit it regulates. This means that there can and most likely will
be tensions between the requirements of democratic justice and the activities
it is intended to condition. In the limiting case, there will be activities that
operate in flat contradiction with the principles I have described. Apart from
the case of parent-child relations to which some attention has already been
devoted, there are football teams, armies, and many other organizational
forms whose purposes seem to defy democratic governance. No doubt one
can always challenge the proposition that such organizational forms must
necessarily be undemocratic, and as the rich literature on the governance of
the firm indicates, we should always be open to creative possibilities for the
democratic management of institutions that seem inherently undemocratic.41
Yet one has to confront the possibility that there will be circumstances in
which there are inescapable tradeoffs between democratic control and the
pursuit of a particular good, be it the gathering of military intelligence, the
running of a professional sports team, or other valuable activity.
One response is to deal with tradeoffs of this kind in the same way that
tensions internal to democratic justice are handled, by recognizing that when
there is more than one thing we value at times we will have to choose among
them. But this should be the last response, not the first. Although there is
never a guarantee that tradeoffs between democratic justice and other goods
can be avoided, the argument for democratic justice bids us to try to find ways
to avoid them. Consider the two examples just mentioned. Congress has
devised oversight mechanisms that, however imperfectly, ensure some demo

cratic accountability of intelligence agencies consistent with their secret
purposes. No doubt we pay a price for them and they could be improved upon,
but the outcome of the Cold War scarcely suggests that our system fared
worse than the Soviet system, in which there was virtually no democratic
accountability of any sort, or indeed that it fared less well than other systems
in the West that until recently have had little or no democratic oversight.42 As
far as professional sports are concerned, there too the situation is less clear-cut
than might at first sight appear to be the case. Although one would not want
everyone on the team calling plays, there are many areas in professional
sports where a measure of democratic control can be achieved without
compromise of athletic purpose. Pay and working conditions are the most
obvious areas; no doubt there are others. To reiterate, the general point is that
the presumption is against undemocratic ways of doing things. It is only a
presumption and it can be overcome, but reasons should be demanded and
the burden of persuasion should always lie with those who would limit
democracy's operation.
Competing Demands of Different Domains
Yet other potential tensions arise, for democratic justice, out of the fact
that it is simultaneously concerned with many domains of civil society. It may
be the case that pursuing democratic justice in one domain makes it more
difficult, perhaps even impossible, to pursue it in other domains. For instance,
participating in governance is part of what democratic justice requires. Yet
there are limits to how much time people have available, so that increased
participatory involvement in one domain may mean diminished participation
in others. This is what Carmen Sirianni has characterized as the "paradox of
participatory pluralism." It arises for anyone who both values democratic
participation and embraces a view of politics that ranges throughout civil
society. We cannot simultaneously maximize participation over all domains.43
The paradox is inescapable for participatory democrats like Sirianni (who
offers no solution to it), but the argument for democratic justice suggests
avenues for dealing with it. Participation is not valuable for its own sake on
my view; rather, it is valuable only as it is pursued in conjunction with the
goods that it conditions. Collective self-governance is important in every
domain of civil society but it is never the most important thing; adherents of
democratic justice should thus always be open to timesaving and other novel
devices to conserve participatory resources. For instance, since the 1970s a
number of writers have explored the use of "deliberative polls": randomly
selected groups that are paid to debate public issues from the selectio
presidential candidates to the governance of school districts.44 Exp
with deliberative polls suggests that they may provide useful mechanis
both exerting democratic control and solving the difficulty, pointed to
Sartori and others, that "knowledge-cognitive competence and control-
becomes more and more the problem as politics becomes more and more
complicated."45 Randomly selected lay groups, that have no particular vested
interest in the outcome in a given area, can invest the time and energy needed
to make informed decisions. Such groups can gather data and listen to expert
witnesses, making use of esoteric knowledge without being held hostage to
it. The decisions they render could be advisory or even binding, at least for
certain matters. From the standpoint of democratic justice the possibilities
offered by deliberative polls are worth exploring because they provide a
potential way out of Sirianni's paradox: they combine citizen control with
the possibility of sophisticated decision making in a complex world, and they
do it in a way that takes account of the economy of time.46
Earlier I suggested that from the standpoint of democratic justice partici-
pation should be seen neither in purely instrumental terms nor as the point of
the exercise in politics. Such devices as the deliberative poll are attractive
because they are an example of a creative institutional response to the goal
of trying to occupy a middle ground between these two views. Everyone
might be expected to participate in some deliberative polls, just as everyone
is expected to sit on some conventional juries. Everyone would know that,
in the bodies they are not involved in, other randomly selected bodies were
sitting with no particular agendas or interest groups of their own that were
being advanced. Everyone would also know that no matter how complex and
technical decisions were becoming, a meaningful element of lay control
would nonetheless be present in all collective decision making. This is
essential for democratic justice.
THE ROLE FOR THE STATE
Apart from the extremes of limiting cases like slavery, the general a
ment for democratic justice does not provide conclusive assessments of
particular decision rules or mechanisms of opposition. Instead, as we have
seen, it generates presumptions and distributes burdens of persuasion in
various ways. That is as it should be. Since the general argument is a
semicontextual, particularities of context are needed to decide when burdens
have been appropriately carried and when presumptions have been rebut-
ted. Democratic justice generates determinate conclusions in particular con-

texts only.
To say this is not, however, to deliver on everything that should reasonably
be expected of the general argument. Invoking the language of presumptions
and burdens of persuasion immediately raises the question: who is to judge
when burdens have been carried and presumptions rebutted? Since the
evidence will often be inconclusive and opinion about it must be expected
often to be divided, just where decision-making authority should be located
is, and is bound to remain, an important general question. The answer is also
partly contextual; different authorities are appropriate in different kinds of
circumstances. But the answer is only partly contextual; from the standpoint
of democratic justice, some general considerations apply.
Antivanguardism and Its Limits
Whenever anyone claims to know how to get to democracy undemocrati-

cally, skepticism is in order for two reasons, one practical and one normative.
The practical reason is that it is doubtful that they can know that they are
right. Just because democratic reforms are typically reactive responses to
particular evils that chart new courses into the future, it is usually difficult to
know what their full consequences will be or what new problems the reforms
will create. For instance, changes in the structure of American family law,
making marriage more of a contract and less of a status, have been motivated
by a desire to undermine the patriarchal family. But it has since become
evident that one of the net effects of these changes has been to render women
increasingly vulnerable to the greater economic power of men in marriage.47
As this becomes apparent, other ways of democratizing family life will be
sought and new experiments tried and modified as and when the obstacles
they generate come into view. Democratizing family life will likely require
changes in the organization of the economy, and perhaps other changes that
are yet to be thought of. Likewise with the debate on democracy in the work-
place, there is now considerable disagreement on which are more effective
in undermining alienating hierarchy: strategies of worker self-management
or plans for employee ownership or part-ownership that leave the structure
of management comparatively untouched. Many varieties of both have been
tried in different industries. It seems certain that no single model will turn
out to be generally applicable and that new possibilities are yet to be tried.48
To take an example from the realm of institutional governance, during the
nineteenth century reasonable salaries and working conditions for politicians
were rightly seen as essential to undermining a system in which government
was a part-time activity for the wealthy. But these improvements have
brought with them new brands of ossified power in the form of professional
politicians with lifetime career aspirations in government. In the United
States, electoral politics have become dependent on money to such a degree
that political elites manage to maintain themselves in positions of power for
life in ways that are at odds with democracy's hostility toward entrenched
hierarchy.49 As a response to this, new democratic reforms are being called
for, geared toward limiting the number of terms politicians can serve and
better regulating the role of money in electoral politics.50
It defies credulity to suppose that in any of these instances democratic
reformers could have understood social processes profoundly enough, or
seen sufficiently far into the future, to have anticipated all the problems and
possibilities that lay ahead. Yet these cases are not exceptional; life has more
imagination than we do, and it will often defeat our best efforts and present
unexpected obstacles and opportunities. The fabric of social life and the
dynamics of historical change are complex and little understood; that is the
reality with which we have to live. Designing democratic institutional con-
straints is thus bound to be a pragmatic business, best pursued in context-sen-
sitive and incremental ways. New activities come into being, technological
change, experience, and the evolution of other causally linked activities all
present fresh problems and generate novel possibilities for democratic gov-
ernance. There are good reasons to be skeptical of anyone who denies this,
whether they harbor a hidden agenda that is being obscured by their van-
guardist pretensions or they are acting out of misplaced faith in their own
prescient abilities.
Means/ends dichotomies are suspect, also, for the normative reason that
they undermine the spirit of democratic justice. Although I have argued that
we should resist the participatory democrat's contention that participation is
valuable for its own sake, we should be no less wary of purely instrumental
conceptions of democracy. Democratic means are never the point of the
exercise, but they are usually of more than mere instrumental value. There is
value to doing things democratically, and there is value to struggling with
how to do things democratically while still achieving one's other goals.
Democratic habits of self-restraint and attention to the needs and aspirations
of others have to be learned through democratic practice; succumbing to the
authoritarianism inherent in means/ends dichotomies should be expected to
undermine it. In this connection, Dewey penned the right maxim for demo-
cratic justice over half a century ago: "Our first defense is to realize that
democracy can be served only by the slow day by day adoption and conta-
gious diffusion in every phase of our common life of methods that are
identical with the ends to be reached."51
The principled refusal to impose solutions from above can provoke the
argument that unless this is done they will not be implemented at all, an
there are, indeed, at least three important classes of exceptions to the in
presumption against vanguardism. The first concerns the provision of pu
goods. As my earlier discussion of education revealed, when the provisio
of public goods is at issue and there are differential capacities for exit, n
local decision rule will likely be effective in diminishing injustice. This
amounts to saying that effective policies will have to be imposed from
above.52 Proponents of "shock therapy" in the transition from communism to
capitalism seem often to take an analogous view. Adam Przeworski argues,
for example, that during transitions from authoritarianism to democracy
unless economic reforms are rammed through from above, those who are
adversely affected by them will mobilize their opposition to them through
the democratic process, scuttling the reforms. Consequently, fledgling demo-
cratic governments "face the choice of either involving a broad range of
political forces in the shaping of reforms, thus compromising their economic
soundness, or trying to undermine all opposition to the [reform] program."
In Przeworski's view, any government "that is resolute must proceed in spite
of the clamor of voices that call for softening or slowing down the reform
program." Because reformers "know what is good" all political conflicts
become no more than a waste of time. Przeworski goes on to point out that
every instance of successful market reform during democratic transitions on
record was implemented by executive decree, remarking that "[t]his potential
is inherent in the very conception of market reforms."53
From the standpoint of democratic justice, the critical question is whether
the reformers really do "know what is good" and pursue it in fact. Much of
what is presented by economic reformers as uncontroversially good might in
fact be controversial, and many economic reforms that are described as public
goods do not meet the technical criteria that require both joint supply and
nonexcludability.54 Whether the sorts of privatization and stabilization poli-
cies that such political economists as Przeworski, Janos Kornai, Jeffrey
Sachs, and others advocate lead to the supply of public goods in this sense is
debatable. No doubt parts of what is provided are public goods, but other
aspects of these policies may amount to little more than mechanisms for
raiding public treasuries by strategically well-placed groups, generating little
or no benefit for anyone else. In such instances, the pursuit of private benefit
may be cloaked in the language of public goods, and opposition to them that
is really a reflection of zero-sum distributive conflict will masquerade as a
collective action problem. What are billed as solutions to it will actually be
partisan policies that help some sectors and hurt others. Democrats who
suspect this is the case with substantial parts of postcommunist privatizations
are bound to find themselves ambivalent, at least, about the "bitter pil
strategies that depend on "initial brutality, on proceeding as quickly as
possible with the most radical measures," and implementing reforms either
by administrative fiat or ramming them through legislatures.55
In circumstances where one does not doubt that a public good is being
supplied, one's democratic moral intuitions are not troubled by decisive
action from above. For instance in the South African constitutional negotia-
tions that led up to the April 1994 elections, it gradually became clear that-
desirable as multiparty roundtable negotiations sounded-they were not
going to produce an agreement on a democratic constitution. Too many groups
had too many incentives to pursue private agendas at the expense of ensuring
that the public good was provided. Consequently, it became evident that, if
a democratic political order was to be put in place, it would have to be
hammered out as an elite pact and then imposed on the society. This is what
transpired in fact, and the reason that democrats the world over applauded as
opponents to the transition were so effectively either marginalized or co-opted
was that almost no one doubted that what the elites proposed to impose-a
democratic constitutional order-was in fact a public good.56
Distinguishing the provision of genuine public goods from spurious ones
is a difficult and controversial business. Often the two will be mixed, making
it even more difficult, as is almost certainly the case with most privatization
plans. Even in the case of the South African constitution it seems clear that
the elites who committed themselves to providing the public good in question
sprinkled in a few benefits for themselves, notably a system of electoral and
parliamentary rules that greatly weakens backbenchers vis-a-vis leaderships,
as well as bribes to particular interest groups to insulate them from the new
political order.57
From the standpoint of democratic justice, the extent to which policies
may legitimately be imposed from above varies with the degree to which
genuine public goods are being provided. As the preceding remarks indicate,
this will often be hotly disputed and ideologically charged, not least because
there will be those who have an interest in obscuring the matter. It may also
be genuinely unclear in certain circumstances. When either of these things is
the case, what we are witnessing is not a failure in the argument for
democratic justice. Rather, it is a failure in understanding of, or agreement
about, whether or not something constitutes a public good. This is not to
diminish the normative importance of the matter, it is only to say that it would
be to expect the wrong kind of thing from any political theory to ask of it that
it resolve contentious empirical questions of political economy. The argument
for democratic justice can be expected to counsel what to do when a certain
fact pattern obtains; it cannot be expected to tell us whether or not the fact
pattern really does obtain. The general argument does, however, counsel us
to regard claims to be providing public goods with suspicion, and to subject
them to what lawyers think of as "strict scrutiny." American courts typically
subject legislative action to this most demanding level of constitutional
scrutiny when the proposed action interferes with a "fundamental" liberty,
usually a freedom protected by the Bill of Rights. Strict scrutiny requires a
showing that the governmental objective is unusually important-that a
"compelling" state interest is at stake-and that it cannot be accomplished in
a less intrusive way.58 By analogy we might say that the undemocratic
imposition of a public good is justified only when the good in question is
essential to the operation of a democratic order and cannot be attained in any
other way. Because those who claim to provide public goods may have
ulterior motives, and because private goods can often masquerade as public
goods, the strong presumption should always be against their imposition
from above.
A second class of exceptions to the general presumption against vanguard-
ism arises when illegitimate hierarchies have been maintained by the state.
For example, in the West the disadvantaged status of women in family life
was sustained.by the common law and other active policies of the state for
centuries. One dramatic legacy of this history is that as recently as the 1950s
throughout the United States a husband could not be prosecuted for raping
his wife. By the mid- 1990s spousal rape was a prosecutable felony during an
ongoing marriage in well over a third of American jurisdictions, the product
of a concerted feminist campaign in state legislatures and courts.59 It would
have been impossible for such changes to have come about without the state's
active involvement, since it is the policies of the state that were at the root of
the injustice in question. Likewise, it took the passage of the married women's
property acts (the first wave of which began in the 1840s) to destroy the
common law rule that had given the husband control, and sometimes title, to
the wife's property and possessions during marriage.'
In such circumstances it will be necessary, and justifiable from the
standpoint of democratic justice, for the state to be centrally involved in
dismantling the unjust system it has created. Women would have been
morally misguided as well as politically shortsighted had they not sought to
enlist public institutions in this struggle to refashion the terms of their
domestic association. Since the unjust hierarchies to which they had been
subjected were direct products of state policies and sustained by the legal
order, it was reasonable to require the state to play an active role in disman-
tling the injustices in question. Likewise, the effects of the Group Areas Act
in South Africa, which led to the forced removals of millions of blacks from
viable communities to desolate deserts, are properly responded to by remedial
action from a democratic South African state.61 The general point here
the more antidemocratic practices have been underwritten by the sta
more powerful is the case for the involvement of state institutions in re
dying the unjust status quo.62
A third class of exceptions arises when domination within a domain is
a direct product of state action, but it is nonetheless sustained by for
external to that domain that can be removed only by state action. This i
Michael Walzer has described as "dominance," the transfer of power
domain of social life where it may be legitimate into another where it i
Walzer contends, for instance, that economic inequality is not objecti
as such, and that it may be justified in the sphere of production for its inc
and other efficiency effects. What is objectionable is that disparities in i
and wealth are all too easily translatable into disparities in the political
domain, the domestic domain, the educational domain, and other areas where
they have no evident rationale.63 This happens because the resources neces-
sary to exercise power tend to be fungible across domains, and Walzer sees
it as one of the appropriate tasks of a democratic state to limit this fungibility.
On this view, laws against buying and selling votes for money can be
defended, for example, even though such laws are inefficient in the econo-
mist's sense. Similarly, refusals by courts to enforce antenuptial agreements
that leave divorcing spouses destitute amount to a refusal by the state to allow
economic disparities that may be justifiable outside the domestic domain to
set the terms of life within it.
From the standpoint of democratic justice Walzer's intuition about this
class of cases is defensible, if for different reasons than those that he supplies.
Whereas for Walzer the reason for trying to prevent domination within a
sphere by those who control goods external to it is rooted in shared meanings
about which goods are appropriate in which domains, from the present
standpoint the justification is rooted in considerations drawn from the politi-
cal economy of power.64 I said earlier that the shape of decision rules should
follow the contours of power relations, not those of memberships. It follows
that when obstacles to democracy within a domain are externally sustained,
it is an appropriate use of state power that it be used to remove such obstacles.
To deny this would amount to abandoning democratic justice in particular
domains to those who have imperial control of fungible resources. In short-
pace Walzer-because causal effects rather than shared membership within
a domain are decisive in legitimating a right to democratic control, it follows
that state action that crosses the boundaries of domains can be justified when
this is necessary to achieve democratic justice within a domain.65
Action by the state to advance democratic reform can be justified, then,
but not as part of any missionary quest on democratic justice's behalf. There
is no secular analogue to "Christianizing the infidels" to justify such

whether by courts, legislatures, or invading armies. Rather, external in
ment can be justified by three principal classes of reasons. First, when
provision of a public good is at stake, imposed solutions may be justifiable,
subject to the caveats I have mentioned. This we might think of as a market-
failure justification. Second, the state may often have an affirmative obliga-
tion to help foster democracy flowing from its historical culpability in creating
and sustaining injustice. Last, when external sources of domination within a
domain can be removed only by state action, this can be justified by reference
to the argument from causal legitimacy.
Legislatures versus Courts
The aspiration to avoid imposed solutions suggests that the presumption

should generally be in favor of doing things through representative institu-
tions rather than courts or other agencies, for the conventional reason that
legislatures are comparatively more democratically accountable. There will
be exceptions to this, but it is the exceptions that stand in need ofjustification.
In this connection the argument for democratic justice exhibits an elective
affinity with the approaches to constitutional adjudication that have been
defended in recent years by Ruth Bader Ginsburg and Robert Burt, and it will
be useful to end with some discussion of their views.
Burt conceives of a constitutional democracy as inescapably committed
to two principles-majority rule and equal self-determination-that have the
potential to conflict with one another. If majoritarian processes are employed
to promote domination of some by others, the contradiction latent in demo-
cratic politics becomes manifest. In such circumstances democracy goes to
war with itself and an institutional mechanism is needed to resolve the
conflict. This is supplied, on Burt's account, by judicial review, understood
as "a coercive instrument extrinsic to the disputants" in a political struggle.
Burt sees judicial review as a "logical response to an internal contradiction
between majority rule and equal self-determination. It is not a deviation from
that theory."66
If the court's legitimate role in a democracy is rooted in this logic of
preventing domination through democratic process, then it follows on Burt's
view that its activities should be limited to dealing with the consequences
of the democratic contradiction. And since preventing domination is the
goal, it also follows that courts should not take up sides in disputes that are
by-products of the democratic contradiction (effectively imposing the wishes
of one group on another). Rather, they should declare the domination that has
610 POLITICAL THEORY/ November 1996
emerged from the democratic process unacceptable and insist that the parties
try anew to find an accommodation. Thus in contrast to what many have seen
as the altogether too timid approach taken by the U.S. Supreme Court in the
school desegregation cases of the 1950s and after, on Burt's view the Court
took the right stand. In Brown v. Board of Education the justices declared the
doctrine of "separate but equal" to be an unconstitutional violation of the
equal protection clause, but they did not describe schooling conditions that
would be acceptable.67 Rather, they turned the problem back to Southern state
legislatures, requiring them to fashion acceptable remedies themselves.68
These remedies came before the Court as a result of subsequent litigation,
were evaluated when they did, and were often found to be wanting.69 But the
Court avoided designing the remedy itself, and with it the charge that it was
usurping the legislative function.
Ginsburg, too, has made the case that when courts try to step beyond a
reactive role they undermine their legitimacy in a democracy. Although she
thinks that it is sometimes necessary for the court to step "ahead" of the
political process to achieve reforms that the Constitution requires, if it gets
too far ahead it can produce a backlash and provoke charges that it is
overreaching its appropriate place in a democratic constitutional order.70 She
and Burt both think that the sort of approach adopted by Justice Blackmun
in Roe v. Wade exemplifies this danger.7' In contrast to the Brown approach
in Roe the Court did a good deal more than strike down a Texas abortion
statute. The majority opinion laid out a detailed test to determine the condi-
tions under which any abortion statute could be expected to pass muster. In
effect, Justice Blackmun authored a federal abortion statute of his own. As
Ginsburg put it, the court "invited no dialogue with legislators. Instead, it
seemed entirely to remove the ball from the legislators' court" by wiping out
virtually every form of abortion regulation then in existence.72
On the Ginsburg-Burt view, the sweeping holding in Roe diminished the
Court's democratic legitimacy at the same time as it polarized opinion about
abortion and put paid to various schemes to liberalize abortion laws that were
under way in different states. Between 1967 and 1973 statutes were passed
in nineteen states liberalizing the permissible grounds for abortion. Many
feminists had been dissatisfied with the pace and extent of this reform. This
is why they mounted the campaign that resulted in Roe. Burt concedes that
in 1973 it was "not clear whether the recently enacted state laws signified the
beginning of a national trend toward abolishing all abortion restrictions or
even whether in the so-called liberalized states, the new enactments would
significantly increase access to abortion for anyone." Nonetheless, he points
out that "the abortion issue was openly, avidly, controverted in a substantial
number of public forums, and unlike the regimen extant as recently as 1967,
it was no longer clear who was winning the battle."73 Following the Brown
model, the Court might have struck down the Texas abortion statute in Roe
and remanded the matter for further action at the state level, thereby setting
limits on what legislatures might do in the matter of regulating abortion
without involving the Court directly in designing that regulation. On the
Ginsburg-Burt view, this would have left space for democratic resolution of
the conflict, ensuring the survival of the right to abortion while at the same
time preserving the legitimacy of the Court's role in a democracy.74
Although the tensions that arise within democratic justice differ from
those that motivate Burt and Ginsburg, in three important respects their view
of the appropriate role for courts in a democratic order fits comfortably within
the general argument developed here. First, they articulate an appropriate
institutional response to the injunction that rather than impose democracy on
collective activities the goal should be to try to structure things so that people
will find ways to democratize things for themselves. By placing courts in a
nay-saying stance of ruling out practices as unacceptable when they violate
the strictures of democratic justice, courts can force legislatures and the
conflicting parties they represent to seek creative solutions to their conflicts
that can pass constitutional muster. Second, the Ginsburg-Burt view is
attractive because it is reactive but directed; it exemplifies the creative
pragmatism that motivates democratic justice. It involves accepting that there
is an important-if circumscribed role for courts in a democracy, yet it does
not make the unmanageable administrative demands on courts that accom-
pany more proactive views of adjudication. On this view a court might
reasonably hold that a given policy should be rejected without stating (indeed,
perhaps without having decided) what policy would pass muster. "This is
unacceptable for reasons a, b, c . . . ; find a better way" is seen as an
appropriate stance for a constitutional court. Finally, by recognizing the
relatively greater legitimacy of legislatures and treating courts as institutional
mechanisms for coping with legislative failure, the Ginsburg-Burt view takes
account of the fact that no decision-making mechanism is flawless. Yet it does
so in a way that is rooted in the idea that democratic procedures should be
made to operate as well as possible, and, when they fail, remedies should be
no more intrusive on the democratic process than is necessary to repair them.
Some will object to this as too minimal a role for reviewing courts, but
democrats have to concern themselves not only with courts that aspire to
advance the cause of democratic justice, as they might reasonably be thought
to have done in Brown and Roe, but also with courts that do not, as was the
case in Dred Scott, the Civil Rights Cases, and Lochner v. New York.7s
Insulated from any further review and lacing, at least in the American context,
in democratic accountability, courts can put decisions of this kind in place
that may not be reversed for decades or even generations. Although it may
thus be wise from the standpoint of democratic justice to embrace an activist
role for a constitutional court, it is equally wise to limit courts to a circum-
scribed and negationist activism.
CONCLUDING REMARKS
My aim in this essay has been to render plausible the case for a demo
conception of social justice. This I have sought to do by building on the
popular view, in which considerations of democracy and justice are inti-
mately linked, rather than conventional academic views of them as funda-
mentally distinct and mutually antagonistic. The account that I offer rests on
the twin commitments to government and opposition in democratic theory,
suggesting that there should always be opportunities for those affected by the
operation of a collective practice both to participate in its governance and to
oppose its results when they are so inclined. These two injunctions should
reasonably be expected to have different implications in different cultures
and, within the same culture, to evolve over time and play themselves out
differently in different domains. They are best thought of as conditioning
constraints, designed to democratize social relations as they are reproduced,
rather than as blueprints for social justice.
This view contains internal tensions, to be sure, but I have tried to show
that these come with the territory in reflecting about the justness of social
arrangements, and to indicate something about how these tensions might best
be coped with consistent with the spirit of the general argument. Beyond this,
I have sought to indicate the main outlines of a view of the state that follows
from my view, and to develop some of its implications for the provision of
public goods and for the state's appropriate role in advancing democratic
justice more generally. I have sketched the basic principles that should guide
state action, as well as the fitting nature and place for judicial review in the
argument for democratic justice. No doubt these arguments raise as many
questions as they settle, but I hope, nonetheless, that I have characterized
the central argument and its motivation sufficiently fully to cast it in an
attractive light.
In 1918 Dewey remarked that any philosophy animated by the striving to
achieve democracy "will construe liberty as meaning a universe in which
there is real uncertainty and contingency, a world which is not all in, and
never will be, a world which in some respects is incomplete and in the making,
and which in these respects may be made this way or that according as men
Shapiro ! ELEMENTS OF DEMOCRATIC JUSTICE 613
judge, prize, love and labor. To such a philosophy any notion of a perfec
complete reality, finished, existing always the same without regard to t
vicissitudes of time, will be abhorrent."76 Democratic justice is conceive
in a similar contingent and pragmatic spirit. Just as there are no blueprin
there are no final destinations. Social practices evolve, as do technologies of
government and opposition, often presenting fresh injustices and novel
possibilities for dealing with them. The challenge is to confront the injustices
and take advantage of the possibilities in a principled and satisfying way.
Democratic justice is intended to help in that endeavor.
NOTES
1. John Dewey, The Public and Its Problems (New York: Henry Holt, 1927), 84-5.
2. See, for example, Guiseppe Di Palma, To Craft Democracies: An Essay on Democra
Transitions (Berkeley: University of California Press, 1990), 23, for defense of the view that th
democratic ideal should be disengaged "from the idea of social progress" if it is to endure, a
Samuel P. Huntington, The Third Wave: Democratization in the Late Twentieth Century (N
man: University of Oklahoma Press, 1991), 165-9, for the argument that political leaders w
sell out on their constituents' demands for social justice are more likely to succeed in conso
dating democratic institutions than those who do not.
3. See my "Three Ways to Be a Democrat," Political Theory 22 (February 1994): 124-
4. I should say at the outset that by using the term foundational I do not mean to take up a
position here in debates about the nature of knowledge and existence. In describing a comm
ment to democracy as foundational, I mean only to suggest that no prior or more basic polit
commitment rightly commands our common allegiance.
5. John Rawls, "Faimess to Goodness," Philosophical Review 82 (1973): 228, and "Justic
as Fairness: Political not Metaphysical," Philosophy and PublicAffairs 14, no. 3 (1985): 223-
esp. 223-6.
6. See Cass Sunstein, "On Legal Theory and Legal Practice," in Nomos XXXVII: Theory
and Practice, ed. Ian Shapiro and Judith Wagner DeCew (New York: New York University Press,
1995), 267-87.
7. Barrington Moore, Jr., Liberal Prospects under Soviet Socialism: A Comparative
Historical Perspective (New York: The Averell Harriman Institute, 1989), 25.
8. Nelson Mandela, "Address to Court before Sentencing," in Ideologies of Liberation in
BlackAfrica 1856-1970, ed. J. Ayo Langley (London: Rex Collins, 1979), 665. On the Chartists,
see Dorothy Thompson, The Chartists (London: Temple Smith, 1984). Generally see Elie
Halevy, The Growth of Philosophic Radicalism (New York: Kelley, 1972).
9. Langley, Ideologies of Liberation, 664.
10. Nor do all liberals agree with one another, any more than all democrats do, over which
of these, combinations of these, or combinations of these and other, reasons they invoke for
adhering to antivanguardist conceptions of the good.
11. For one conventional statement of this view, see William Riker, Liberalism against
Populism (Prospect Heights, IL: Waveland, 1982). It might be objected that the depiction of
liberalism in the text is something of a caricature in that it deals with academic rather than popular
conceptions, and only a subset of academic conceptions of liberalism at that. This is co

though I would contend that it is an expansive subset, ranging at least from the fears of m
tyranny expressed through the state that can be found in Mill's On Liberty (1859; re
Indianapolis, IN: Hackett, 1978) and de Tocqueville's Democracy in America (1832; reprint,
Garden City, NY: Doubleday, 1969), through the libertarian liberalisms of Riker, Nozick, and
Buchanan and Tullock discussed in the text, and the nonlibertarianism antistatism embraced by
Judith N. Shklar, Ordinary Vices (Cambridge, MA: Harvard University Press, 1984).
12. Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), 4.
13. For extensive discussion of this question, see my The Evolution of Rights in Liberal
Theory (New York: Cambridge University Press, 1986), chaps. 4-6, and "Three Fallacies
Concerning Majorities, Minorities, and Democratic Politics," in Nomos XXXII: Majorities and
Minorities, ed. John Chapman and Alan Wertheimer (New York: New York University Press,
1990), 79-125.
14. See Michael Walzer, Spheres of Justice (New York: Basic Books, 1983), 3-20, and
Alasdair Maclntyre, After Virtue, 2d ed. (Notre Dame, IN: University of Notre Dame Press,
1984), 181-203.
15. James Buchanan and Gordon Tullock, The Calculus of Consent: Logical Foundations
of Constitutional Democracy (Ann Arbor, MI: Ann Arbor Paperbacks, 1962), 77, 73-6.
16. Brian Barry, PoliticalArgument 2d ed., (1965; reprint, Herefordshire: HarvesterWheat-
sheaf, 1990); Douglas W. Rae, "Decision-Rules and Individual Values in Constitutional Choice,"
American Political Science Review 63, no. 1 (1969): 40-56, 51; Michael Taylor, "Proof of a
Theorem on Majority Rule," Behavioral Science 14 (May 1969): 228-3 1. When the number of
voters is odd, the optimal decision rule is majority rule, n over two, plus one-half; when n is
even, the optimal decision rule is either majority rule (n over two plus one), or majority rule
minus one (simply n over two). Generally, see Dennis C. Mueller, Public Choice II (New York:
Cambridge University Press, 1989), 96-111.
17. Even if we accept the contractualist metaphor, the logic of Buchanan's and Tullock's
defense of unanimity rule can be shown to break down once time and externalities are taken into
account. See Douglas W. Rae, "The Limits of Consensual Decision," American Political Science
Review 69 (1975): 1270-94.
18. It might appear that no-fault divorce destroys the marriage contract qua contract entirely,
since it is terminable at the will of either party. But such a conclusion (i) ignores the fact that
conventional unanimity rule operates in marriages unless and until they reach the point of
dissolution, and (ii) conflates the grounds for divorce with the terms of divorce (and in particular
the distribution of costs that courts will impose on divorcing parties). In fact many countries,
and some American states, that embrace some form of no-fault divorce do not go all the way
with it. Instead they insist that the court find that "irretrievable breakdown" has occurred, for
which purpose the judge may take various factors including the wishes of both parties into
account. See Mary Ann Glendon, Abortion and Divorce in Western Law (Cambridge, MA:
Harvard University Press, 1987), 64-81.
19. See Robert Dahl, A Preface to Economic Democracy (Berkeley: University of California
Press, 1985), 111-35.
20. See Elaine Spitz, Majority Rule (Chatham, NJ: Chatham House, 1984), 135-215.
21. See my "Three Ways to Be a Democrat," 142-4.
22. This is discussed at length in my Political Criticism (Berkeley: University of California
Press, 1990), 252-61.
23. The $50 million figure is reported by Tim Rinne, "The Rise and Fall of Single-Payer
Health Care in Nebraska," Action for Universal Health Care 3, no. 10 (May 1995): 4-5. See also
Tom Hamburger and Ted Marmor, "Dead on Arrival: Why Washington's Power Elites
Consider Single Payer Health Reform," The Washington Monthly (September 1993): 27
24. Liberals take the basic unit of the nation state for granted, treating it as a kind of
voluntary association writ large, as has often been pointed out in criticism of Rawls. S
Rawls, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971), 371-82,
"The Law of Peoples," Critical Inquiry 20 (Autumn 1993): 36-68. No doubt this is often a
consequence of the liberal proclivity for thinking in contractualist terms. For an illustration of
the communitarian view of membership as the basic trumping good, see Walzer, Spheres of
Justice, 29, 31-63.
25. My contention that the causally based view is more defensible than the going alternatives
is compatible with a number of recent arguments whose purpose is to decenter membership-
based sovereignty as the decisive determinant of participation, and to replace it with systems of
overlapping jurisdiction in which different groups of persons are seen as sovereign over different
classes of decisions. See Thomas Pogge, "Cosmopolitanism and Sovereignty," Ethics 103
(October 1992): 48-75; Alexander Wendt, "Collective Identity-Formation and the International
State," American Political Science Review 88, no. 2 (June 1994): 384-96; and William Antholis,
"Liberal Democratic Theory and the Transformation of Sovereignty" (Ph.D. diss., Yale Univer-
sity, 1993).
26. Barrington Moore, Jr., Liberal Prospects, 8. See also Archibald S. Foord, His Majesty's
Opposition 1714-1830 (Oxford: Oxford University Press, 1964).
27. Even before the advent of the modern literature on public choice, Schumpeter had
exposed the logical flaws in the Rousseauist idea of a general will, concluding that "though a
common will or public opinion of some sort may still be said to emerge from the infinitely
complex jumble of individual and group-wise situation, volitions, influences, actions and
reactions of the 'democratic process,' the result lacks not only rational unity but also rational
sanction." Joseph Schumpeter, Capitalism, Socialism and Democracy (New York: Harper and
Row, 1942), 253.
28. Huntington regards at least two turnovers of power following elections as necessary for
a country's being democratic. See The Third Wave, 6-7. See also Robert Dahl, Polyarchy:
Participation and Opposition (New Haven, CT: Yale University Press, 1971).
29. Stone focused his attention on "statutes directed at particular religious ... or racial
minorities" and circumstances that tend "seriously to curtail the operation of those political
processes ordinarily to be relied upon to protect minorities," United States v. Carolene Products
Co., 304 U.S. 144 (1938), 152 n. 4.
30. John Hart Ely, Democracy and Distrust (Cambridge, MA: Harvard University Press,
1980). Ely described his argument as purely procedural, designed to repair defects of democratic
process. But as critics have pointed out and the discussion below makes clear, it is obviously a
substantive argument. See R. Smith, Liberalism and American Constitutional Law (Cambridge,
MA: Harvard University Press, 1985), 89-91, 170-4.
31. Charles Beitz, "Equal Opportunity in Political Representation," in Equal Opportunity,
ed. Norman E. Bowie (Boulder, CO: Westview, 1988), 155-74. It should not be thought that
Carolene Products logic is the exclusive preserve of the political left. For instance, Riker and
Weingast employ it to criticize taxation of property: "What protection is there against members
of today's majority from providing private, redistributive benefits to themselves under the guise
of public purposes and at the expense of some minority of owners and the efficiency of
production? Why is the abridgement of a minority's economic rights less troubling than an
abridgement of the same minority's political rights?" William H. Riker and Barry R. Weingast,
"Constitutional Regulation of Legislative Choice: The Political Consequences of Judicial
Deference to Legislatures," The Hoover Institution, Stanford University, Working Pape

(December 1986), 6.
32. Beitz, "Equal Opportunity," 168.
33. The quixotic political commitments that follow from the injunction to overthrow all
hierarchy everywhere have been explored in Roberto Unger's multivolume Politics (New York:
Cambridge University Press, 1987). For criticism of his argument, see my "Constructing
Politics," Political Theory 17 (August 1989): 475-82.
34. Wisconsin v. Yoder, 406 U.S. 205 (1972). From the standpoint of democratic justice,
Yoder was thus wrongly decided, although it would have been a more difficult case had the
parents pressed their best understandings of their children's interests rather than their own.
35. This is not to say that all polygamous regimes fare equally poorly from the standpoint
of democratic justice. Polygamous regimes from which there is no realistic chance of escape (as
when they are enshrined in a country's legal system as the only available form of marriage) fare
worse than polygamous regimes that are tolerated but not obligatory, and from which escape is
legally possible and not prohibitively expensive. Even in these circumstances there is always the
possibility that voluntary adherents have been brainwashed, of course, and arguments to this
effect cannot be dismissed out of hand. But their proponents will have to come to grips with the
eloquently reasoned denials of this that have been put forward by some Mormon women. It has
been argued, for instance, that polygamous marriage makes it possible for women to have both
a career and a family so that polygamy "is good for feminism." See Elizabeth Joseph, "My
Husband's Nine Wives," New York Times, April 9, 1991, A22.
36. The Oneida Perfectionists, founded in 1848 in Oneida, New York, by John Henry Noyes,
rejected all forms of private property, and extended their belief in community property to
community property in persons. Like the Mormon polygamists, they were persecuted by the
state, eventually abandoning their commitment to complex marriage in 1879. See Carol Weis-
brod, "On the Breakup of Oneida," Connecticut Law Review 14 (Summer 1982): 717-32. In fact,
the community was run in an authoritarian manner by Noyes, who decided unilaterally who could
marry, suggesting that the community would have been suspect on a number of grounds from
the standpoint of democratic justice. See Spenser Klaw, Wiithout Sin (New York: Allen Lane, 1993).
37. Rawls, A Theory of Justice, 42-5, 61-5, 82-9, 151-61.
38. It is not difficult, for example, to demonstrate the existence of contradictory imperatives
flowing from Rawls's lexical rankings. See T. M. Scanlon, "Rawls's Theory of Justice," in
ReadingRawls, ed. Norman Daniels (Bristol, UK: Basil Blackwell, 1975), 169-205; H.L.A. Hart,
"Rawls on Liberty and Its Priority," idem, 230-52; and Benjamin Barber, "Justifying Justice:
Problems of Psychology, Politics and Measurement in Rawls," idem, 292-318.
39. This is the view defended in Richard Ameson and Ian Shapiro, "Democracy and
Religious Freedom: A Critique of Wisconsin v. Yoder," in Nomos XXXVIII: Political Order, ed.
Ian Shapiro and Russell Hardin (New York: New York University Press, forthcoming).
40. From the standpoint of democratic justice, the Supreme Court thus reached the right
result in Bob Jones University v. United States, 461 U.S. 574 (1983), when it held that the federal
government may legitimately deny tax-exempt status to institutions that would otherwise qualify
but which engage in racial discrimination.
41. For a useful summary of recent literature, see Henry Hansman, "When Does Worker
Ownership Work?" Yale Law Journal 99 (June 1990): 1749-816.
42. Robert Dahl has argued that analogous skepticism is in order toward claims that
democratic control of nuclear arsenals and development interferes with their efficient deploy-
ment. See Controlling Nuclear Weapons: Democracy versus Guardianship (Syracuse, NY:
Syracuse University Press, 1985), 33-51.
43. Carmen Sirianni, "Learning Pluralism: Democracy and Diversity in Feminist Or

tions," in Nomos XXXV: Democratic Community, ed. John Chapman and Ian Shapiro (N
New York University Press, 1993), 283-312.
44. See James Fishkin, Democracy and Deliberation: New Directions for Democratic
Reform (New Haven, CT: Yale University Press, 1991).
45. Giovanni Sartori, The Theory of Democracy Revisited, vol. 1 (Chatham, NJ: Chatham
House, 1987), 119-20.
46. Deliberative polls must confront the difficulty that whoever sets the agenda may exert
disproportionate influence on the outcome, but this is a difficulty that every decision-making
procedure must confront. It is not the weakness in democratic theory that proponents of deliber-
ative polls are intending to resolve, though they, like proponents of other decision-making
mechanisms, need to be concerned about it.
47. See Susan Okin, Justice, Gender, and the Family (New York: Basic Books, 1989), 134-69.
48. See Charles Sabel and Jonathan Zeitlin, "Historical Alternatives to Mass Production,"
Past and Present 108 (1986): 133-76; and Hansman, "When Does Worker Ownership Work?"
49. See Herbert Alexander, Financing Politics: Money, Elections and Political Reform
(Washington, DC: Congressional Quarterly Press, 1976) and Frank J. Sorauf, Inside Campaign
Finance (New Haven, CT: Yale University Press, 1992).
50. In this connection a small but not insignificant victory was achieved for democracy in
March 1990 in Austin v. Michigan State Chamber of Commerce, 110 S. Ct. 1391, when the
Supreme Court cut back on the Buckley v. Valeo 424 U.S. 1 (1976) rule, which had held that
although contributions to political campaigns may be limited by legislation, limiting expenditures
constitutes a violation of the free speech clause of the First Amendment. In Austin the Court held
that some corporate expenditures on political speech may be regulated. As far as term limits are
concerned, there is considerable scholarly debate as to how bad the incumbency problem is and
whether or not term limits would be a solution to the problem of the ossification of power in
professional hands. They might, for example, lead to a net transfer of power from politicians to
bureaucrats, as Morris Fiorina suggests in Divided Government (London: Macmillan, 1992),
53-9.
51. John Dewey, "Democratic Ends Need Democratic Methods for Their Realization," New
Leader 22 (October 1939) reprinted in John Dewey, The Political Writings, ed. Debra Morris
and Ian Shapiro (Indianapolis, IN: Hackett, 1993), 206.
52. For elaboration, see Jennifer Hochschild, The NewAmerican Dilemma: LiberalDemoc-
racy and School Desegregation (New Haven, CT: Yale University Press, 1984).
53. Adam Przeworski, Democracy and the Market (Cambridge, England: Cambridge Uni-
versity Press, 1991), 183-4. See also Janos Kornai, The Road to a Free Economy: Shifting from
a Socialist System (New York: Norton, 1990); and Jeffrey Sachs, "The Transformation of Eastern
Europe: The Case of Poland" (The Frank E. Seidman Lecture, Rhodes College, Memphis,
Tennessee, September 26, 1991); Boris Pleskovic and Jeffrey Sachs, "Political Independence
and Economic Reform in Slovenia," in The Transition in Eastern Europe, ed. Oliver Blanchard,
Kenneth Froot, and Jeffrey Sachs, vol. 1 (Chicago: University of Chicago Press, 1994), 191-220.
54. "A pure public good has two salient characteristics: jointness of supply, and the
impossibility or inefficiency of excluding others from its consumption, once it has been supplied
by some members of the community." Mueller, Public Choice II, 11.
55. Przeworski, Democracy and the Market, 183-4.
56. On the collapse of the roundtable negotiations and the emergence of an elite pactbetween
the National Party and ANC leaderships, see my "Democratic Innovation: South Africa in
Comparative Context," World Politics 46, no. 1 (October 1993): 138-41.
57. As far as political elites sprinkling in benefits for themselves is concerned, the 1993
constitution requires that any member of parliament who ceases to be a member of his or her
political party will also cease to be a member of parliament, being replaced by someone else
from the party's parliamentary list. As for bribes, all civil service jobs and salaries were
guaranteed for at least five years following the transition, and in the last weeks before the election
President de Klerk transferred some three million acres of land to Zulu king Goodwill Zwelitini
in order to prevent their falling under the control of the new national government following the
April 1994 elections. New York Times, May 24, 1994, A6. For further discussion, see Courtney
Jung and Ian Shapiro, "South Africa's Negotiated Transition: Democracy, Opposition, and the
New Constitutional Order," Politics & Society 23, no. 3 (September 1995): 269-308.
58. See Laurence H. Tribe, American Constitutional Law, 2d ed. (New York: Foundation,
1988), 251-75.
59. On the changing law of marital rape in the United States, see Michael Freeman, "If You
Can't Rape Your Wife, Who[m] Can You Rape? The Marital Rape Exception Re-examined,"
Family Law Quarterly 15, no. 1 (Spring 1981): 1-29; Deborah Rhode, Justice and Gender
(Cambridge, MA: Harvard University Press, 1989), 249-51; Rene I. Augustine, "Marriage: The
Safe Haven for Rapists," Journal of Family Law 29, no. 3 (1990-1): 559-90; Sandra Ryder and
Sheryl Kuzmenka, "Legal Rape: The Marital Exception," John Marshall Law Review 24 (1992):
393-421. On the English evolution of the exception, see P. M. Bromley and N. V. Lowe, Family
Law, 7th ed. (Salem, NH: Butterworths, 1987), 109-12.
60. See H. H. Clark, The Law of Domestic Relations in the United States, 2d ed. (Saint Paul,
MN: West, 1988), 589.
61. For an account of the extent and effects of these policies, see Helen Suzman, In No
Uncertain Terms (New York: Knopf, 1993), 65-212.
62. This says nothing about which state institutions are most appropriate for the purpose,
whether courts, legislatures or executive agencies. See below.
63. Walzer, Spheres of Justice, 3-30.
64. I have noted elsewhere that the appeal to shared meanings fails because these are
invariably in contention. "Three Ways to Be a Democrat," 130-5.
65. It should be evident from my earlier discussion of public goods and state culpability that
I also think Walzer is mistaken in thinking that preventing dominance is the only legitimate basis
for the imposition of solutions by the state.
66. Robert A. Burt, The Constitution in Conflict (Cambridge, MA: Harvard University Press,
1992), 29.
67. Brown v. Board of Education 1, 347 U.S. 483 (1954).
68. Brown v. Board of Education II, 349 U.S. 294 (1955).
69. Burt, Constitution in Conflict, 271-3 10.
70. Ruth Bader Ginsburg, "Speaking in a Judicial Voice" (Madison Lecture, New York
University Law School, March 9, 1993, mimeo), 30-8. See also Nomination of Ruth Bader
Ginsburg to Be an Associate Justice of the United States Supreme Court: Report Together with
Additional Views, Exec. Report 103-6-93-1 United States Senate.
71. Roe v. Wade, 410 U.S. 113 (1973).
72. Ginsburg, "Speaking in a Judicial Voice," 32.
73. Burt, Constitution in Conflict, 348.
74. Ibid., 349-52. The Ginsburg-Burt approach was finally adopted by the Supreme Court
in Planned Parenthood of Pennsylvania v. Casey, 112 S. Ct. 2791 (1992). By affinming the
existence of a woman's fundamental constitutional right to an abortion, recognizing the legiti-
macy of the state's interest in potential life, and insisting that states may notpursue the vindication
of that interest in a manner that is unduly burdensome to women, the Court set some basic
parameters within which legislatures must now fashion regulations that govern aborti
Casey dissenters are right to point out that there will be a degree of unpredictability and c
as different regulatory regimes are enacted in different states and tested through the cou
S. Ct. 2791, at 2866 (1992). On views of adjudication thatencourage efficiency and clarit
all else this will appear to be a reprehensible invitation to further litigation. On the Ginsb
view, however, that Casey invites litigation may be a cost worth paying. It places the b
coming up with modes of regulating abortion that are not unduly burdensome on democra
elected legislatures, and forces them to do this in the knowledge that the statutes they en
be tested through the courts and thrown out if they are found wanting. These issues are t
further in my introduction to Abortion: The Supreme Court Decisions (Indianapolis, IN: Ha
1995), 1-23.
75. Dred Scott v. Sandford, 60 U.S. 393 (1856); In re Civil Rights Cases, 109 U.S. 3 (1883);
and Lochner v. New York, 198 U.S. 45 (1905).
76. From an address to the Philosophical Union of the University of California in November
1918, reprinted in Dewey, The Political Writings, 44.
Ian Shapiro is professor of political science at Yale. His new book, Democracy's Place,
will be published by Cornell University Press this fall.

Elements of Democratic Justice: Ian Shapiro Yale University

Uploaded by

Copyright:

Available Formats

Elements of Democratic Justice: Ian Shapiro Yale University

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Elements of Democratic Justice: Ian Shapiro Yale University

Uploaded by

Copyright:

Available Formats

ELEMENTS OF DEMOCRATIC JUSTICE

DEMOCRATIC MOVEMENTS DERIVE much of their moral authority

POLITICAL THEORY, Vol. 24 No. 4, November 1996 579-619

aspire to get on with our collective lives in democratic ways on my account,

TWO DIMENSIONS OF DEMOCRATIC JUSTICE

actually rule."7 In the modern world at least, democratic movements have

If democracy is understood to require that the people be sovereign over

the fundamental question of astronomy is whether or not there ought to be

deliberation, which, in turn, are believed most likely to lead to discovery of

conceded that those purposes and activities are inevitably in contention?

against other imperatives. But the general argument counsels suspicio

particular collective decision. These difficulties should not therefore count

Barrington Moore contends that "the existence of a legitimate and, to some

Below I argue that the imperative to make effective opposition pos

surprising. To say this is not, however, to respond to the difficulty that

circumstances, or about how these constraints should be enforced. By itself,

Finally, the general argument for democratic justice directs us to attend to

SOME DIFFICULTIES CONSIDERED

The preceding elaboration of the two central dimensions of demo

Conflicts Internal to Democratic Justice

Any argument for an internally complex set of principles must co

system of governance in a domain may conflict with the presumptive

accompanied established churches might also counsel that there is wisdom

conduct themselves in victory, the stronger is the obligation on the de

Tradeoffs between Democratic Justice and Other Goods

Additional sources of tension arise from my argument being premised on

devised oversight mechanisms that, however imperfectly, ensure some demo

Competing Demands of Different Domains

THE ROLE FOR THE STATE

ted. Democratic justice generates determinate conclusions in particular con-

Antivanguardism and Its Limits

Whenever anyone claims to know how to get to democracy undemocrati-

is no secular analogue to "Christianizing the infidels" to justify such

Legislatures versus Courts

The aspiration to avoid imposed solutions suggests that the presumption

conceptions, and only a subset of academic conceptions of liberalism at that. This is co

Deference to Legislatures," The Hoover Institution, Stanford University, Working Pape

43. Carmen Sirianni, "Learning Pluralism: Democracy and Diversity in Feminist Or

You might also like