Political autonomy is a democratic value – a democratic instantiation of collective self-
determination. Its great champion was Rousseau. Most people find Rousseau’s ideal of political
autonomy wildly unrealistic for modern societies. The main purpose of this paper is to present
an account of political autonomy that shows that it has at least a shot of being a desirable and
feasible ideal for modern societies. After completing this task, the paper considers briefly the
extent to which political autonomy, if accepted as a sound ideal for modern societies, can
provide support for certain exclusionary liberal principles, such as public reason and state
neutrality. The conclusions drawn on this issue are modest and hedged.
My topic is political autonomy. My primary aim is to characterize this value in a way
that shows that it has at least a shot of being a desirable and feasible ideal for modern political
societies. My secondary aim is to discuss its normative force. In particular, I am interested in
exploring whether, and to what extent, this value can provide substantial support for certain
(exclusionary1) liberal principles, such as the principle of state neutrality or the principle of
public reason. My conclusions are modest and hedged.
Political autonomy is not the same value as personal autonomy. A person can be
autonomous whether or not he lives in a political society that realizes political autonomy.
Political autonomy obtains only in a democratic society. More precisely, it obtains only when
those who are subject to political authority (i) participate, or have the option to participate, in its
exercise (on roughly equal terms) and (ii) authorize it, or can authorize it, as reflecting their own
will or reason. Political autonomy is thus a specifically democratic instantiation of political self-
The roots of this democratic value go back to Rousseau. He was its great champion and
its most influential proponent. The solution to the fundamental problem of politics, Rousseau
argued, requires a people to fully achieve political autonomy. The fundamental problem of
politics is to “find a form of association that defends and protects the person and goods of each
associate with all the common force” while at the same time ensuring that each associate
“uniting with all, nevertheless obeys only himself and remains as free as before.” The solution to
this problem is a political association with terms that it would be reasonable for each associate to
endorse. If each associate reasonably authorizes, or can reasonably authorize, the terms of the
political association that protects and defends him, and if each associate participates in the rule of
the association on roughly equal terms, then each will remain free. And when each associate
remains free, political autonomy is achieved.
Note that in describing political autonomy, and Rousseau’s statement of it, I have left an
important ambiguity in place. This concerns the nature of the required authorization. Must it
actually take place among the associates, or is it sufficient if it could occur? And if the latter is
the case, then what sense of possibility is expressed by the modal “could”? Depending on the
answers given to these questions, the value of political autonomy will be more or less demanding
and more or less difficult to achieve. A good case can be made that Rousseau believed that
actual authorization was necessary. Call this the actualist view. By contrast, the hypotheticalist
view requires only that it be possible for all associates to endorse the terms of the political
association that defends and protects them.
The hypotheticalist view has become the favored view in recent discussions of political
autonomy. Seeking to update Rousseau’s social contract, Rawls claimed that in a democratic
society – a society in which the people rule themselves and possess “an equal share” of its
“coercive political power” – “political power should be exercised, at least when constitutional
essentials and matters of basic justice are at stake, in ways that all citizens can publicly endorse
in the light of their own reason.” (JF, pp. 90-91 ) The last clause in the sentence expresses the
key requirement that all citizens in a democratic society should be able to endorse the basic terms
of their political association. The fact that some citizens, who could reasonably endorse the
basic terms of their political association, fail to do so would not, on this understanding, show that
the society did not achieve political autonomy.
Let us follow Rawls and other contemporary writers in assuming that the hypotheticalist
view of authorization is the favored view. There are two reasons for doing so. First, we want to
consider the prospects for grounding certain exclusionary liberal principles on the value of
political autonomy and the hypotheticalist view fits better with this ambition than the actualist
view. Second, political autonomy on the hypotheticalist view is easier to achieve than it is on the
actualist view and it thereby avoids some of the practical problems that bedevil actual
The hypotheticalist view of authorization, as I mentioned, needs to specify the sense of
possibility implicit in the notion of possible endorsement. This is a difficult issue. The view
also must specify whether the endorsement in question is merely permitted or required. This is a
less difficult issue, for if the endorsement were merely permitted, then arguably it would be too
permissive. As Scanlon has observed, many arrangements might be such that it would not be
unreasonable for people to accept them, even if also it would not be unreasonable for them to
reject them. For this reason, on the hypotheticalist view, we should favor the stricter demand
that the terms of the political association are ones that, reasonably, the associates are required to
accept. This is equivalent to the claim that the terms must be such that no associate could
reasonably reject them.2
This leaves the difficult issue of how to construe the “could” in the could not reasonably
reject demand. The problem can be framed by distinguishing associates as they are from
idealized versions of them. If the terms of a political association are ones that the associates
could not reasonably reject, then it is one that their idealized counterparts would authorize. The
problem then becomes how to fix the degree of idealization. Too much idealization will
undermine the whole point of hypothetical endorsement. Too little will fail to capture the
thought that the endorsement must be reasonable. Faced with this problem, many writers have
sought to identify a set of background commitments that people in democratic societies could
reasonably affirm. Idealization must not efface these commitments. Thus, on this common
view, the requirements of what could not be reasonably rejected (and hence could be authorized)
are indexed to the background reasonable commitments of democratic citizens.
This common view can be developed in myriad ways. The details need not detain us.
The key point for now is that the hypotheticalist view of authorization, at least as it has been
defended by its recent proponents, presupposes reasonable pluralism with respect to background
commitments. Furthermore, any idealization that is imposed on reasonable endorsement must
leave this reasonable pluralism in place. The hypotheticalist view is thus closely aligned with
contractualist principles of political legitimacy. Still, the democratic value of political autonomy
should not be identified too tightly with contractualist political legitimacy for two main reasons.
First, it may be possible for a political association to satisfy the standard of contractualist
political legitimacy while failing to realize political autonomy. Rawls hints at this possibility
with the following example.
[S]uppose (we wildly imagine) that the Prussian chancellor of Kant’s day, with
the support of the King, acts to ensure that all laws enacted are in accord with
Kant’s principle of the social contract. If so, free and equal citizens would – let us
say on due reflection – agree with them. Since citizens do not themselves freely
discuss, vote on, and enact these laws, however, citizens are not politically
autonomous and cannot thus regard themselves. (“Reply to Habermas,” p. 411)
Kant’s principle of the social contract is a contractualist principle of political legitimacy, but, if
Rawls is right, then it can be satisfied by citizens who are not politically autonomous. Political
autonomy requires, in addition, that all citizens participate, or at least have the option to do so, in
the process that generates the laws that apply to them. Second, it is possible to accept that
political autonomy is a valid ideal and yet reject contractualist principles of political legitimacy.
This claim may seem confused, since the achievement of political autonomy presupposes that
democratic citizens can reasonably endorse the terms of their political association and the
reasonable endorsement demand appears to express the requirements of contractualist political
legitimacy. But, as I explain below, the appearance is illusory if political autonomy and
contractualist legitimacy are values with a different normative status.
Political autonomy is a democratic value that is achieved when two conditions are jointly
satisfied. There is no guarantee, however, that the two conditions – what I will now dub the
participation condition and the collective authorization condition – will not come into conflict.
To illustrate, let us continue Rawls’ example. Suppose that if the Prussian chancellor were to
withdraw from his supervision of the laws and let the citizens themselves freely enact them, then
the citizens would enact laws that violated Kant’s principle of the social contract. In such
unhappy circumstances, the value of political autonomy would not be attainable and a choice
would need to be made as to which component of political autonomy – participation or collective
authorization – should take precedence over the other.3
The possibility of internal conflicts of this kind are important. They show that appeals to
political autonomy only have force in societies in which the two components are jointly
satisfiable. Still, even when political autonomy is attainable, it can conflict with the achievement
of other goods and values. As will become clear below, these external conflicts, as I shall call
them, are important in helping us to gauge the normative force of the value.
The two conditions I have distinguished can be satisfied in different ways. Consider the
collective authorization condition. It can be held to apply only at the level of fundamental
political institutions (viz. Rawls’ constitutional essentials and basic justice) or it can be held to
apply also to all political institutions and decisions. Call the former view modest and the latter
robust. Likewise, the participation condition can be held to apply only to fundamental
constitutional decisions, such as the decision to establish a monarchy or a democracy; or it may
be taken to apply to all laws. Call the former view constitutional and the latter comprehensive.
Combining these distinctions, there are four interpretations of political autonomy that we should
Collective Authorization Modest (1) (2)
Robust (3) (4)
(1) and (3) require participation only at the fundamental constitutional level. In Rawls’ example,
let us imagine, contrary to historical fact, that the Prussian monarchy enjoyed the ongoing
democratic endorsement of the Prussian people. Participatory assemblies of the sort Rousseau
recommended periodically convened and endorsed it. If that had occurred, then perhaps it would
have been sufficient for the participation condition to be satisfied, but if the Prussian citizens had
established constitutional democracy and in addition freely discussed and voted on all the laws
that bind them, then their realization of political autonomy would have been more substantial.
Presumably, that is why Rawls thought that the participation condition should apply to all laws. 4
Let us agree that the participation condition must be satisfied comprehensively. That
leaves (2) and (4). Why insist on modest rather than robust collective authorization? One reason
for doing so is the thought that modest authorization is more attainable than robust endorsement.
Requiring citizens to all reasonably endorse every law may seem to set the bar too high. Against
this concern, it can be said that if the fundamental political institutions meet the collective
authorization condition, then all laws passed within these institutions do so as well. In effect,
reasonable endorsement transmits down from the fundamental political institutions to ordinary
law. Call this claim transmission. If we accept transmission, then robust authorization will look
compelling. If we reject transmission, then we will need to confront more directly the concern
that robust authorization is too demanding.
For present purposes, we do not need to choose between (2) and (4). Modest and robust
political autonomy, let us allow, are both viable interpretations of the value.
Many people will strongly doubt that political autonomy is achievable for us even on the
modest interpretation. Unlike Rousseau’s ideal republic, modern societies simply contain too
much diversity to make the pursuit of the value a sensible goal. This is obviously an important
concern, one I will address below. But a deeper objection to political autonomy should be
addressed first. The objection holds that even if political autonomy were achievable for modern
societies, it would not be desirable to pursue it in practice.
Proponents of this objection – call it the desirability objection – can argue that the
realization of political autonomy is not an appropriate goal for first-person political deliberation.
Politics involves the exercise of coercive power over others. As such, it is vital that the terms of
political association be as just as they can be. For this reason, citizens should aim at justice in
their political deliberations about whom or what to support. They should not favor participatory
decision making procedures if there is good reason to believe that these procedures will be less
good at yielding just outcomes over time than less participatory procedures. Similarly, they
should not strive to meet the collective authorization condition, for doing so may lead them to
compromise justice for consensus.
The desirability objection raises a number of important issues. Can political autonomy
conflict with justice, or is it an essential component of justice? If the latter were true, then
citizens would need to aim at political autonomy insofar as they aimed at justice. And, in fact,
many contemporary writers contend that justice requires some form of democracy. They will
insist that the satisfaction of the participation condition is a necessary element of justice. It
might be claimed further that justice requires the collective authorization condition to be
satisfied. If a law is just, then it is reasonably endorsable by all. But this latter claim, as we will
see, is not one that proponents of political autonomy are well positioned to accept. We should
allow that justice and reasonable endorsement, in principle, can come apart. A successful
response to the desirability objection, accordingly, must not insist that the pursuit of political
autonomy can never conflict with the pursuit of justice.
A successful response to the objection is not hard to discern, however. Even if citizens
should not aim at political autonomy directly, they still could regard its realization as a valuable
by-product of their politics. Comparing two political orders that are equally just, they could rank
one over the other because it alone achieves political autonomy. But if they can recognize the
value of political autonomy in this indirect way, then they should be open to the possibility that it
could trump justice in some circumstances. A slightly less just political order that achieved
political autonomy, for instance, might be preferable to the more just political order that did not.
And if this were true, then good citizens would need, at least on occasion, to aim at political
autonomy directly rather than justice. Justice-loving citizens, we might say, do not have to be
fanatics about justice.
The upshot is that we can accept the desirability of political autonomy without
determining its importance relative to other political values. If the pursuit of political autonomy
and the pursuit of justice pull in opposing directions in a given set of circumstances, then we will
need to decide how to balance them in these circumstances. That difficult issue will be taken up
below. Yet even on the extreme assumption that political autonomy is always subordinate to all
other political values, its achievement can still be regarded as desirable.
Rousseau held that political autonomy is only achievable in a small republic whose
members adhere to a civil region and share a thick conception of the good life. Such a political
community is not a live option for large modern societies. Thus the real concern behind the
desirability objection may be not that political autonomy has no value, but rather that serious
efforts to realize it in the modern world will be repressive and exclusionary.
A response to this concern must show that political autonomy is achievable in modern
societies and that it can be achieved without enforced uniformity. Rawls held that in a well-
ordered society “everyone has a similar sense of justice and in this respect a well-ordered society
is homogeneous.” (TJ, p. ) But he emphasized that this does not mean that a well-ordered
society must be homogeneous with respect to conceptions of the good or to other matters. The
pressure of pluralism forces a retreat to justice. A politically autonomous society must be
organized around a shareable conception of justice and nothing more.
We do not need to view the retreat to justice as a lamentable retreat. Pluralism can be
good with respect to conceptions of the good, while being bad with respect to conceptions of
justice. To defend this asymmetry, we can follow Rawls and invoke Humboldt’s ideal of a
society that brings forth the full development of human talents and powers. We have reason to
value other people’s diverse conceptions of the good since we cannot realize all that is valuable
ourselves. Call this Humboldt’s claim. Its truth would explain why political autonomy with
pluralism (on the good) is not an ersatz version of Rousseau’s ideal.
A problem now presents itself. Not every conception of the good is valuable in the way
expressed by Humboldt’s claim. Some conceptions of the good, including some which are
consistent with justice, add nothing to the full and valuable development of human powers and
talents. Nonetheless, some of these conceptions are (likely) reasonable in the sense implicated in
the hypotheticalist view of the collective authorization condition.
The problem is not simply that Humboldt’s claim does not line up neatly with this
condition. It is rather that the appeal to this claim does not adequately express the
accommodation to pluralism that is required. An adequate characterization of political autonomy
for the modern political world, it is often said, must acknowledge the limited “scope of practical
reason.”5 Fully reasonable people who live under free conditions can accept misguided
conceptions of the good. That fact, if it is a fact, may be unfortunate, but it is a consequence of
the limits of practical reason; and it must be acknowledged if the pursuit of political autonomy is
not to become the dangerous business of enforced homogeneity that those who reject political
autonomy suspect it will become.
Suppose that we accept these claims. Then we must press the retreat further. For the
limits of practical reason, if accepted, also explain why good faith efforts by fully competent
reasoners will lead them to disagree on questions of justice. This will be case, even if the
reasoners in question are all committed to supporting a conception of justice that other citizens
can collectively authorize. A conception of political autonomy for the modern world, it now
must be said, must abandon the claim that every citizen should have a similar sense of justice.
The homogeneity that Rawls claimed characterizes a well-ordered society will not come about
under free conditions.
What then is left after this further retreat? The popular answer today is that there exist a
family of conceptions of justice, and that if any member of the family is selected by democratic
means, then the institutions that this conception regulates will be such that they can be authorized
by all. Political autonomy is thus achievable, even in the face of disagreement over justice and
the good. Call this weak political autonomy to contrast it with the strong political autonomy that
is realized when citizens share a conception of justice. Given reasonable pluralism, weak
political autonomy will not be achieved unless citizens aim to achieve it. (It is extremely
unlikely that it would result as a by-product of each citizen pursuing justice as he sees it.) Here
then is the picture that we have arrived at. Each reasonable citizen aims at a conception of
justice that, or so she believes in good faith, could be reasonably endorsed by all. Reasonable
citizens as a group agree on a set of such conceptions, but disagree over which member of the set
is the best. Still, since they agree on the set, they can agree that any member of the set, if
selected by the right kind of procedure, could serve as a legitimate conception of justice in the
sense that it could be authorized by all. Under the pressure of pluralism, this weak instantiation
of political autonomy becomes as close as we in the modern world can come to the Rousseauian
ideal of citizens giving the law to themselves.
So construed, political autonomy may be a valid ideal for modern societies. It is not, I
think, obviously utopian in the pejorative sense. Still, more needs to be said. There are different
ways to understand the nature of this value. Most importantly, it can be understood either to
generate requirements or, as I will put it, aspirations. This difference in the kind of demands that
a value generates I will refer to as its normative status.
Normative requirements are internally related to permissions. If some course of action is
required for a particular agent on a particular occasion, then he is not permitted not to do it. It is
possible that political autonomy generates requirements for societies capable of achieving it.
The value might explain why modern societies are morally required to satisfy the two conditions
we have been discussing – the participation condition and the collective authorization condition.
Failure to satisfy either or both conditions would render illegitimate (i.e. morally impermissible)
the exercise of political authority in these societies.
By contrast, if political autonomy is an aspirational ideal, then it will not generate
requirements to comply with its component conditions. The failure to realize political autonomy
will be a shortcoming of the society, but it will not undercut the moral permissibility of the
political authority exercised in the society. This point explains why, as I hinted at earlier, it is no
contradiction to affirm the value of political autonomy while rejecting the contractualist principle
of political legitimacy that requires political institutions to satisfy the collective authorization
condition. The contractualist principle in question articulates a requirement for morally
permissible institutions, but if political autonomy is an aspirational ideal, then it will generate no
A good case can be made that political autonomy is best viewed as an aspirational ideal.
The case appeals to the very concern that guided the characterization of the value in response to
what I termed the pressure of pluralism. That concern is that in characterizing the value of
political autonomy we should be sensitive to its feasibility. The move from actualist to
hypotheticalist collective authorization, the retreat from Rousseau’s ideal of a community united
on ethics and religion to one united only on justice and the retreat from strong to weak political
autonomy were all spurred by the need to reconcile political autonomy with the pluralism of the
Aspirational values are values to be approximated. They represent a form of excellence
that is achieved to a greater or lesser extent. By contrast, values that generate requirements are
best understood as ones that either are or are not realized. Thus, if political autonomy generates
a requirement that all citizens must be able to reasonably endorse the terms of their political
association, then if this requirement were not fully satisfied, then the value would not be realized.
However, if political autonomy instead generates aspirational demands, then as its two
conditions are more fully satisfied, it will be realized to a greater extent. This scalar dimension
of the value can be understood as follows. Political autonomy will be realized to a greater extent
as more individual citizens that have an equal opportunity to participate in the making of the
laws that bind them can reasonably endorse them.6 Plainly, then, if political autonomy is
characterized as an aspirational value, then it will be more plausible to view it as a realistic
political goal for modern societies. Its full realization may seem improbable, but a substantial
realization of the value may not be out of reach for those societies committed to it.7
Viewing political autonomy as an aspirational ideal, however, has an important
consequence for its power to support exclusionary liberal principles. The realization of political
autonomy becomes a desideratum, not a value that generates requirements, and this means that,
at most, it can support exclusionary liberal principles as a desideratum, one that (likely) can be
defeated by other considerations of political morality. This brings us finally to the issue of the
normative force of political autonomy.
To gauge its normative force, we need to consider possible conflicts between political
autonomy and other values. With this in mind, let us revisit the desirability objection discussed
above. This objection highlights the potential conflict between political autonomy and justice. It
is possible that efforts to achieve political autonomy never obstruct efforts to achieve justice.
This no conflict view may be accepted by some proponents of political autonomy. But, as a
general claim, it is not very plausible. Efforts to satisfy the participation condition, in some
circumstances, will make it more likely that unjust decisions will be reached. The same can be
said of the collective authorization condition. That condition is satisfied, on the weak view of
political autonomy, when one member from the set of reasonably endorsable conceptions of
justice, after being selected by an appropriate process, regulates the fundamental political and
legal institutions of a society. Now suppose that the selected conception, call it J1, is one of
several reasonably endorsable conceptions and that one of these alternatives, call it J2, is a
superior conception of justice. Next suppose that some of the most compelling arguments –
compelling both in terms of their soundness and their motivational power – would be excluded
by exclusionary liberal principles grounded on the value of political autonomy. In this scenario,
efforts to achieve political autonomy would impede efforts to achieve justice.
One response to this possibility is to sidestep it and claim that political autonomy applies
only under conditions in which political autonomy and justice do not conflict. But this response
does not help us to think about the normative force of the value. A second response is to assign
priority to one or the other value. Justice trumps political autonomy or political autonomy
trumps justice. A third response is to deny any general priority to either value and to allow that
sometimes one takes priority over the other and sometimes the other takes priority over it. On
this third response, everything depends on the context and the issues in play.
The third response, in my judgment, is closest to the truth; but it is also the response that
provides the least guidance. It is not at all clear how to determine the weights of conflicting
values. Still, if either it or the view that justice trumps political autonomy is accepted, then
political autonomy cannot support general exclusionary principles. At most, it can provide
support for a modest idea: political institutions and decisions that satisfy such principles are
desirable. Depending on the weight given to the realization of political autonomy and depending
on how likely and how significant are the conflicts between justice and political autonomy, this
support will be greater or lesser.
I have drawn a number of distinctions to bring the value of political autonomy into
sharper focus. To review, these include:
Strong / Weak
Robust / Modest
Requirements / Aspirations
I have argued that we should embrace weak over strong political autonomy. And I have claimed
that political autonomy is best understood as an ideal that grounds aspirations, not requirements.
I have remained neutral on whether political autonomy is best understood as robust or modest.
This issue, I have said, turns on the plausibility of transmission, which is a claim that cannot be
addressed here. (Many proponents of political autonomy, however, have accepted transmission.
They have claimed that only the fundamental political institutions of a society must meet the
collective authorization condition.) I have said almost nothing about how much weight the value
of political autonomy has – i.e how strong is its capacity to override other values and ideals. I
have claimed only that it is implausible to hold that it always overrides claims of justice.
Consider now a view that affirms the weaker or more modest fork in each of the above
contrasts. Such a view would hold that political autonomy does not require its citizens to share
either a conception of the good or a conception of justice, but only to view the conception of
justice in force in their society as one that has been appropriately selected and one that is in a set
of conceptions of justice that are all reasonably endorsable. Further such a view would hold that
the collective authorization condition need only apply at the fundamental political level. Finally,
such a view would claim that political autonomy is an aspirational ideal.
Would this view provide substantial support for exclusionary liberal principles? It would
permit the state to engage in excluded state action when such state action was justified by
considerations or values that properly override the value of political autonomy. A proponent of
this view, who also assigned only modest weight to the value, could accept that political
autonomy is a genuine value – and reject the desirability objection discussed above – but deny
that any general exclusionary principles could be derived from it. Contrast this view with a
second view that is the same as the first except that it holds that political autonomy is a very
weighty aspirational ideal, one that only in very rare circumstances can be overridden. On this
second view, political autonomy might provide substantial support for exclusionary liberal
Nothing I have said here tells us which of these two views have a stronger claim to
acceptance. The more that one thinks that efforts to realize political autonomy will impede or
obstruct the realization of other important political values, such as justice or the promotion of
excellence, the more one will incline toward the first view. By contrast, those who accept the
second view, in all likelihood, believe that the pursuit of political autonomy does not conflict
significantly in practice with the pursuit of justice and they likely believe, in addition, that there
are independent reasons that establish that the state should not pursue other conflicting ideals.
These prior beliefs and commitments, and not the value of political autonomy alone, likely
explain why they think that it can contribute significantly to the grounding of exclusionary
Exclusionary principles exclude a specified class of considerations from playing any role in the
justification of political arrangements, laws, decisions, etc. Excluded state action is state action that is
ruled out by these principles.
Henceforth, I will use the phrase “could reasonably accept” to mean “could not reasonably reject.”
One might conclude that neither component has value if both cannot be satisfied; but this would be an
Does Rawls favor comprehensive participation? Not straightforwardly, since he does not require
ongoing democratic participation at the constitutional level. However, he does claim that while “those
already living in a just constitutional structure cannot found a just constitution . . . they can fully reflect on
it, endorse it, and so freely exercise it in all ways necessary.” (Reply to Habermas, p. 403)
J. Cohen, “Moral Pluralism and Political Consensus,” in Philosophy, Politics, Democracy, pp. 38-60.
Possibly, and in addition, the value will be realized to a greater extent the more citizens actually
participate in making the laws that bind them.
Political autonomy may provide support for efforts to find broadly based justifications for political
institutions and decisions, where broadly based justifications are not ones that satisfy the more demanding
“no one could reasonably reject” standard.
University of Arizona