. 9
( 13)


trates on the reasonableness of basic deliberative capacities and dispositions that
are needed in order to make adequate use of public reason. I will argue
that a focus on deliberative capacities and dispositions is demanded by the
idea of liberalism without agreement.
If we concentrate on the reasonableness of deliberative capacities and
dispositions, autonomous participation in practices of public reasoning
presupposes wholehearted commitment to basic civic values of freedom,
equality, rule of law, peaceful social and political cooperation, the bur-
dens of judgment, and an understanding of the circumstances of poli-
tics. Although it requires that one, in practice, be loyal to a good enough
framework for cooperation, it does not require that one be committed
to a publicly favored interpretation of the civic practices it holds in place. It
rather requires that one develop capacities and dispositions that enable
one to recognize and be loyal to the authority of non-ideal solutions, to
express competently one™s own interpretation of civic ideals in reciprocal
democratic deliberation over problems generated by non-ideal solutions,
and to be critical of and willing to learn from one™s own and others™ po-
litical conceptions in reciprocal deliberative exchanges in light of the
general civic ideal to let the needs and interests of all citizens matter
equally in social and political affairs. Again, given the circumstances of
politics, there need not be substantive agreement on the exact require-
ments of this general civic ideal. But liberalism without agreement has
to admit that a minimal agreement on its importance as a condition of a
reasonable civic disposition is necessary.
According to the reasonableness-of-opinions-and-arrangements un-
derstanding, to be reasonable is to be committed to a “thick” set of
values and principles that specify a canon of reasonable opinions and
arrangements. On this view, it is assumed that public culture is shaped
by substantive values, institutional arrangements, and ways of reasoning
Political Autonomy and Agonistic Citizenship 259

that have proved themselves in a historical, consensus-oriented process of
trial and error and must be regarded as legitimate outcomes of the use of
public reason. These outcomes of the use of public reason over time will
undoubtedly specify an understanding of capacities and dispositions, too,
but they will do so in a way that is more restrictive “ and possibly backed up
by a claim to the objectivity of the canon of opinions and arrangements “
than the capacities-and-dispositions view thinks necessary.
The opinions-and-arrangements view does not recognize the funda-
mental tension between non-ideal but authoritative solutions and a plu-
rality of ideal conceptions of justice. It makes the mistake of understand-
ing dominant interpretations of political cooperation that are backed
up by democratic majority decision as interpretations that all reasonable
citizens should wholeheartedly accept. It confuses the need for practical
and authoritative solutions to empirical problems of political cooperation
with the aim for objective principles that can function as yardsticks for
reasonable and autonomous political action. For that reason, this view
runs the risk of excluding from view the political vocabularies, needs,
and interests “ and therefore the agonistic political autonomy “ of those
citizens who hold ideal conceptions that do not ¬t with the authoritative
practical solutions to problems of cooperation. It lacks what John Christ-
man calls “emphatic respect” for all authentic articulations of political
values and conceptions that citizens judge to be important input to rea-
soning practices “ that is, it lacks “an attempt at emphatic grasping of the
subjectivity and motivations of others.”37
Agreement-based political liberalism has in recent years often been
criticized for defending a notion of public reason and political auton-
omy that is too substantive to live up to its aspirations as a freestand-
ing political doctrine. Often, political liberalism has been criticized for
silently presupposing individualistic, post-traditionalist lifestyles that eas-
ily ¬t with the normative framework that political liberalism proposes
to be inherently better than other lifestyles. The claim is that, inadver-
tently, political liberalism formulates criteria of autonomy and reason-
ableness that those who hold individualist and post-traditionalist opin-
ions, and favor attendant legal and political arrangements, can answer
to far more easily than citizens with more collectivist and possibly tradi-
tionalist ideas of the good.38 The best way in which political liberalism
can answer this criticism is by ¬rmly stating that a notion of public rea-
son should be understood from a capacities-and-dispositions view rather
than from an opinions-and-arrangements view. The most important cri-
terion for living up to requirements of the use of public reason is that
Bert van den Brink

citizens “ irrespective of their conceptions of the good “ be disposed to
engage in reciprocal democratic deliberations that will result in good
enough democratic decisions, not that they will in the end come to favor
this or that opinion or arrangement that issues from public delibera-
tion. That general disposition “ the requirements of which I spelled out
earlier “ is a necessary condition for civilized ways of engaging in inter-
pretative con¬‚ict over both ideal and non-ideal understandings of consti-
tutional essentials and justice. The point is as simple as it is demanding.
Yet it is crucial to our understanding of liberalism without agreement.

IV Civic Virtue and the Need for Endurance and Responsiveness
In a liberal society without agreement, in which objective principles for
right action are lacking, the stability and justice of political interactions
becomes strongly dependent on the civic dispositions citizens act from.
But in such a society, the exercise of civic virtue cannot be understood as
following pre-given rules of civic conduct or as derived from an uncon-
tested view of human excellence. Where there is no ¬xed and stable set of
fully just political principles, the question as to what it means in practice
to act as an autonomous, competent citizen who aims to foster the com-
mon good of all citizens has no de¬nitive answer. Acting for the common
good cannot be understood as acting according to one™s own or dominant
legal understandings of the common good. For, as was noted before, one
cannot be sure that these understandings are correct. One of the conse-
quences of this is that acting autonomously is not just about giving oneself
and others acceptable laws in light of clear criteria of what one considers
to be right; it is also about discovering time and time again, in interac-
tion with others, what one really believes and what one can or cannot
accept about those beliefs.39 Finding out which dispositions and virtues
are required in which situation becomes a continuous self-critical task for
citizens who aim to make competent autonomous use of their reasoning
capacities. Citizens™ civic dispositions and the autonomy of their actions
will be expressed, tested, and revised in their encounters over practical
questions of distribution and opportunity (justice), of mutual respect
(recognition), of the ordering of public debates (reasonableness), of giv-
ing all parties their due in political decisions (reaching compromises), of
accepting or rejecting utterances and actions of their fellow citizens (tol-
erance), and so on. In such encounters, citizens will test the acceptability
of (1) their own civic competencies and ideals in light of (2) the compe-
tencies and ideals of fellow citizens and (3) predominant but contestable
Political Autonomy and Agonistic Citizenship 261

interpretations of civic competencies and ideals as laid down in law. As a
result, revisions of all three dimensions may become necessary. The cit-
izen who contests alternative interpretations of civic competencies and
ideals will entice a critical test of her own interpretations. Her allowing
such a test to take place is a sign both of her reasonableness and of her
civic virtue; it will warrant that interpretative con¬‚icts will be more than a
mere collision of self-righteous wills.40 Furthermore, allowing the test to
take place is a condition of developing a civic attitude that is conductive
to learning processes that enable the internalization of authentic civic
dispositions that are more than functional requirements of an existing
political order. For good reasons, liberal theory has always been wary of
such functionalistic understandings of virtue, which require citizens to
subordinate their action to quasi-unquestionable but arbitrary forms of
power. Given its suspicion of arbitrary claims to sovereignty, the agonistic
account of civic virtue does not involve that requirement.
Agreement-based liberalism has often tried to free citizens from the
burden of developing good political dispositions (most famously per-
haps in Kant™s re¬‚ections on the nation of devils). However, for lack of a
substantive agreement that would make the near total proceduralization
of virtue (justice, reciprocity, reasonableness, tolerance, and so on) in
strong institutions a viable option, liberalism without agreement cannot
do that. Indeed, it is exactly because of the lack of an agreement-based
common framework that much responsibility for the quality of politi-
cal cooperation in a society falls on the shoulders of citizens and their
What is remarkable about many liberal accounts of civic virtue is that
although they are quite detailed, they do not say much about how they
may guide citizens through a pluralistic condition in which aspects of
the very basic framework of cooperation comes up for deliberation and
amendment time and time again. Liberal theorists of civic virtue are surely
right that all citizens at times need to exercise virtues such as courage,
justice, loyalty, law-abidingness, tolerance, self-restraint, conscientious-
ness, perseverance, and independence of thought.41 But because these
theorists think about these and other virtues against the background of
quite substantive understandings of an inherently stable and nearly just,
agreement-based liberal society, they do not stress suf¬ciently that the
exercise of these virtues within a liberal framework does not require the
same kinds of attitudes and actions from all citizens. Let me elaborate
on this point by introducing to the debate over civic virtues two general
civic dispositions that I believe help citizens who are positioned on the
Bert van den Brink

map of society in speci¬c ways to exercise political autonomy and civic
virtue in their own ways “ ways that do not surrender to arbitrary power
relations. The ¬rst of these I label “civic endurance” and the second,
“civic responsiveness.”42
Civic endurance is the disposition of being willing to carry the burdens
of citizenship that come with being a member of a political community the
given principles and concrete policies of which one cannot wholeheart-
edly agree to. Civic responsiveness, on the other hand, is the disposition
of being forthcoming to those who tend to carry these burdens “ in our
societies, members of cultural and religious minority groups, citizens in
vulnerable socioeconomic positions, and citizens with “alternative” tastes
and lifestyles, especially. We may call these dispositions general virtues of
civic cooperation, since “ or so I claim “ they are crucial to the general
civic competencies that decide whether or not citizens are able to act
from more speci¬c civic competencies and virtues in ways that ¬t well
with the circumstances of politics in pluralistic societies.
Although in some way or another, all citizens at times need to exer-
cise civic endurance, it is likely that the members of social, cultural, and
political minorities in society will experience the burdens of endurance
most strongly. Let us assume that members of such minorities, as citizens,
often have the experience that their prospects of leading a good life in
light of their own political and more comprehensive conceptions of the
good are somewhat under threat because of their lack of effective politi-
cal power. They may have equal political rights, but may at the same time
feel that the dominant liberal-democratic terms of civic cooperation and
dominant but discriminatory informal standards of evaluation in society
make it impossible for them to effectively claim their legitimate needs
and interests in political deliberation. They may also feel that they lack
the social and educational resources they need in order to stand up for
themselves and to gain information that is relevant to gaining suf¬cient
insight in the reasons for their condition. Such claims, which are hardly
ever claims to the total illegitimacy of political arrangements (and all
the more reasonable because of that), are of course familiar from the
debates over feminism, multiculturalism, and special political and cul-
tural rights.43 Depending both on the seriousness of their condition and
the level of their own civic-mindedness, citizens who endure the conse-
quences of being in a less favorable position in society may respond in
a number of ways.44 First, they may passively endure the consequences of
their position in society “ that is, regard it as an unchangeable condition.
Second, they may actively endure their situation “ that is, regard it as a
Political Autonomy and Agonistic Citizenship 263

condition that is caused by a mistaken, contestable, but in principle al-
terable interpretation of civic ideals in their society. Third, they may not
want to endure their situation at all and, for instance, resort to the use of
civil disobedience or even political violence.
I call the middle position of active endurance of one™s position in
society one of civic endurance “ the endurance of burdens that come
with civic cooperation in a mode that still relies on the possibility of re-
pairing certain aspects of these burdens through civic interaction within
the limits set by the law. Note that the other two attitudes are not al-
ways to be condemned: passive endurance may be a prudent strategy of
sheer survival in a totalitarian society, and resort to civil disobedience
and even political violence is not necessarily morally illegitimate. They
are to be condemned only to the extent that they are engaged in by
members of a civic community that is suf¬ciently responsive to the claims
of those who endure actively (I will come to that shortly). Those who
endure actively still rely on their perhaps threatened, but not wholly ab-
sent, political autonomy as de¬ned both by their legal status and by the
not self-righteous acknowledgment of their moral and legal right to po-
litical autonomy by at least some of their fellow citizens who have the
power to change, through political action, the condition of those who
endure arbitrary but repairable burdens of membership of the political
As Russell Bentley and David Owen put it, civic endurance is “a dynamic
political way of being that captures an agonistic sense of belonging.”45 On
the one hand, this way of being is directed at making oneself heard to one™s
fellow citizens with the intention of repairing unequally distributed civic
burdens, and thereby repairing the integrity of the civic community as
a whole. It captures an agonistic sense of belonging and of autonomous
membership because it necessitates a reciprocal civic competition over
the best evaluation of the situation at hand and the best options for repair-
ing the shortcoming of the civic order. On the other hand, it is directed,
for each of the participants in the political encounter at hand, at dis-
covering a political self-understanding that ¬ts with one™s speci¬c position
within the civic order. So civic endurance may be said to be a matter of
productively shaping one™s civic relations to others as well as one™s self-
understanding as a citizen. The citizen who actively endures civic burdens
understands that the civic order as a whole, in its current arrangement,
is by nature contested and non-ideal, and that this exact circumstance
makes it possible (and necessary) to contest the constellation that is re-
sponsible for the arbitrary but repairable civic burdens she carries.
Bert van den Brink

Like civic endurance, civic responsiveness is a virtuous civic disposition
that must be delimited from two possible reactions to the claims for fairer
distribution of civic burdens: blindness to and repression of such claims. Just
as is the case with civic endurance, civic responsiveness concerns one™s
relation to the civic order as a whole, one™s civic relations with others,
as well as one™s own self-understanding as a citizen. As to the relation to
the civic order as a whole, the responsive citizen understands that this
order is, by its very nature, a contested one. By recognizing this, she
also recognizes, ¬rst, the contested nature of her own civic beliefs and
ideals and those of others; second, the gap that is to be expected between
“pure” civic ideals and actual civic practices; and, third, the circumstance
that for some groups in society this gap “ and the burdens that come
with it “ will be bigger than for others. For her relation to other citizens,
this implies that the responsive citizen “ let us say, a civically minded
member of the majority culture in society “ will investigate to what extent
other citizens™ political claims may be justi¬ed even though they do not
¬t well with dominant arrangements and opinions in society (remember
the section on the use of public reason). With respect to her own civic
self-understanding, this implies that the responsive citizen is willing to
accept that not only the current state of society, but also her own political
conception of it, may well be blind to legitimate needs and interests
of citizens whose political voice is not easily heard.46 In this threefold
sense, civic responsiveness enables the citizen to make better political
judgments “ that is, to exercise her political autonomy in a way that is
true to the actual circumstances of politics in her society instead of being
true to a powerful but partly arbitrary interpretation of core civic ideals.
Together, civic endurance and civic responsiveness warrant a political
sense of reality and ¬‚exibility of the citizen that, like a sense of justice and a
sense of the good, are preconditions of autonomous political judgment.
The citizen who civically endures the burdens that come with citizenship
recognizes that her lack of consent to the way in which her position
in society is arranged ¬ts with normal circumstances of politics. Instead
of giving up her reliance on the normal circuits and routines of civic
cooperation, she still trusts that she will be able to thematize “ through
the help of social movements, political representatives, the courts, and
so on “ her condition in a way that will eventually meet with a responsive
and self-critical attitude of her fellow citizens. Civic responsiveness is a
similarly ¬‚exible and generous disposition in face of disagreement. It
enables one to withstand the huge attraction of simply not re¬‚ecting on
the uncivil or even immoral consequences of one™s position of power
Political Autonomy and Agonistic Citizenship 265

in society. The responsive citizen is generous enough to admit, in light
of a critical evaluation of her civic ideals and vis-` -vis the burdens of
cooperation carried by some of her fellow citizens, that she cannot accept
the legitimacy of the very civic order that privileges her. Seen in this
way, the citizen™s disposition to act from civic responsiveness and civic
endurance is a sign of her civic integrity “ that is, of her testing her own,
others, and legally sanctioned civic interpretations in light of the ideal to
let the legitimate needs and interests of all citizens matter equally. This
civic integrity presupposes a considerable amount of self-restraint and
self-sacri¬ce.47 For the citizen who acts from the virtue of civic endurance,
it presupposes the willingness to remain patient and civil in her political
conduct even when the limits of loyalty are well in sight. For the citizen
who acts from responsiveness, it presupposes the willingness to live up to
insights that may result in her having to give up some of the privileges
she enjoys. Finally, where democratic majorities are not responsive to the
needs of minorities, it is not morally in order to expect these minorities
to exercise civic endurance. And where minorities do not want to endure
even perfectly reasonable burdens of citizenship, there is a clear limit
to the responsiveness they can expect from majorities “ a limit that is
de¬ned by the lack of the very disposition towards civic cooperation that
is a condition of all fruitful civic interactions.48
One last word on more speci¬c civic virtues. Where laws and regu-
lations that are supposed to protect the civil peace become subject to
interpretative con¬‚ict themselves, citizens and their political representa-
tives are often asked to remain tolerant, just, exercise self-restraint, loyalty,
law-abidingness, and so on. Given the importance of stability and social
order, this is understandable. However, such appeals to the exercise of
civic virtue remain quite empty if they do no make it clear what the call
for civic virtue implies for differently positioned citizens. A re¬‚ection on
the requirements of civic endurance and responsiveness can help answer
that question. The point is easily made through dramatic examples: dur-
ing racial riots, a call for tolerance implies very different concessions and
future courses of actions for different persons and groups involved; in
a situation of unequal distribution of basic goods such as food, water,
and medical care, calls for restraint and justice imply different courses of
action for different people. But if one thinks of less dramatic examples
from political struggles over socioeconomic and identity-related issues,
the same point goes for virtues such as courage, loyalty, law-abidingness,
perseverance, and so on. It is through our dispositions for civic responsive-
ness and endurance that we learn to exercise civic virtue and autonomous
Bert van den Brink

political judgment and action in ways that are appropriate to our own and
others™ situation as members of internally con¬‚icted political communi-
ties that are not, but often aim to be, well-ordered and just.

V Conclusion
I have argued that liberalism without agreement is a viable alternative to
variants of liberalism based on the assumption that a ¬rm agreement on
the requirements of a common framework for social and political coop-
eration is a necessary condition for civic cooperation. The need to make
this argument arises from the many serious doubts that liberal and other
political theorists have raised as to the appropriateness of agreement-
based liberalism in a pluralist age. In particular, an understanding of the
deeply pluralistic “circumstances of politics” (Waldron) provides us with
good reasons to leave agreement-based liberalism and its understanding
of the requirements of civic cooperation behind. I have shown that lib-
eralism without agreement is based on quite different ideas about the
legitimacy of the civic order, the nature of political autonomy, the use
of public reason, and civic virtue than agreement-based liberalism is. Fi-
nally, I have introduced two general civic dispositions “ civic endurance
and civic responsiveness “ and explained how these may help us to un-
derstand the requirements of the exercise of civic virtue and political
autonomy. Although the story I have told is quite general in nature, and
calls for further elaboration, I hope that it has made it clear that liberal-
ism without agreement may well be a viable alternative to the immensely
important, but in its theoretical self-understanding mistaken, tradition
of agreement-based liberalism.

1. I have bene¬ted from comments on earlier drafts of this chapter by mem-
bers of audiences in Frankfurt am Main, Tilburg, and Southampton. Special
thanks are due to Russell Bentley, Rainer Forst, Axel Honneth, Bart van Klink,
David Owen, Hildegard Penn, Morten Raffnsøe Møller, Sanne Taekema,
Bertjan Wolthuis, the editors of the present volume, and four anonymous
referees for Cambridge University Press.
2. Marilyn Friedman, “John Rawls and the Political Coercion of Unreasonable
People,” in: The Idea of a Political Liberalism: Essays on Rawls, ed. by Victoria
Davion and Clark Wolf (New York: Rowman & Little¬eld, 2000), pp. 16“33;
here: p. 31. Emphasis in the original.
3. John Rawls, Political Liberalism (New York: Columbia University Press, 1993),
p. 137.
Political Autonomy and Agonistic Citizenship 267

4. Rawls™s “hermeneutical turn” in Political Liberalism clearly serves the purpose
of developing a normative political theory for members of contemporary
Western, liberal constitutional democracies (or even narrower: the United
5. This is, of course, John Rawls™s method. The most fundamental account of
the role of consent in constitutional-democratic political thought is probably
J¨ rgen Habermas, Between Facts and Norms: Contributions to a Discourse Theory
of Law and Democracy, trans. by William Rehg (Cambridge, MA: MIT Press,
6. Joseph Raz, “Government by Consent,” in: Raz, Ethics in the Public Domain:
Essay in the Morality of Law and Politics (Oxford: Clarendon, 1994), pp. 339“
7. Most famously, of course, by Michael J. Sandel in Liberalism and the Limits of
Justice (Cambridge: Cambridge University Press, 1982).
8. One of the main problems with the notion of an overlapping consensus is
that although you and I may agree that liberty, equality, solidarity, equity,
and pluralism should be core values in our society, we may hold radically
different views about all these values. One politically liberal author who sees
this problem very clearly is J. Donald Moon, Constructing Community: Moral
Pluralism and Tragic Con¬‚icts (Princeton: Princeton University Press, 1993),
pp. 75ff.
9. James Tully, “The Agonic Freedom of Citizens,” Economy and Society, 28/2
(1999): 161“182; here: 170.
10. Jeremy Waldron, Law and Disagreement (Oxford: Oxford University Press,
1999), pp. 105“106. The quotation from Rawls is from “The Domain of
the Political and Overlapping Consensus,” in The Idea of Democracy, ed. by
D. Copp, J. Hampton, and J. Roemer (Cambridge: Cambridge University
Press, 1993), p. 246.
11. To be fair, it is not easy to determine to what extent Rawls thinks agree-
ment on the substance of, for instance, principles of justice, primary
goods, and the requirements of the burdens of judgment is important. It
seems that, over the years, agreement on substance became less important
for him, while a conscious use of public reason had become ever more
12. All quotations from Waldron, Law and Disagreement, p. 106.
13. See, for the metaphysical issues looming behind this assumption, Bernard
Williams, Shame and Necessity, Berkeley: University of California Press, 1993,
pp. 161ff. Williams may be said to give a genealogy of the historically con-
tingent core idea of Enlightenment political thought “. . . that the relations
of human beings to society and to each other, if properly understood and
properly enacted, can realize a harmonious identity that involves no real
loss” (p. 162).
14. See, on this point, Immanuel Kant, “On the Common Saying: That May Be
Correct in Theory, but It Is of No Use in Practice,” in: Kant, Practical Philosophy,
trans. and ed. by Mary J. Gregor (Cambridge: Cambridge University Press,
1996), pp. 296“297 especially.
15. Waldron, Law and Disagreement, p. 108.
Bert van den Brink

16. See, for a feminist analysis of the dangers involved, Marilyn Friedman™s
contribution to the present volume (Chapter 7), “Autonomy and Male
17. I stress the moral capacities since they are of crucial importance to both the
conception of political autonomy of the Rawlsian brand of agreement-based
liberalism I discuss and liberalism without agreement. See, for several ac-
counts of the importance of moral autonomy in liberalism in the present
volume, Rainer Forst, “Political Liberty: Integrating Five Conceptions of Au-
tonomy” (Chapter 10); Gerald Gaus, “The Place of Autonomy Within Liber-
alism” (Chapter 12); Jeremy Waldron, “Moral Autonomy and Personal Au-
tonomy” (Chapter 13); and John Christman, “Autonomy, Self-Knowledge,
and Liberal Legitimacy” (Chapter 14). Of course, other cognitive capaci-
ties relating to rationality, self-control, freedom from pathology, adequate
information, and motivational effectiveness of one™s beliefs are relevant to
the exercise of political autonomy as well. See Christman, Chapter 14 in the
present volume, Section I.
18. As to the question of authentic identi¬cation, I agree with John Christman
that in political reasoning, the authority of an individual citizen with respect
to her authentic beliefs is not of an “epistemic” but of a “personal” or “norma-
tive” nature (see John Christman™s Chapter 14 in the present volume, Sec-
tions III and IV). We accept the authority of someone expressing her “own”
beliefs not because we have reason to assume that she is the best judge of the
truth of what she believes, but because she is in the best position to express
her own value commitments, which she considers to be of importance to
public deliberations over political matters. The epistemic authority in these
matters falls to intersubjective deliberation over the appropriateness of her
beliefs. This is an important point with respect to a question posed by John
Rawls that will be discussed later “ namely, how autonomous political action
can be prevented from being too self-righteous.
19. John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press,
1971), p. 515.
20. Ibid., p. 516.
21. Ibid., p. 519. Similarly, in Political Liberalism, p. 77, Rawls states that it “. . . is
in their public recognition and informed application of the principles of
justice in their political life, and as their effective sense of justice directs, that
citizens achieve full [political] autonomy. Thus, full autonomy is realized by
citizens when they act from principles of justice that specify the fair terms of
cooperation they would give to themselves when fairly represented as free
and equal persons.”
22. See, for the several aspects of autonomy mentioned here, Jeremy Waldron™s
reconstruction of autonomy in Section III of his contribution to the present
volume, “Moral Autonomy and Personal Autonomy” (Chapter 13). His re-
construction there focuses on similarities between the notions of personal
and political autonomy, especially the importance of moral autonomy to
both these notions. I agree with him that in political theory, a radical sepa-
ration between personal and political autonomy cannot be made. See also
Bert van den Brink, The Tragedy of Liberalism: An Alternative Defense of a
Political Autonomy and Agonistic Citizenship 269

Political Tradition (Albany, NY: State University of New York Press, 2000),
Part 2, where I argue that John Rawls™s notion of political autonomy presup-
poses that citizens orient themselves to certain moral values even in their
personal, “pre-political” lives.
The change in position is most obvious in “The Idea of Public Reason Re-
visited,” in: The Law of Peoples, Cambridge, Mass.: Harvard University Press,
See note 21.
Isaiah Berlin, “Two Concepts of Liberty,” in: Four Essays on Liberty (Oxford:
Oxford University Press 1969), pp. 118“172.
Charles Taylor, “What™s Wrong with Negative Liberty,” in: Philosophy and the
Human Sciences. Philosophical Papers Vol. 2 (Cambridge: Cambridge University
Press, 1985), pp. 211“229.
Philip Pettit, Republicanism: A Theory of Freedom and Government (Oxford:
Clarendon Press, 1997).
Tully stresses this game element of the agonistic understanding of politics in
“Agonic Freedom,” passim.
John Gray, “Agonistic Liberalism,” Enlightenment™s Wake (London: Routledge,
1995), p. 68.
Hannah Arendt, “What is Freedom?” in: Arendt, Between Past and Future
(Harmondsworth: Penguin, 1985), p. 165. See Michel Foucault, “Truth and
Power,” in: Foucault, Power/Knowledge: Selected Interviews and Other Writings,
1972“1977, ed. by Colin Gordon (New York: Pantheon Books, 1980),
pp. 109“133; here: 121. See Tully, “Agonic Freedom” for a detailed account
of the similarities and differences of Arendt™s and Foucault™s accounts of
agonistic freedom.
See note 18.
John Tomasi, Liberalism Beyond Justice (Princeton: Princeton University Press,
2001), p. 54.
See Arendt, “What is Freedom?” pp. 165ff.; Tully, “Agonic Freedom,” p. 162.
Jeremy Waldron, “Moral Autonomy and Personal Autonomy,” Chapter 13 in
the present volume (p. 307).
Rawls, Political Liberalism, p. 226: “Accepting the idea of public reason and its
principle of legitimacy emphatically does not mean . . . accepting a particular
liberal conception of justice down to its last details of the principles de¬ning
its content. We may differ about these principles and still agree in accepting a
conception™s more general features. We agree that citizens share in political
power as free and equal, and that as reasonable and rational they have a
duty of civility to appeal to public reason, yet we differ as to which principles
are the most reasonable basis of public justi¬cation.” In Political Liberalism,
Rawls seems torn between this view and what is probably his older and more
static view that a well-ordered society “is a society in which everyone accepts,
and knows that everyone else accepts, the very same principles of justice . . . ”
(p. 52).
See Russell Bentley and David Owen, “Ethical Loyalties, Civic Virtue and
the Circumstances of Politics,” Philosophical Explorations 4 (2001): 223“239;
Friedman, “Rawls and Political Coercion”; J. Donald Moon, Constructing
Bert van den Brink

Community; David Archard, “Political Disagreement, Legitimacy, and Civility,”
Philosophical Explorations 4 (2001): 207“222; van den Brink, The Tragedy
of Liberalism.
Christman, “Autonomy, Self-Knowledge, and Liberal Legitimacy,” Chap-
ter 14, Section III, in the present volume.
See, for instance, the literature mentioned in note 36.
Cf. Diana T. Meyers™ distinction between self-de¬nition and self-discovery
in her contribution to the present volume, “Decentralizing Autonomy: Five
Faces of Selfhood,” Chapter 2, Section V.
For a similar point, see David Owen, Nietzsche, Politics and Modernity (London/
Thousand Oak, CA: Sage, 1995), pp. 142ff.
Stephen Macedo, Liberal Virtues: Citizenship, Virtue, and Community in Liberal
Constitutionalism (Oxford: Clarendon Press, 1990), pp. 275“276; William
A. Galston, Liberal Purposes: Goods, Virtues, and Diversity in the Liberal State
(Cambridge: Cambridge University Press, 1991), pp. 221ff.
To my knowledge, Bentley and Owen, in their essay “Ethical Loyalties, Civic
Virtue and the Circumstances of Politics” (note 36), are the only other au-
thors who use those terms in roughly the same way as I do. I thank them for
acknowledging the in¬‚uence of my lecture “Endurance as a Civic Virtue”
(University of Southampton, December 2000) on their work. In return, I
want to acknowledge the way in which their discussion of civic endurance
and responsiveness has shaped the ideas I present here, especially the sugges-
tion to delimit what they call the “virtues” of endurance and responsiveness
from attendant vices.
For quite different analyses of the exact forms of repression and lack of po-
litical voice see, for instance, James Tully, Strange Multiplicity (Cambridge:
Cambridge University Press, 1994) and Will Kymlicka, Multicultural Citizen-
ship, (Oxford: Oxford University Press, 1995).
See also Bentley and Owen, “Ethical Loyalties,” p. 236.
I cannot go here into the interesting parallels between responsiveness as a
civic disposition and as a virtue of “responsive” law, as developed by Philip
Selznick in The Moral Commonwealth: Social Theory and the Promise of Community
(Berkeley: University of California Press, 1993), pp. 463“471.
See Philip Selznick, “Civilizing Civil Society,” in: Anton van Harskamp and
Albert W. Musschenga, eds., The Many Faces of Individualism (Leuven: Peeters,
2001), pp. 171“185.
Although I cannot argue for this here, I think that my account of endurance
and responsiveness ¬ts with some of the normative consequences that seem
to follow from Axel Honneth and Joel Anderson™s quest for an understanding
of civic freedom and equality in terms of a multivocal model of interpersonal
communication and mutual recognition. With the help of the notions of civic
endurance and responsiveness, I try to gain insight into situation-dependent
forms of concrete mutual recognition that can help restore the multivocality
and effective freedom and equality of the members of a pluralistic and less
than just political community. See Anderson and Honneth™s contribution
Political Autonomy and Agonistic Citizenship 271

to the present volume, “Autonomy, Vulnerability, Recognition, and Justice”
(Chapter 6). For a partly recognition-theoretic argument regarding civic re-
spect among members of majority and minority cultures see my “Democratic
Reasoning and Reasons of One™s Own,” in Maureen Sie, Marc Slors, Bert van
den Brink (eds.), Reasons of One™s Own (Aldershot: Ashgate, 2004), pp. 69“

The Place of Autonomy within Liberalism

Gerald F. Gaus

I Introduction
My concern in this chapter is the place of autonomy within liberalism,
understood as a public morality.1 To what extent is liberal morality neces-
sarily committed to some doctrine of autonomy, and what is the nature of
this doctrine? I begin (Section II) by brie¬‚y explicating my understand-
ing of liberalism, which is based on the fundamental liberal principle “ that
all interferences with action stand in need of justi¬cation. Section III
then defends my ¬rst core claim: given a certain compelling view of the
nature of moral reasons, the fundamental liberal principle presupposes
a Kantian conception of morally autonomous agents. I then consider
(Section IV) an implication of the fundamental liberal principle when
applied to public morality and the law “ that an interference with liberty
must be justi¬ed to everyone. This public justi¬cation principle, I argue,
constitutes a version of Kant™s categorical imperative; thus liberalism is
committed to not only autonomy of the will (Section III) but a substantive
morality of autonomy. By the end of Section IV, I will have shown that
liberal morality is committed to what may be broadly deemed a “Kantian”
conception of moral autonomy.
In Section V, I show how this necessary presupposition of moral au-
tonomy in liberal public morality implies a further commitment to one
interpretation of the much-discussed ideal of “personal autonomy.” It is
often maintained that the ideal of personal autonomy is independent of
moral or “Kantian autonomy”: the commitment to one is said not to entail
a commitment to the other. Kantian autonomy is understood as a meta-
physical idea concerned with free will, or more generally a presupposition

The Place of Autonomy within Liberalism 273

of the very possibility of moral responsibility, while personal autonomy is
typically understood as a character ideal, focusing on the value of critical
self-re¬‚ection on one™s desires, values and plans, or the value of choos-
ing one™s way of life for oneself, or perhaps the value of self-control.2
To be sure, most acknowledge that Kantian moral autonomy and per-
sonal autonomy are in some way related “ after all, they both go by the
label of “autonomy.” Both are about self-direction or self-government.3
Nevertheless, most advocates of what we might call “liberal autonomy” “
according to which the justi¬cation of liberalism is grounded on the no-
tion of personal autonomy “ seem more intent on distinguishing the two
ideas of autonomy than showing their connections. Section V challenges
this: the case for personal autonomy, I argue, derives from the case for
moral autonomy. They are distinct, but by no means independent, no-
tions of autonomy “ a position not unlike Ranier Forst™s in Chapter 10
of the present volume. My claim is, I think, less radical than the thesis
Jeremy Waldron defends in Chapter 13: while I wish to stress the connec-
tions between moral and personal autonomy, I nevertheless rely on the
distinction between them.4
Insofar as a commitment to autonomy is bound up with liberal pub-
lic morality, many liberal autonomists appear to think that it is personal
autonomy that is really crucial. Distinguishing personal autonomy from
Kantian autonomy is typically part of a project claiming that a liberal polit-
ical morality can be based on the former.5 By and large, those who would
construct liberal political morality on autonomy seek to build on personal
autonomy.6 Thus, for example, one commentator tells us that although
“Kant™s strong and metaphysically controversial conception of autonomy”
seems unable to “play the role of providing a suf¬ciently non-sectarian
basis for liberalism,” those “conceptions connected with the value of self-
re¬‚ection” are much more widely accepted, and may well provide the ba-
sis of non-sectarian liberalism.7 To be sure, some who advance personal
autonomy justi¬cations of liberal morality and the liberal state give at
least a passing acknowledgment to Kant™s conception of autonomy.8 More
importantly, Rawls is, on the whole, an exception to this common privi-
leging of personal autonomy: Kantian autonomy, understood as a type of
moral power, plays a fundamental role in Rawls™ liberalism (though it is
certainly also true that he moved away from a Kantian “comprehensive”
view as his political liberalism evolved).9 In any event, I aim to show in
this chapter that though Kantian and personal autonomy are related,
and commitments to both are part of liberal public morality, the Kantian
notion is more fundamental than the ideal of personal autonomy.
Gerald F. Gaus

II The Fundamental Liberal Principle
Stanley Benn asks us to

Imagine Alan sitting on a public beach, a pebble in each hand, splitting one
pebble by striking it with another. Betty, a causal observer, asks him what he is
doing. She can see, of course, that he is splitting pebbles; what she is asking
him to do is to explain it, to redescribe it as an activity with an intelligible point,
something he could have a reason for doing. There is nothing untoward about her
question, but Alan is not bound to answer it unless he likes. Suppose, however, that
Betty had asked Alan to justify what he was doing or to give an excuse for doing it.
Unlike explanations, justi¬cations and excuses presume at least prima facie fault,
a charge to be rebutted, and what can be wrong with splitting pebbles on a public
beach? Besides, so far as we can tell, Alan is not obliged to account to Betty for his
actions. . . .
Suppose Betty were to prevent Alan from splitting pebbles by handcuf¬ng him or
removing all the pebbles within reach. Alan could now quite properly demand a
justi¬cation from Betty, and a tu quoque reply from her that he, on his side, had
not offered her a justi¬cation for splitting pebbles, would not meet the case, for
Alan™s pebble splitting had done nothing to interfere with Betty™s actions. The
burden of justi¬cation falls on the interferer, not on the person interfered with.
So while Alan might properly resent Betty™s interference, Betty has no ground
for complaint against Alan.10

Benn observes a basic asymmetry between acting and interfering with
the actions of another. Alan does not have to justify his pebble splitting
to Betty: he is under no standing requirement to show Betty that he has
good reasons for what he is doing. On the other hand, it is required
of Betty that she justify to Alan interfering with his actions, or stopping
him from what he is doing, or in some way restricting his actions. This
is essentially what Joel Feinberg has called the “presumption in favor of
liberty”: “liberty should be the norm, coercion always needs some special
The liberal tradition in moral and political philosophy maintains that
each person has a moral claim to do as he wishes until some justi¬cation
is offered for limiting his liberty.12 As liberals see it, we have liberty to act
as we see ¬t unless reason can be provided for restriction. Call this the
fundamental liberal principle:
1. A person is under no standing obligation to justify his actions;
2. Interference with, or restriction of, an other™s action requires
justi¬cation; unjusti¬ed interference or restriction is unjust, and
so morally wrong.
The Place of Autonomy within Liberalism 275

The presumption underlying the liberal principle is essentially justi-
¬catory: it regulates justi¬catory discourse about the morality of action,
and ties moral wrongness to the lack of required justi¬cation. (This is not
to say that this justi¬catory presumption in favor of liberty itself does not
have to be defended; that indeed would be a question-begging error.) It
matters greatly, then, on whom the onus of justi¬cation is placed: who
must bear the justi¬catory burden? As Benn says, “justi¬cations and ex-
cuses presume at least prima facie fault, a charge to be rebutted.” If I have
no justi¬catory burden I am permitted to act without justi¬cation, for I
have no charge to rebut, no case to answer. If the onus is on you, the fail-
ure to justify condemns your act. Conceivably, a conception of morality
might place the onus on the actor: “never act unless one can meet the
justi¬catory burden by showing that one is allowed to act.” The liberal
insists that moral persons have no such general burden to bear, though
of course they may in special contexts in which a restriction already has
been established (say, in contexts of trusteeships). Thus, unless you oc-
cupy a special role, such as a trustee, if I object to what you are doing, it
is of no avail to demand, “Show me why you should be allowed to act.” As
Locke said, all men are naturally in “a State of perfect Freedom to order
their actions . . . as they see ¬t . . . without asking leave, or depending upon
the Will of any other Man.”13 My objection must take the form of a claim
that your action is immoral, or inconsiderate, or dishonorable “ you must
answer the case that your act is not eligible. But I bear the justi¬catory
burden of establishing this case. Ranier Forst objects that, contrary to the
fundamental liberal principle, “there is a standing duty to justify morally
relevant actions” (see Chapter 10 of the present volume, fn 24). How-
ever, by the time we have established that an action ± is morally relevant
we have, ex hypothesi, justi¬ed a limitation on freedom (or, alternatively,
a case to be answered). That there is a justi¬ed moral rule prohibiting or
regulating ± implies, of course, that the justi¬catory burden has already
been met. Morality, for the liberal, is as much in need of justi¬cation as
any other restriction on action,14 but once justi¬ed, moral prescriptions
shift the onus back to the agent (he now has a case to answer), as Forst
rightly observes.
My main aim in this chapter, however, is not to defend the funda-
mental liberal principle (but see Section Va); rather, I seek to examine
its presuppositions. More precisely, I am concerned with what must be
the case about reasons for actions and agents if the fundamental liberal
principle is to serve as a moral principle for governing social life.
Gerald F. Gaus

III The Liberal Principle and Moral Autonomy
IIIa What Reasons Do We Have? The Radical Instrumentalist Model
My concern, then, is the sorts of agents and their reasons that are pre-
supposed by the fundamental liberal principle. What must be true for
the fundamental liberal principle to be the basic moral principle? To be-
gin, suppose that all reasons for action are instrumental reasons.15 The
core of the instrumental model is the intuition that in rational action, an
agent necessarily seeks the best available result, with “best” being under-
stood in terms of what she cares about, her goals, her purposes, and so
on. This is the idea behind one conception of rational action qua util-
ity maximization, which is often taken to be much the same as saying
that an agent has “purposes that her action is designed to promote.”16
I shall follow Robert Nozick in taking the idea of goal pursuit as the
core of instrumental rationality; indeed, as Nozick observes, “it is nat-
ural to think of rationality as a goal-directed process.” So according to
the basic “instrumental conception, rationality consists in the effective
and ef¬cient achievement of goals, ends, and desires. About the goals
themselves, an instrumental conception has little to say.”17 I explicitly do
not refer to “preferences” here, as the idea of a preference is ambigu-
ous between something akin to a goal, purpose, or end (in which case
“preference” would be suitable) and something akin to an overall reason
for action, in which case it is axiomatic that all reasons for action are in-
tended to satisfy preferences (which is a broader idea than instrumental
Elsewhere I have speci¬ed this instrumental model in more detail.18
For now, let us work with a straightforward formulation:

Instrumental Rationality: Betty has a (good) purely instrumental reason to β if and
only if given her option set, β best secures her goals (ends, etc.).19

Therefore, if Betty performs some alternative action β — , β — cannot be
justi¬ed by appeal to instrumental rationality.
Suppose, then, not only a world in which each is always guided by,
and only by, her instrumental rationality, but a conceptual world in
which there simply is no other type of reason for action. The only rea-
son for action that anyone ever has or could have, given a set possi-
ble acts, is a reason to do that which best promotes her goals, achieves
her ends, and so on. Many think we live in such a world: they are con-
vinced that instrumental rationality subsumes all of rationality. This is the
world of orthodox rational-actor theory and, through that theory, many
The Place of Autonomy within Liberalism 277

moral theories, such as David Gauthier™s.20 It is, I shall argue, an illiberal

IIIb The Basic Case
Suppose in this world of purely instrumental reasons, Betty interferes
with Alan™s actions. Betty is a successful predator, and manages to force
Alan to do what she wants. She gains; he loses. She does not seek to
justify her actions to him, nor is she concerned that he is harmed; her
instrumental reasons unambiguously instruct her to invade. Given all she
cares about, the act “invade Alan” best promotes her goals, so she follows
her best reasons and invades. This is a manifest injustice; it is a considered
judgment of liberals that Betty does wrong. Our question is this: in such
a world, can Alan invoke the basic liberal principle, insisting that unless
Betty justi¬es her intervention, she acts unjustly? If he cannot do so, then
the fundamental liberal principle presupposes some world other than
the world of purely instrumental reasons.

IIIc The Rejection of a Radical Externalist Account of Moral Obligation
For Alan to sensibly invoke the fundamental liberal principle in the
world of purely instrumental reasons, it would have to be possible for
him to claim that Betty™s unjusti¬ed invasive action is ipso facto wrong,
even though she has no reason to refrain from her invasion. That is, he
(and we) would have to accept:

Radical Externalism: Betty can have a moral obligation to refrain from act β even
if there is no reason for her not to β.21

Radical Externalism should be rejected: it denies a necessary connection
between a moral obligation to not β and a reason to not β. If Radical
Externalism holds, then even if Betty has no reason at all to refrain from
invading Alan, she still can have a moral obligation to refrain; the moral
obligation itself provides no reason for her to refrain. It should be stressed
that Radical Externalism is indeed a truly radical form of externalism. It
goes considerably further than would those externalists who would insist
that Betty can have a moral obligation to refrain from β even if she does
not have a motivating reason to refrain from β (for an externalist can
admit that her lack of motivating reason to β does not itself show that there
is no reason for her to β; see, however, Section IIIe). Indeed, Radical
Externalism goes beyond the typical externalist claim that Betty can have
a moral obligation not to β even if there is no way that, given her epistemic
Gerald F. Gaus

position, desires, and so on, she could have access to a reason not to β.22
In fact, most externalists, as well as all internalists, deny Radical External-
ism “ that is, they deny that Betty can have a moral obligation to refrain
from β when there is no reason whatsoever for her to refrain from β.
They thus accept:

Modest Internalism: Betty has a moral obligation to refrain from β only if there is
a reason for her not to β.

Strong conceptions of externalism are consistent with accepting Modest
Internalism (which goes to show just how modest a form of internalism it
is). On these more plausible externalist accounts, just as there can exist a
moral obligation whether or not a person knows about it or is motivated
to act on it, so too can there be a reason to act on this obligation whether
or not a person is aware of it, or is motivated to act on it.23
Radical Externalism denies, to use Michael Smith™s term, a “platitude”
about morality: that morality is part of practical reason in at least the
weak sense that an ideally rational agent, who was aware of all the reasons
for action that there are, would necessarily have reasons to act on her
moral obligations.24 To accept Radical Externalism is to hold that correct
moral judgments need not imply reasons of any sort “ motivating or
otherwise “ to do anything about them. Of course, some do think this.
Radical expressivists seem to believe that moral judgments are simply
affective expressions that have no tie to practical reason; radical realists
conceive of moral judgments as simply claims about certain moral facts
that have no implications for what agents have reason to do. But these
are strange views, which are not even embraced by most expressivists and
realists. If morality is not about what agents have reasons to do, it is hard
to understand what it is about.
Now if we accept, as we should, Modest Internalism, Alan “ in our
world of purely instrumental reasons “ cannot coherently claim that Betty
is under a moral obligation or duty to refrain from interfering with him
without justi¬cation. More simply, he cannot claim that Betty acts wrongly
in our case, for, ex hypothesi, Betty™s only reasons are to advance her own
goals, and these are reasons that unambiguously endorse invading him.
Thus, according to Modest Internalism, if Betty has no reason to refrain
from invading, she cannot have an obligation to refrain, and so she does
not act wrongly. If that is so, we have a contradiction: the fundamental
liberal principle deems Betty™s act wrong, but on the supposition that all
reasons are instrumental reasons, she does not act wrongly. Given Modest
The Place of Autonomy within Liberalism 279

Internalism, we need to give up either the fundamental liberal principle
or the purely instrumental view of practical reason.

IIId A Challenge to an Assumption: The Convergence Thesis
Some may seek to remove the apparent contradiction by challenging
the assumption that, all things considered, instrumental reasons endorse
Betty™s unjusti¬ed invasion. Following Hobbes, a number of contempo-
rary moral and political philosophers have argued that agents such as
Betty would ¬nd themselves in intractable con¬‚ict, which would frustrate
the pursuit of their goals, and so our assumption is false. The contradic-
tion, then, might be said to depend on a false assumption that instru-
mental reasons endorse unjusti¬ed interference.
For this reply to be effective there must be no case in which instru-
mental reason instructs Betty to wrongly invade Alan simply to effectively
advance her goals. If there is just one case in which such predation is
instrumentally rational, the inconsistency we have been discussing arises.
Thus we need an argument for the universal convergence of instrumental
reason and applications of the liberal principle. It seems pretty doubtful
that a successful argument along these lines will be forthcoming, but
let us grant the convergence assumption. Let us suppose that a project
such as Gauthier™s succeeds in showing that, given facts about human
society and human nature in world W, for all individuals in W of purely
instrumental reasons, it will always be the case in W that one will have
reason to refrain from interference (unless that interference can be jus-
ti¬ed to the person being interfered with). This, though, would still only
show that within W the contradiction would not arise. But our concep-
tion of liberal morality is not limited to W. Our understanding of morality
commits us to some (I am assuming for now) counterfactual judgments
(think of Judith Thomson™s trolley cases, or her violinist case).25 That
those cases do not arise in our world does not show that our moral con-
cepts need not apply to them. Now so too with the fundamental liberal
principle: even if we live in W, the principle covers at least proximate
possible worlds, including those in which instrumental reasons lead us
to invade others without justi¬cation. If in these counterfactual cases
the contradiction arises, we can conclude that our understanding of the
liberal principle still presupposes that not all reasons are instrumental
reasons. The conceptual point about the presuppositions of the funda-
mental liberal principle thus cannot be met by showing that there exists
a world W in which the problem does not arise, even if we happen to
live in W.
Gerald F. Gaus

To be sure, as we entertain more and more outlandish counter-
factuals “ consider possible worlds that are further and further from our
own “ our concepts loose their grip, our ability to apply them becomes
attenuated, and we become confused about what to say. This is certainly
a severe problem with many of the so-called “mental experiments” de-
signed to “test” our “moral intuitions,” or, more accurately, the criteria
for applying moral notions. In these worlds of incredible machines where
hitting one button or the other causes amazing chains of events, our
normal concepts are apt to leave us unsure about what to say.26 Surely,
though, that is not a relevant objection here. Our empirical world is one
in which what best advances people™s goals often enough con¬‚icts with
refraining from interfering with others; it is the assumption of universal
convergence that pushes us into unfamiliar territory.
This criticism of the convergence assumption could be avoided by
showing that the tie between the fundamental liberal principle and
agents™ goals is not contingent. Drawing on a theory of value, it might
be suggested, say, that because everyone™s true goal is to respect others,
and because the fundamental liberal principle is an expression of this
(or, perhaps, a means to it), in all relevant possible worlds “ those with
the correct theory of value “ there will always be an instrumental reason
not to unjusti¬ably interfere with others. Serious problems confront this
proposal. As I have argued elsewhere, goal-based and principled-based
reasons are not the same, nor can one be reduced to the other.27 If this is
so, then converting the fundamental liberal principle into a goal would
not account for the types of reasons it implies. I will not, however, in-
sist on this somewhat complicated point here. For our present purposes,
we can reject the suggestion as it is clearly inconsistent with liberal the-
ory. Liberalism denies that each of us has the overriding goal of being a
good liberal, or that our overriding goal is to abide by liberal principles.
(It should be stressed that for the present suggestion to work, abiding
by the liberal principle must be one of our highest ranked goals, capa-
ble of giving instrumental reasons to forgo our other goals by refraining
from interfering with others.) Although liberals do indeed insist that in-
dividuals are capable of putting aside their various goals to abide by the
principle of non-interference, this is not because they believe that our
primary goal is not to interfere with others. Liberals conceive of individ-
uals as possessing a diverse array of goals and ends; they do not “ certainly
need not “ advance a theory of value according to which an overriding
goal of everyone is to abide by the fundamental liberal principle.
The Place of Autonomy within Liberalism 281

IIIe Acting upon Reasons for Action: A Standard Internalist
Claim About Reasons to Act
We must conclude that, given Modest Internalism, the fundamental lib-
eral principle is incompatible with a world in which all reasons to act
are instrumental. Liberalism supposes that there are moral reasons (to
refrain from interfering with others) that are not ways to achieve goals.
Now to accept that there exists a reason R to refrain from β commits
one to also accepting that, supposing no competing reasons outweigh R
(are more important than R, rank higher than R, and so on28 ), a rational
agent who is aware of R will refrain from β on the grounds of R, or because
of R.29 Reasons for action are part of practical rationality, and practical
rationality guides the action of rational agents. A form of internalism that
goes beyond Modest Internalism (Section IIIc) about reasons for action
is compelling: there is an internal “ necessary “ connection between R™s
status as a reason and R™s being acted upon. Let us call this:

Standard Internalism: If R is a reason to refrain from act β, it must be the case that,
barring overriding reasons, a rational agent who is aware of R will refrain from β
because of R.

Philippa Foot apparently rejects this; as she sees it, “an agent may fail to
be moved by a reason, even when he is aware of it.”30 On her view, one can
be aware that R is a reason to β, and yet not β. Now of course this can be
the case if the agent is characterized by a failure of practical rationality;
what is called “weakness of will” can be understood as a failure to act
on one™s best reasons.31 However, one who fails to be moved by the best
reasons for action of which she is aware always suffers from a defect of
rationality: a practically rational person™s actions are guided by her rea-
sons. This is not merely asserting the de¬nition that one is practically ra-
tional if and only if one is moved by one™s reasons. We possess an implicit
concept of rational agents,32 and according to this concept, someone who
asserts that “Yes, R is a reason to β that applies to me in the present context,
but what does that have to do with my actually being moved to β?” does not
understand what it means to say that R is a reason for action. “Yes, I have a
reason not to steal, but what does this have to do with my actually refrain-
ing from stealing?” is not to exploit a distinction in our understanding of
reasons for action and motivation; it expresses conceptual confusion.33
(To make it intelligible, we might suppose that the speaker is claiming that
though R is typically taken as a reason, she is actually denying it.) It is
thus mistaken to assert that one may, without inducing conceptual
Gerald F. Gaus

puzzlement, claim that one just happens to be unmoved by one™s rea-
sons, say, because one lacks the desire to be rational.34 Once we have
established that a person acknowledges that R is her best reason, and it is
a reason to β, we do not need additional premises to explain her β-ing.
Indeed, her not β-ing is what calls for further explanation: we are apt to
invoke a special account of breakdowns of rationality to make not β-ing

IIIf Moral Autonomy as a Property of the Rational Will
We thus arrive at our ¬rst conclusion: given a plausible internalist concep-
tion of reasons for action, the fundamental liberal principle presupposes
that there are reasons for agents to set aside their instrumental reasons
and abide by the principle (Sections IIIa“d), and that (when they are
the best reasons) rational agents act on these reasons (Section IIIe). This
is to endorse a Kantian “ though, of course, not Kant™s “ conception of
moral autonomy.
As in Kant™s view, autonomy is analytically connected with practical
reason. As Kant understood it, to attribute autonomy to an agent just
is to attribute to her the capacity to be moved by a practical principle,
endorsed by practical reason, which does not make reference to her needs
or interests.35 To be autonomous is to have the capacity for one™s will to be
determined by moral practical reason.36 Autonomy, then, is a property
of the will. Our analysis of the presupposition of claims based on the
fundamental liberal principle has led us to conclude that the principle is
intelligible only if individuals have the capacity to be guided by practical
reasons that do not derive from promoting their goals, ends, and so on.
Again following Kant, only because we are cognizant of the demands of
morality do we know that we are able to be guided by reasons for action
that do not derive from furthering our goals or ends “ that is, we possess
moral autonomy.37
Susan Wolf seems to pose an objection. She has argued that the ability
to act on reason is to be distinguished from a conception of freedom as
autonomy.38 As I understand her, she would depict the view endorsed
here as a “Reason View,” not an “Autonomy View,” of moral responsibil-
ity. Whereas an Autonomy View locates moral freedom and responsibility
in one™s option to do or not do one™s moral duty, for the Reason View
“[w]hat matters is rather the availability of one very particular option,
namely the option to act in accordance with Reason.”39 Certainly my po-
sition has much in common with Wolf™s Reason View. And it is certainly
The Place of Autonomy within Liberalism 283

true that claims about counter-causal freedom are important in Kant™s
thinking about autonomy. The argument presented here is silent about
such freedom, so it does not capture all of what Kant meant to include in
the concept of “autonomy.” Acknowledging all that, it should be noted
that interpretations of Kant, and Kantian conceptions of morality, differ
in the ways they relate the will to reason, and the relative priority they as-
sign to one or the other. For example, in contrast to my account, Christine
Korsgaard appears to give a much more central place to the idea of the
will; in some ways, reason seems secondary to the idea of the will in her
interpretation.40 The Kantian view defended here, in contrast, takes as
central to Kant™s conception of autonomy the idea that an autonomous
will is one determined by moral reason, and that we are free when we
act rationally in this way. That, in my view, is the central feature of the
concept of moral autonomy.
It is worth stressing just how important to Kant™s understanding of
autonomy is what we might call the “metaphysics of reasons.” Kant dis-
tinguished between, on the one hand, reasons of morality and, on the
other, reasons that might be variously described as those of prudence,
reasons concerned with one™s subjective interests as a sensuous being, or
reasons of self-love.41 As is commonly observed, Kant conceived of the
latter category too narrowly: we need not suppose that the reasons po-
tentially opposed to moral reasons are necessarily sel¬sh or self-centered.
A more adequate contrast is between reasons devoted to pursuing that
which we see as good, and so endeavor to obtain (valued states of affairs,
cherished objects, goals, ends), and those moral reasons that demand we
set aside our pursuit of the good or valuable.42 In contrast to instrumental
reasons, moral reasons do not confront us as hypothetical, because they
do not depend on our af¬rmation of a goal or an end. They confront
us as imperitival and categorical.43 Regardless of our ends or goals, they
demand that we do the right thing.
I follow Kant in distinguishing heteronomous moralities from au-
tonomous moralities.44 Attempts to derive the moral from the rational,
where the latter is understood simply in terms of instrumental rationality,
are heteronomous.45 On such views, morality is simply a device for ef¬-
cient goal pursuit. Such moral theories are ultimately unsuitable; their
denial of autonomy renders them at a loss as to how to account for our abil-
ity to refrain from pursuing our concerns and values, and our demands
that others do so as well. Rational agents approach being psychopathic
when their reasons are consumed by their own ends.46
Gerald F. Gaus

IIIg The Moral Autonomy of the Claimant
Thus far I have argued that in appealing to the fundamental liberal prin-
ciple, Alan necessarily supposes that Betty possesses moral autonomy. Can
Alan, though, insist that while he supposes that Betty possesses moral au-
tonomy, he is simply an instrumental reasoner, and so non-autonomous?
No: not only must Alan suppose that Betty is an autonomous agent, he
must also suppose that he possesses moral autonomy.
To continue our example, assume that Alan invokes the fundamental
liberal principle against Betty. We have seen that he supposes that she
possesses moral autonomy, in the sense that she can act on moral reasons
rather than on her goal-based ones. The basic liberal principle, though,
does not prohibit all interference; it puts the onus of justi¬cation on
Betty, who would interfere with Alan. Now assume that she meets this
burden. Betty offers a justi¬cation for interfering with Alan™s act ± of the
form: “Reason R justi¬es a moral prohibition of your act ±; you ought
not to ±, and if you seek to ±, I may legitimately β “ that is, stop you.”
What can we say about the nature of this reason?
1. Well, it could be claimed that for the true liberal, there really are
no such reasons as R purports to be. One might think that the genuinely
liberal view is that it is never permissible to interfere with a person™s lib-
erty. Call this the absolutist interpretation of the fundamental liberal principle:
the onus of justi¬cation can never be met. On the face of it, the abso-
lutist interpretation appears too strict, as it apparently prevents liberals
from endorsing a right to private property, or rights to bodily integrity.
It would seem that the liberal would want to claim that it is justi¬able
when I interfere with your liberty by asserting my rights to private prop-
erty or bodily integrity. If you are using your liberty to hit me on the
head, or steal my acorn, it would appear that the liberal must think in-
terference with your liberty is justi¬able. Yet, if the absolutist interpreta-
tion excludes interference with liberty, it would appear to exclude such
liberal rights and their defenses. Some seek to rescue the absolutist in-
terpretation from this criticism, though, by insisting that one™s property
rights de¬ne one™s liberty rights. As Jan Narveson argues, you own your
eyes, and that is why they cannot be removed, and because you own your
arm, it is up to you to decide whether to lift it or not. To be free to do
something is just to be free to use what is yours “ your property; so all
your freedom rights concern your property. Indeed, Narveson claims,
“it is plausible to construe all rights as property rights.”47 If so, then a
person™s liberty is interfered with if and only if his property rights are
The Place of Autonomy within Liberalism 285

Not only does this absolutist interpretation depend on the identi¬ca-
tion of liberty with property rights “ which, I think, can be shown to be
implausible48 “ but, in addition, it requires a claim about the compossi-
bility of property-liberty rights.49 If it were possible for my valid property
right to X to con¬‚ict with your valid property right to Y, then somebody
must interfere with someone else™s property right (and, so liberty); but
since, on the absolutist interpretation, there are no reasons that could
justify an interference, it would follow that someone must do wrong.50
I shall not pursue this option further. Unless one can show that all lib-
erty rights are property rights and that property rights are compossible
(or else accept that in some cases wrongdoing is unavoidable) “ and I
do not believe these can be shown “ the absolutist interpretation is not
2. Betty might justify a prohibition of Alan™s ± on paternalistic
grounds, claiming that her present interference can be justi¬ed because
it better promotes Alan™s own values, goals, projects, and so on. That is,
Betty might appeal simply to Alan™s instrumental reasons. Now if Betty
takes this route, and shows that Alan™s instrumental reasons endorse the
prohibition, she advances a paternalist justi¬cation; the justi¬cation of
the prohibition of ± (and/or her interference, β) is that it advances
Alan™s goals.51 This is worthy of note. If we accept (a) a world of purely
instrumental reasons along with (b) Modest Internalism, the only justi-
fying reason we could give another for limiting his freedom would be
paternalistic. Suppose the convergence thesis held (Section IIId): moral
principles are justi¬ed on the ground that everyone™s instrumental rea-
sons support following them. Everyone™s goals are advanced by following
the principles. Now suppose that Betty seeks to restrict Alan™s liberty by
appealing to these principles; she wants to claim that he is acting wrongly,
and so morality justi¬es stopping him. Given Modest Internalism, her jus-
ti¬cation must be that his reason for accepting her interference is that
his own ends are advanced by the interference. Thus, in a Gauthierish
moral world, all justi¬cations meeting the burden of the fundamen-
tal liberal principle collapse into paternalistic reasoning. To be sure, it
would also be the case that the regulation would advance Betty™s goals
(ex hypothesi, which is why she has a reason to β), but her claim that Alan™s
action is wrong, and so that he has a reason to refrain from ±, must be
a claim that his reason for not ±-ing is that it fails to advance his own
It is certainly an odd account of liberal morality that would col-
lapse all moral justi¬cations into paternalism. Paternalism is, at best, an
Gerald F. Gaus

uncomfortable ¬t with liberalism.52 Only those under the spell of a the-
ory of practical reasoning according to which all reasons are instrumen-
tal would even attempt to construe liberal morality in this way. Typically,
when one seeks to justify interfering with the liberty of another, it is not
being claimed that the action is bad for him, but that even if it is the best
thing he can do to advance his own goals, this use of his liberty is wrong,
typically because it unjusti¬ably harms others.
3. Suppose, then, that Betty justi¬es her interference (β) on the
ground that Alan™s act (±) frustrates her goals, and that is her reason
for stepping in. But suppose that, under the sway of the purely instru-
mental theory of practical reasons, Betty accepts that in a case in which
± advances Alan™s goals but thwarts hers, he has reason to ±, and she has
reason to β “ that is, interfere with his ±. Notice that in this case, Betty
only justi¬es an interference with Alan: she does not justify the claim that
Alan ought to refrain from ±, or that it is wrong to ±. So she asserts simply
a Hohfeldian liberty to interfere.53 According to Wesley Hohfeld, Betty
has a liberty to β if and only if Alan has no claim against Betty that she
not β. It also follows that if Alan has no claim that Betty refrain from β,
then she has no duty to Alan to not β. For Hohfeld, when we talk about
a person having a right to do something, we sometimes mean that she is
merely at liberty to do so; she has no duty to refrain. But merely to have a
liberty to do something does not imply that you have a claim that others
not interfere. The classic example is the liberty of two pedestrians to pick
up a dollar bill laying on the sidewalk. Neither has a duty to refrain from
picking it up, but neither has a claim on the other to stand aside and let
her pick it up. Such “naked liberties” often characterize competitions;
people have the liberty to win, but no one has a claim to win. So Betty
could simply be asserting a moral liberty to β (that is, interfere with Alan™s
±), while also accepting that Alan has a moral liberty to ±.
Could it be the case that all justifying reasons are such permissions?
If so, liberal morality would contain simply one moral duty, the duty not
to interfere with the actions of another without justi¬cation. In speci¬c
cases, this moral duty would be met by justifying reasons that give one
permission to intervene, though the person interfered with would have
permission to resist. This certainly wouldn™t be much of a morality. Every
justi¬cation of an interference would be a justi¬cation of a competition
or struggle. But liberals endorse rights to property, and rights to bodily in-
tegrity “ and these are not plausibly understood as mere permissions, but
as claims on others to act or refrain from acting. In contrast to liberties,
The Place of Autonomy within Liberalism 287

claim rights imply duties on the part of others not to interfere (or to act);
we might call them rights in the strict sense. To have a right in the strict
sense is to be able to demand that others respect it: they have a duty to re-
spect it, and so are not at liberty to ignore it. One™s rights, then, concern
what is owed to you, and so people are not free to decline honoring your
4. By far the most compelling view, then, is that liberalism recognizes
reasons that (a) justify interference with Alan™s liberty, (b) do not sim-
ply appeal to his own instrumental reasons, and (c) justify claims on
Alan that he has a duty to honor, not mere permissions to interfere.
Again, insofar as reasons are practical (Section IIIe), it follows that fully
rational agents will act on these reasons. Thus, when invoking the fun-
damental liberal principle against Betty, Alan not only supposes that as
a rational moral agent she possesses moral autonomy, and can act on
reasons to set aside her values, but he must also conceive of himself as
a morally autonomous agent: one that accepts, and acts upon, reasons
for action that Betty may give him that justify her interference, and this
not only in the weak sense that Betty may show that she is at liberty to
interfere, but in the stronger sense that she has claims upon Alan that
require him to refrain from blocking her interference, or refrain from
the use of his liberty to which she objects. The basic liberal principle,
then, supposes a relationship among morally autonomous agents. Both
are capable of setting aside their instrumental reasons and acting on

IV Public Reason, The General Will, and Autonomous Legislation
IVa Post-Enlightenment Public Reasoning
Kant conceives of moral autonomy as both a property of the will that is
presupposed by morality and as a substantive moral principle.54 Thus far,
I have been concerned with autonomy in the former sense; I shall now
argue that the fundamental liberal principle also leads to a substantive
morality of autonomy.
The fundamental liberal principle requires that interference be justi-
¬ed. We have arrived at the following conclusion: such justi¬cations are
possible (Section IIIg, 1); they present reasons that do not simply appeal
to the goals of the person being interfered with (Section IIIg, 2); they are
not typically merely permissions (Section IIIg, 3), and so typically consti-
tute claims on the person being interfered with (Section IIIg, 4). It is on
Gerald F. Gaus

this last category of reasons for interference that I shall now focus. I have
also argued that for R to be a reason justifying Betty™s claim on Alan, it
must be the case that a rational Alan would act on R (Section IIIe). So
interpreted, the fundamental liberal principle implies that the justi¬ca-
tion for interfering with Alan must be recognized by a rational Alan as
a reason. When advancing a moral claim on Alan “ that he has a duty
not to use his liberty in some way, and so Betty™s interference with him is
justi¬ed “ she is appealing to Alan as a morally autonomous agent, one
who can act on moral reasons even if that requires putting aside his in-
strumental reasons. In the second stage of the argument, in which Betty is
seeking to meet the onus of justi¬cation by showing that her interference
is justi¬ed, she now occupies the role of claimant trying to show that Alan
has a duty to her (recall that on Hohfeld™s analysis, Betty™s claim right on
Alan implies a duty of Alan toward Betty). Thus, all our conclusions about
how Alan must conceive of Betty as an autonomous agent now apply to
how Betty must see Alan: both assume that the other as well as himself or
herself possess moral autonomy.
Betty™s assertion that Alan does wrong by ignoring her claim on him
presupposes that he has a reason to act on this claim. Betty™s justi¬cation,
then, must be a justi¬cation addressed to Alan as a rational moral agent.
She is barred from presenting a consideration C as a justi¬cation of her
interference if C would not be acknowledged by a rational Alan as pro-
viding him with a reason to act. It is not suf¬cient that C is a reason for
Betty “ that would not in itself show that Alan has a reason, and only if
Betty can claim that Alan has a reason to act can she intelligibly claim
that he does wrong by ignoring her claim (Section IIIc).
To be sure, if, as some assert, R is a reason for Betty if and only if it is a
reason for all rational agents, this is a distinction without importance.55
Any consideration that is a reason for Betty necessarily would be a reason
for any rational agent, and so for Alan. If Betty knows her own reasons,
then she would know his too. Justifying to herself would be equivalent
to justifying to him. (Indeed, there would be no “justi¬cation to,” only
“justi¬cation that”). Some recent philosophers have sought to uphold
this position, or one that approximates it, by appealing to Wittgenstein.
Adopting his argument against private language, they seek to show that
reasons must be inherently public, and shared.56 I cannot examine this
Wittgensteinian-inspired argument here57 (or the allied arguments of
some pragmatists). Suf¬ce it to say for our present purposes that even
if in some sense all reasons are social, and so there are no entirely pri-
vate reasons, this would not show that all reasons are shared among all
The Place of Autonomy within Liberalism 289

members of society. Languages are public and shared, but within a soci-
ety, people speak different languages and numerous dialects. So even if
reasons must be shared with some others, it would still be entirely possi-
ble, and indeed, likely, that some fully rational people will not share my
reasons. If so, then the distinction between justifying to Alan and justifying
to Betty becomes real and important.
This distinction is brought to the fore by what might be appropriately
deemed “the post-Enlightenment insight.” On one plausible view, the
European Enlightenment of the seventeenth and eighteenth centuries
was based on the supposition that the use of human reason produces,
over the long term, convergence on truth in morals as well science.58
The free inquiry of scientists was thought to produce agreement because
(1) the truth is the same for everyone, (2) reason is a shared capacity of
all human beings, and (3) the norms of good reasoning are universal.
Thus, people reasoning correctly about the world will arrive at the same
answer. Any premise p that is true for one person is necessarily true for
all others; if the inferential rule (p & [p ’q])’q is valid for one person,
it is necessarily valid for all. The true and valid results of one person™s
reasoning are thus necessarily true and valid for all. Now, as Rawls puts it,
the post-Enlightenment insight is that it is a “permanent feature of the
public culture of a democracy” that the free exercise of human reason
leads us to embrace a diversity of reasonable moral and religious views.59
The fundamental feature of the political culture of such societies is that
not even rational citizens would share all the same reasons.
To be sure, Enlightenment ¬gures such as Kant recognized the ubiq-
uity of disagreement. Despite his belief that the free exercise of human
reason could reveal universal moral principles, Kant also believed that
on many questions concerning the good and justice, actual people come
to divergent conclusions. For Kant, relying on one™s individual judgment
about justice characterizes “the state of nature” “ “even if men were to
be ever so good natured and righteous before a public lawful state of
society is established, individual men, nations and states can never be
certain they are secure against violence from one another because each
will have the right to do what seems just and good to him, entirely indepen-
dently of the opinion of others.”60 For Kant, reason tells us that if we
are to avoid such con¬‚ict, we must submit to a lawful public authority to
adjudicate disputes about justice. It is plausible, though, to understand
Kant as insisting merely that politics must come to grips with our failure
to be rational; our errors in understanding what reasons we actually have
lead to con¬‚icts of private judgment that government must adjudicate.
Gerald F. Gaus

At least on one interpretation of Kant, it would seem, (fully) ratio-
nal agents would recognize the same reasons for action. Recall, however,
the distinction advanced in Section IIIe between a mere de¬nition of
a rational agent as one who is moved by the best reasons, and the con-
ceptual claim that, given our understanding of rational agents, they will
be moved by their recognition of the best reasons. On this latter view,
the idea of a rational agent is not simply derived from a notion of best
reasons, but relies on a substantive model of good deliberation, evidence
gathering, and so on. If the former, de¬ntional, idea is employed, then
certainly if (1) R is a reason for anyone it is a reason for everyone, and
(2) fully rational agents (by de¬nition) recognize and act upon the best
reasons, then (3) if Alan and Betty disagree whether R is a reason, at
least one of them is not fully rational (though they both might still qual-
ify as “reasonable”). But the idea of being fully rational really does no
work here; it follows entirely from the notion of best reasons. We do,
though, possess a notion of the rational that is not simply derivative of
our understanding of what is the best reason. A rational person takes
into account all the relevant available evidence, makes no errors when
evaluating it, makes all the correct inferences, and so is not subject to var-
ious distortions of deliberation or action (for example, he is not under
the in¬‚uence of drugs or compulsions), and so on. It is still a demand-
ing ideal, much more demanding than being simply a reasonable person
(although it does not require omniscience; rational people do not know
all there is to be known). Nevertheless, we can apply it even when we
do not know what is the best reason. If a person displays the virtues of
rational deliberation and action and none of the vices, then, given our
understanding of a rational agent, we should conclude that he quali¬es as
On this understanding of rational agency, even if we accept premise
(1) in the previous paragraph, it does not follow that if Alan and Betty
disagree on whether R is a reason, at least one of them must have ex-
perienced a failure of rationality “ that is, not be fully rational. Even if
there is a truth to the matter, fully rational people can arrive at differing
conclusions. If so, then even if there is a truth to the matter, Betty cannot
advance her conclusion C as a justi¬cation to Alan on the grounds that
if it is a justifying reason for her, it must, ipso facto, be a reason any fully
rational Alan would be moved by. Given this, Betty must present justi¬ca-
tions that are addressed to others “ that is, she must seek to show that her


. 9
( 13)