Full text of DOI:10.1017/S0007123400000132

B.J.Pol.S. 30, 291–311
Copyright 2000 Cambridge University Press
Printed in the United Kingdom
Liberalism , Autonom y and Stability
ALAN APPERLEY*
Supporters of liberal neutrality distinguish between ‘weak’ conceptions of autonomy which
operate as background features of liberal democratic regimes, and ‘strong’ conceptions, which
amount to conceptions of the good. These latter are to be excluded from the political realm on
the grounds that in order to protect and promote a conception of the good, in the context of a
pluralistic society, the state would have to resort to illiberal methods. The result of this will be
the destablization and fragmentation of the regime. In this article I argue two things: first, that
autonomy ought to be understood, not as a neutral background assumption of liberal theory, but
as a partially comprehensive conception of the good in its own right; secondly, that protecting
and promoting autonomy need not lead either to illiberalism or to the destabilization and
fragmentation of liberal democratic society.
Many liberals acknowledge the importance of autonomy in their political and
moral theories. Liberals, however, differ markedly in their understanding of this
concept and about the role it should occupy in their theories. Rawls, for example,
assumes that we have a higher order interest in autonomy, understood in part
as the capacity ‘to form, to revise, and rationally to pursue a conception of one’s
rational advantage or good’.1 Yet he is critical of the ‘ethical’ autonomy of Mill
and Kant, as well as that of contemporary liberal theorists such as Joseph Raz.2
According to Rawls, the liberal state ought to be ‘neutral’ amongst rival
conceptions of the good life. Rawls criticizes Mill, Kant and Raz because they
treat autonomy as a substantive conception of the good, whereas he takes
autonomy to be a non-controversial background cultural feature of liberal
democratic societies. Given that liberal democratic societies are marked by a
plurality of conceptions of the good, Rawls believes that a state that attempts
to promote a conception of the good will have to resort to coercion in order to
suppress rival conceptions, and this he believes will threaten the stability and
unity of a well-ordered liberal democratic, pluralistic society. Similarly, in a
recent defence of liberalism, David Johnston has argued that, while the liberal
state may pursue some kinds of autonomy as compatible with reasonable
value-pluralism, some conceptions of autonomy are so strong that they
undermine reasonable value-pluralism and are thus illiberal.3
My task in this article is two-fold. First, I want to show with respect both to
* Division of Politics, University of Wolverhampton. I would like to thank Albert Weale, the
anonymous referees and my colleagues Kevin Magill and Meena Dhanda for their comments on
earlier drafts of this article.
1
John Rawls, Political Liberalism (New York: Columbia University Press, 1993), pp. 19, 72.
Joseph Raz, The Morality of Freedom (Oxford: Clarendon Press, 1986).
David Johnston, The Idea of a Liberal Theory: A Critique and Reconstruction (Princeton, NJ:
2
3
Princeton University Press, 1994), p. 78.

292
APPERLEY
Johnston’s humanistic liberalism and Rawls’s political liberalism, that the
account of autonomy they employ is in fact stronger than they take it to be.
Indeed, I want to suggest that it is plausible to see it as a substantive conception
of the good in its own right. Strictly speaking, and in Rawls’s terminology, I
want to suggest that autonomy ought to be understood as a partially
comprehensive conception of the good. Secondly, I want to suggest that
promoting autonomy as a conception of the good need not issue either in
despotic paternalism, as Johnston fears, or in the destabilization and fragmen-
tation of society, as Rawls fears. My contention will be that the kind of good
autonomy is entails a presumption against the use of coercive or other illiberal
methods of social control (though of course it can never rule these out
absolutely). If I am right in this, I will have gone a considerable way towards
meeting Johnston’s and Rawls’s concerns.
I: AUTONOMY
Liberalism, Agency and Autonomy
In The Idea of a Liberal Theory David Johnston defends what he calls a
‘humanist liberalism’. Autonomy plays a central role in this defence, so it is
important to see the way in which Johnston develops his account of this concept.
He begins by distinguishing three forms of autonomy relevant to his enquiry –
autonomy as agency; moral autonomy; and personal autonomy, this latter
understood as a capacity for critical self-appraisal. A liberal state may defend
and promote the first two of these, but a state that defends all three forms of
autonomy defends autonomy in a strong sense. For Johnston, the strong sense
of autonomy is so strong that it is, he believes, incompatible with reasonable
value pluralism.4 Before we consider this claim, let us see what is entailed by
the first two forms of autonomy.
Johnston argues that liberalism is characterized by three core values – that
individual human beings are important; that individuals are to count equally in
terms of ‘whatever features make us worth counting’; and that individuals are
agents. Agents are ‘creatures who are capable of conceiving and of trying to
bring to fruition projects and values’.5 To be an agent is to be autonomous in
this minimal sense. These three core values, however, are not of equal standing
for it is clear that agency-autonomy is more fundamental than the others.
Individuals matter – they are worth counting – because they are agents; they are
as agents equally worthy even if their individual projects and values differ from
one to the next. Moreover, they are worth counting just because they are agents
and not because of the character of the particular projects that each pursues. The
foundational role of agency is reflected in the fact that Johnston makes
4
Johnston, The Idea of a Liberal Theory, p. 138. I return to the issue of what is ‘reasonable’ in
the next section.
5
Johnston, The Idea of a Liberal Theory, p. 68.

Liberalism, Autonomy and Stability
293
agency-autonomy a presupposition of both moral autonomy and personal
autonomy.
Moral autonomy presupposes agency-autonomy but it is not the same thing,
for moral autonomy is achieved when one has an ‘effective sense of justice’,
which involves recognizing ‘that other human beings are agents like yourself,
with projects and values of their own, projects and values that may impose limits
on the things that you can do in pursuit of your own projects and values’.6
Whereas agency autonomy is an essential element in the idea of a human being,
moral autonomy is not. We can without contradiction conceive of a person in
whom a sense of justice is absent. When it is present, it may be too weak to
prevent our other interests from overwhelming it, or it may be too narrow,
ranging only over our families, tribes, or some other group of which we are a
member. A liberal society, however, is marked by an inclusive sense of justice.7
These two types of autonomy – agency-autonomy and moral autonomy –
correspond, as Johnston recognises, to Rawls’s two features of moral
personality.8
By arguing that liberalism is marked by an inclusive sense of justice, Johnston
here seeks to establish a link between the assumption of agency, drawn so thinly
as to exclude almost no-one, and liberal justice. What underlies this link is the
plausible assumption that liberal democratic regimes are marked by heterogene-
ity. As Johnston puts it, ‘liberal societies embody, and liberal theories
presuppose … the assumption of reasonable value pluralism’.9 Similarly, for
Rawls liberal democratic regimes are marked by ‘the fact of reasonable value
pluralism’.10 Beginning from the assumption (or fact) of heterogeneity, liberals
must pitch their arguments in such a way that they can appeal to the diverse
groups – interest, ethnic, cultural, religious and so on – that comprise modern
liberal democratic regimes. Thus the account of agency (or moral personality)
is presented as a relatively uncontroversial starting-point from which to begin
theorizing about politics in liberal democratic regimes. As Rawls acknowledges
in constructing the original position in his A Theory of Justice, the stronger the
assumptions are upon which the theory is built, the less likely it is that they will
be widely shared and the more controversial will the principles of justice upon
which they depend become.11 One might say that the thicker the assumptions,
the more partisan will be the resulting arguments.
Paradoxically, however, this strategy opens liberalism up to the communitar-
ian critique, for the more general become the premisses upon which liberal
arguments turn, it is argued, the less purchase they have on the substantive lives
of the plethora of groups to which they are addressed. Liberalism, it is claimed,
6
Johnston, The Idea of a Liberal Theory, pp. 72–3.
Johnston, The Idea of a Liberal Theory, p. 86.
John Rawls, A Theory of Justice (Oxford: Oxford University Press, 1972), pp. 505ff; Political
7
8
Liberalism, pp. 19, 81.
9
Johnston, The Idea of a Liberal Theory, p. 26. Emphasis in original.
Rawls, Political Liberalism, p. xix.
Rawls, A Theory of Justice, p. 18; Political Liberalism, pp. 127, 157.
10
11

294
APPERLEY
sacrifices the language of membership and solidarity in the interests of
generality, and thus fails to address individuals in terms that they themselves
recognize. Liberals have responded in a variety of ways to these criticisms, and
I do not intend to review these responses here.12 I am concerned, however, with
the question of how one gets from general claims, such as those concerning
agency and moral personality made by Johnston and Rawls, through to claims
about liberal institutions and liberal justice. After all, even members of
non-egalitarian, non-liberal societies are agents capable of formulating and
pursuing projects. Liberalism presumes that all individuals are equally entitled
to respect, regardless of caste, class, race, religion or gender and hence is
properly understood as universalist. Again, individuals are all worth counting,
and worth counting equally, because they are agents. But one may be no less
of an agent in a non-liberal society. How does one get from universalistic
premisses such as these to particularist claims about liberal democratic regimes?
Before we tackle this issue, we should recognize that the universality of these
assumptions is potentially in conflict with liberalism’s aspiration to inclusive-
ness, for although its characterizations are asserted of all human beings, not all
the human beings over whom the characterizations range accept or value the
liberal account of what human beings are like. This is true even of the attribution
of agency. It is possible, for example, to reject altogether the idea that human
beings are agents in any meaningful sense, and to see them instead as merely
vehicles in the service of ‘selfish’ genes, or as instruments of a divine power.13
Such a rejection might well be unusual (more so now than a century ago) but
it is not unthinkable, and if the conception of agency upon which justifications
ofliberalinstitutionsandliberalprinciplesaretobebasedisrejected, thenliberal
arguments may find that they have no soil in which to germinate. It is possible,
of course, that one might arrive at liberal institutions and principles via some
other route, such as an evolutionary biology or divine edict. Rawls allows that
the necessary conditions of liberalism, including the liberal sense of justice, will
have emerged as a ‘normal result of the exercise of human reason within the
framework of the free institutions of a constitutional democratic regime’.14 For
Johnston, the link is made by connecting agency-autonomy (present in any
society) to a particular conception of moral autonomy, a conception specified,
in part at least, in terms of inclusiveness.
But even if we accept that human beings are agents in Johnston’s sense, we
might deny that this is what makes them worth counting. The issue here is the
12
For overviews of the debate between liberals and communitarians, see Stephen Mulhall and
Adam Swift, Liberals and Communitarians (Oxford: Basil Blackwell, 1992); Shlomo Avineri and
Avner de-Shalit, eds, Communitarianism and Individualism (Oxford: Oxford University Press,
1992); and Ellen Frankel Paul, Fred D. Miller Jr and Jeffrey Paul, eds, The Communitarian Challenge
to Liberalism (Cambridge: Cambridge University Press, 1996).
13
Ignatius Loyola, founder of the Jesuit order, considered that in pursuing God’s work he must
be as ‘a corpse which has neither intelligence nor will’ (cited in Friedrich A. Hayek, The Constitution
of Liberty (London: Routledge, 1960), p. 423, n. 12).
14
Rawls, Political Liberalism, p.xviii, see also pp. 4, 36–7. I return to this point below in the
section after next.

Liberalism, Autonomy and Stability
295
relative importance of the assumption in establishing further claims. Since
Johnston argues for a humanistic liberalism it is not surprising to find that he
wishes to mark human beings off from the rest of the animal world (and, we
might add, from nature as such) as worthy of special concern.15 Yet although
a liberal such as Bentham would not have denied that human beings are agents
in Johnston’s sense, he evidently thought that what makes human and
non-human beings worth counting is that they feel pain. One might therefore
agree with Johnston that human beings have agency-autonomy but disagree that
this is what makes individuals worth counting.
Yet even if we accept with Johnston that it is agency that makes us worth
counting, we would still need to say what it is about agency that leads us to
endorse a liberal society, as opposed to any other kind of society. As Johnston
remarks, agency-autonomy is a capacity which it is not difficult to develop since
the conditions for its development are not stringent.16 He also remarks that it
is ‘inconceivable that any society would not seek to secure the conditions
required for all its members to become agents in this rudimentary sense’.17 Yet
so rudimentary is this sense of agency that it is not clear why, on these grounds,
weshouldendorsealiberalsocietyoveranyother. Afterall, non-liberalsocieties
are what Kymlicka has called ‘contexts of choice’ just as liberal societies are.18
They are contexts within which certain choices have meaning, and others do not.
Choices that are meaningful in the context of non-liberal societies might not
have the same meaning within a liberal society and vice versa, but this fact in
itself provides us with scant reason to prefer liberal societies over non-liberal
societies.
We might say, for example, that in a liberal society there is greater scope for
conceiving projects or that the particular kinds of choice open to us are in some
way preferable to those in a non-liberal society, and this might even be true. But
we cannot say why this is preferable solely in terms of the weak account of
agency provided by Johnston. We need an account of what human flourishing
might consist in, or of what flourishing in a particular social context might mean
in order to make the case for having greater scope rather than less or for having
this range of choices rather than that. Such accounts of human flourishing are
conceptions of the good and I contend that in liberal societies the liberal
conception of the good can be partly explicated through an account of
autonomy.19 I will return to this point below, but for now all we need to recognise
is that this account of autonomy will have to be more substantive than is
Johnston’s weak account of agency.
15
Johnston allows that individuals may formulate and pursue ‘projects and values that are not
designed simply to affect our own experiences’ (The Idea of a Liberal Theory, p. 68, also pp. 81,
89) and this may include concern for animal welfare or the environment.
16
Johnston, The Idea of a Liberal Theory, p. 72.
Johnston, The Idea of a Liberal Theory, p. 79.
Will Kymlicka, Liberalism, Community and Culture (Oxford: Clarendon Press, 1989),
17
18
p. 166–7.
19
For Charles Taylor autonomy is a ‘hypergood’ indicating an ‘inherently conflictual’ ethical
outlook (Charles Taylor, Sources of the Self (Cambridge: Cambridge University Press, 1989), p. 65).

296
APPERLEY
Autonomy and Rawlsian Political Liberalism
Having shown that Johnston’s weak account of autonomy is more controversial
than he takes it to be, I now turn to a discussion of Rawls with the same objective
in mind. Like Johnston, Rawls treats the autonomous life as an uncontroversial
background assumption of his theory of justice.20 It is a ‘general fact’, says
Rawls, that the political culture of a liberal democratic society ‘normally
contains, at least implicitly, certain fundamental intuitive ideas from which it
is possible to work up a political conception of justice suitable for a
constitutional regime’.21 The two moral powers that characterize Rawls’s
account of moral personality – the capacity to form, revise and pursue a
conception of the good (Johnston’s agency autonomy) and the capacity for a
sense of justice (Johnston’s moral autonomy) – are part of the ‘shared fund of
implicitly recognized basic ideas and principles’ of the public culture of a liberal
democratic society.22 This public culture is itself born out of a ‘certain political
tradition’ in which the idea of society as a ‘fair system of cooperation over time,
from one generation to the next’ is fundamental, as is the idea that citizens are
‘free and equal’.23
Despite this shared fund of ideas and principles, a liberal democratic regime
will nevertheless be characterized by reasonable disagreement over conceptions
of the good. Rawls calls this the ‘fact of reasonable pluralism’ and distinguishes
this from the fact of pluralism as such. The point of this distinction is in part
to limit the scope of neutrality, for the liberal state need only be neutral amongst
reasonable conceptions of the good.24 It follows that if liberal democratic
societies are marked by the fact of reasonable pluralism, then ‘reasonableness’
must itself be a fact implicit in the public political culture of such regimes. This
must be so for, as we shall see, Rawls relates it directly to that aspect of moral
personality which sustains the sense of liberal justice (which, as we have seen,
Johnston calls moral autonomy); and the account of moral personality is
amongst the assumptions from which the theory of justice as fairness is worked
up. Given that Rawls bounds his neutrality with an appeal to the idea of
reasonableness, we need to say something about what it is and what work it
performs in his theory.
Rawls distinguishes the reasonable from the rational and he does this in
part on the grounds of the two moral powers identified above. One is
20
Rawls, Political Liberalism, p. 178.
Rawls, Political Liberalism, p. 38, n. 41.
Rawls, Political Liberalism, pp. 8, 13–14, 15, 18–19, 43, 78, 79, 90, 97. Several commentators
21
22
have pointed out that the account of moral personality given here is broadly derived from Kant and
as such is neither uncontroversial nor political, where ‘political’ is contrasted with ‘metaphysical’,
as it is in Rawls’s ‘Justice as Fairness: Political Not Metaphysical’, Philosophy and Public Affairs,
14 (1985), 223–51; Political Liberalism, pp. 3–46. For discussions of this see William A. Galston,
Liberal Purposes: Goods, Virtues and Diversity in the Liberal State (Cambridge: Cambridge
University Press, 1991), pp. 118–39; Patrick Neal, ‘Justice as Fairness: Political or Metaphysical?’
Political Theory, 18 (1990), 24–50.
23
Rawls, Political Liberalism, p. 13–14.
Rawls, Political Liberalism, pp. 190–5; Johnston, The Idea of a Liberal Theory, p. 26.
24

Liberalism, Autonomy and Stability
297
rational in so far as one has a conception of the good which it is rational to pursue
(agency autonomy). One is reasonable (in part) in so far as one is willing ‘to
propose fair terms of cooperation and to abide by them providing others do so’
(moral autonomy).25 The rational is directed towards one’s own conception of
the good, while the reasonable is directed towards the public culture.26 More
specifically, it is directed at the public political culture, since Rawls is insistent
that this is all political liberalism requires.
But the reasonable is only partly characterized by the willingness to propose
and abide by fair terms of co-operation. The ‘second basic aspect’ of the
reasonable is a willingness to accept the ‘burdens of judgement’. Rawls asks:
how is it that reasonable people who ‘share a common human reason, similar
powers of thought and judgement [and who] can draw inferences, weigh
evidence, and balance competing considerations’ can yet arrive at widely
divergent conceptions of the good?27 One answer to this apparent conundrum
is that there are many gaps in the evidence that is available to us and there are
limitations bearing on our reasoning, arising both from within our rational
conceptions of the good, and from the reasonable claims of others. Reasoning
is a complex process and it is not surprising that it should result in a diversity
of reasonable conceptions of the good.
A reasonable doctrine of the good then, is one that accepts the burdens of
judgement and consequently accepts the legitimacy of a plurality of rival but
similarly reasonable conceptions of the good. Accepting the burdens of
judgement, one will conclude that it is unreasonable to employ the coercive
power of the state to impose one’s conception of the good on everyone else.28
The political conception of justice that arises from these considerations is
therefore one that does not take a stand on the relative merits of reasonable
conceptions of the good.29 Conceptions of the good are reasonable if they are
political in this sense and view the public, political realm as an area where
controversy over freedom, equality, the fair terms of co-operation and, of
course, thecentralityofautonomy, hasbeensettled. Therefusaltodosoamounts
to unreasonableness because it betrays a willingness to exploit the ‘deep
divisions latent in society’ for partisan ends and this ‘dangerously increases the
insecurity and hostility of public life’.30
To get to this point Rawls has had to make a number of contentious empirical
claims about the public political culture of liberal democratic regimes, and these
claims have not gone unchallenged. George Klosko, for example, has pointed
out that Rawls provides almost no evidence to support his empirical claims and
25
Rawls, Political Liberalism, pp. 54, 49.
Rawls, Political Liberalism, p. 53.
Rawls, Political Liberalism, p. 55.
Rawls, Political Liberalism, pp. 54, 60, 61, 138.
Rawls, Political Liberalism, p.xxi, 15, 57n, 67, 95, 127, 157, 161.
Rawls, Political Liberalism, p. 161. Rawls elsewhere (p. 152) suggests that, should such
26
27
28
29
30
circumstances arise, we are not required to maintain neutrality.

298
APPERLEY
that where empirical evidence exists it does not obviously support Rawls’s
case.31 Rawls points out, for example, that amongst our ‘settled convictions’ is
a belief in the virtue of toleration32 and his theory of justice holds that the basic
liberties (rights) are lexically prior to the distributive claims of the difference
principle and will not be traded for other goods.33 But Klosko points out that
many Americans display an intolerance towards, and a willingness to
circumscribe the rights and liberties of, various groups casting doubt on the
accuracy of Rawls’s characterization of the context to which his argument is so
carefully fitted.34 It is true, of course, that Rawls does not claim that the shared
assumptions he identifies are explicitly held by those to whom his argument is
addressed. All he needs to claim is that these assumptions are latent in the public
political culture, awaiting only the right kind of argument or method to bring
them to the fore.
But even if Rawls is right concerning the (latent or implicit) existence of these
shared assumptions, it does not follow that they provide a non-controversial
background against which a political conception of justice might be developed.
Rawls, for example, accepts that empirical evidence ‘is conflicting and complex,
and thus hard to assess and evaluate’.35 It may well be the case that judgements
concerning the nature and importance of shared assumptions within the public
political culture of a democratic society will themselves be controversial. We
might well agree on their existence, but nevertheless differ widely in the
meaning or relative importance we accord to this fact. Michael Walzer, for
example, claims in his Spheres of Justice to be offering an interpretation of the
shared understandings of American political culture, the same political culture
to which Rawls’s arguments are addressed, yet his account of that culture is not
only at odds with Rawls’s at many points, but is widely held to be controversial
in its own right.36 Moreover, as J. Donald Moon has pointed out, the existence
of shared understandings is no guarantee of the stability that Rawls prizes so
highly.37 Infact, sharedunderstandingsmightprovidethenecessarybackground
against which a controversy can be understood as a controversy.
Rawls wishes to purge the political conception of justice of appeals to truth
31
George Klosko, ‘Rawls’s “Political” Philosophy and American Democracy’, American
Political Science Review, 87 (1993), 348–59, p. 348.
32
Rawls, Political Liberalism, p. 122; A Theory of Justice, p. 19.
Rawls, A Theory of Justice, pp. 152, 244, 250, 542.
Klosko, ‘Rawls’s “Political” Philosophy’, pp. 352–3.
Rawls, Political Liberalism, pp. 56–7.
33
34
35
36
Walzer claims that, based on his interpretation of the meaning of social goods in American
society, the appropriate institutional arrangement would be that of ‘a decentralized democratic
socialism’ including ‘a strong welfare state … a constrained market … [and] workers’ control of
companies and factories’ (Michael Walzer, Spheres of Justice: A Defence of Pluralism and Equality
(Oxford: Basil Blackwell, 1983), p. 318). Walzer’s claims are clearly at odds with the popular view
of contemporary American society as individualistic, market-driven and broadly libertarian.
37
J. Donald Moon, Constructing Community: Moral Pluralism and Tragic Conflicts (Princeton,
NJ: Princeton University Press, 1993), p. 25.

Liberalism, Autonomy and Stability
299
in favour of appeals to reasonableness because he believes that appeals to truth
will be controversial.38 Yet, as Jean Hampton has pointed out, Rawls appears
to treat such values as freedom and equality, reasonableness and the autonomous
life as true and correct for a liberal democratic regime.39 Rawls might respond
to this criticism by invoking a distinction between facts and values. If the shared
intuitions he appeals to are basic non-controversial building-blocks untainted
by value, then the argument for political liberalism might well go through. Yet
we have seen that claims about autonomy, to take just one example, are not
uncontroversial. And Rawls does more than merely point to the existence of
these shared assumptions; he tacitly endorses them too. After all, it is surely
possible to agree that they exist but think that this is a bad thing and that liberal
democracy is undesirable. Rawls’s claims regarding the burdens of judgement,
it is true, explicate his account of reasonable pluralism. But they also show why
thepoliticalrealmcannotbepurgedofcontroversyandreasonabledisagreement
around these ‘basic’ values of liberal democracy.
Having shown that for both Johnston and Rawls, the notion of autonomy is
a controversial value I want now to suggest that there is good reason to view
the autonomous life as a conception of the good in its own right. Specifically,
I want to show that it ought to be understood as a partially comprehensive
conception of the good. This is Rawls’s own term, and so I begin with an
explication of what he means by this.
Autonomy as a Conception of the Good
Rawls treats the two aspects of moral personality as equivalent – both elements
are required in order to support the theory of justice as fairness. Yet as we have
seen, Johnston indicates that they are not equivalent and that while it is difficult
to imagine a society in which agency-autonomy was absent, it is not difficult
to imagine a society in which the liberal sense of justice (or, more generally,
moral autonomy) is either underdeveloped, non-inclusive or absent altogether.
The question that Johnston’s discussion raises is: ought liberals to be concerned
about the development of the required sense of liberal justice? Rawls might of
course be right to assume, as he does, that there are sufficient moral resources
38
Rawls thinks that excluding controversy from the political realm is necessary to secure stability.
But this amounts to politics without politics, for it may be precisely those issues that are controversial
and divisive that are most likely to be the ones that people want to place on the political agenda. See
Chandran Kukathas and Philip Pettit’s discussion of this in their Rawls: A Theory of Justice and Its
Critics (London: Polity Press, 1990), pp. 148–9. See also Peter Singer, Democracy and Disobedience
(Oxford: Oxford University Press, 1973), pp. 84–92.
39
Jean Hampton, ‘The Moral Commitments of Liberalism’ in David Copp, Jean Hampton and
John Roemer, eds, The Idea of Democracy (Cambridge: Cambridge University Press, 1993),
pp. 292–313. See also Simon Caney, ‘Anti-perfectionism and Rawlsian Liberalism’, Political
Studies, 43 (1995), 248–64, pp. 252–61. Larmore explicitly states that the norms of rational dialogue
and equal respect, and the principle of neutrality they justify, are to be understood as ‘the correct
and valid norms’ of liberal democratic society (Charles Larmore, ‘Political Liberalism’, Political
Theory, 18 (1990), 339–60, pp. 353–6).

300
APPERLEY
in society with which to sustain the liberal sense of justice. We have already
noticed that Rawls believes that the necessary resources will have worked
themselves into the culture of liberal democratic regimes as a ‘normal result of
the exercise of human reason within the framework of the free institutions of
aconstitutionaldemocraticregime’. Moreover, theseresourceswillbeprotected
and sustained by the basic rights and liberties contained in the first principle of
justice whose purpose is ‘to assure that everyone can adequately develop these
powers and exercise them fully over the course of a complete life as they so
decide’.40
However, we might wonder about the extent to which the framework of
institutions has in fact been free. It is surely plausible to argue that the legal and
political institutions of liberal democratic regimes operate systematically in the
interests of certain sections of society – candidates might include whites, males,
capitalists, the middle classes and so on – and against the interests of others. I
cannot consider the arguments here, but such concerns cannot be dismissed out
of hand given the importance to Rawls’s argument of the role of such
background features. Moreover, bearing in mind Rawls’s comments about the
burdens of judgement, we might wonder whether or not there is a correct or at
least a non-controversial way in which to understand the idea of ‘free
institutions’. There are few ideas more controversial in normative political
theory than that of freedom. Yet even if those institutions were free, and free
in a correct way or non-controversial way, we might still wonder about the
relationship between the constitutional, political realm and society more
generally. It is this latter point that I will address in this section. Rawls argues
that the political conception of justice as fairness should not be understood as
a comprehensive conception of the good in its own right. If it were to be so
understood, implementing it would destabilize the liberal democratic regime.
Given the concern I express over the development and maintenance of the liberal
sense of justice (moral autonomy) I will argue that since liberals cannot be
indifferent to the extra-political sources of moral development they must
necessarily understand autonomy in a strong sense, which is to say as a
comprehensive conception of the good in its own right.
Rawls defines comprehensive conceptions of the good in relation to the
political realm, and this distinction is made in terms of scope.41 Comprehensive
conceptions of the good are general when they apply to a wide range of subjects.
The political conception of justice is not general but particular in that it applies
only to the basic structure of a well-ordered society. Comprehensive
conceptions of the good include ‘conceptions of what is valuable in human life,
and ideals of personal character, as well as ideals of friendship and of familial
and associational relationships, and much else that is to inform our conduct’.42
Since this is a matter of scope, conceptions of the good may be more or less
comprehensive. Rawls distinguishes, for example, between wholly comprehen-
40
Rawls, Political Liberalism, p. 202.
Rawls, Political Liberalism, pp. 13, 175.
Rawls, Political Liberalism, p. 13.
41
42

Liberalism, Autonomy and Stability
301
sive conceptions and partially comprehensive conceptions. The former cover
‘all recognized values and virtues within one rather precisely articulated
system’, whilst the latter ‘comprises a number of, but by no means all,
nonpolitical values and virtues and is rather loosely articulated’.43 The political
conception of justice is not comprehensive in either of these two senses for ‘by
definition, for a conception to be even partially comprehensive, it must extend
beyond the political’.44 What all reasonable comprehensive conceptions of the
good will have in common – the point at which they will ‘overlap’ – is the
political conception of justice. It is in the non-political elements that the ‘fact’
of plurality lies.
Elsewhere, Rawls expands on the political/non-political distinction. For
example, the political realm ‘is distinct from the associational, which is
voluntary in ways that the political is not; [and] it is distinct from the personal
and the familial, which are affectional, again in ways the political is not’.45
Clearly, the political conception of justice, in so far as it is restricted to the realm
of the political, need not apply to the personal, the familial or the associational.
Were it to do so the ‘political’ conception of justice would, in effect, become
an at least partially comprehensive conception of the good in that it would
promote an ideal of just ‘personal … familial and associational relationships’.
It seems clear that Rawls, who now takes the ‘problem of stability’ to be
‘fundamental to political philosophy’, fears that this would run the risk of
destabilizingsociety.46 Ifwecanshowgoodreasonfornotexcludingthefamilial
and associational realms from the conception of justice then we will have gone
some way towards establishing it, in Rawls’s own terms, as a partially
comprehensive conception of the good in its own right. The so-called
background features of the conception of justice would then stand revealed as
an essential element in that partially comprehensive conception of the good. It
is to this task that I now turn.
The Familial Realm
Susan Moller Okin has pointed to the contentious role of the family in Rawls’s
theory of justice. She has remarked upon what she sees as a failure on Rawls’s
part to treat the family as itself a political institution to which the principles of
justice ought to apply.47 Rawls, for example, accepts as a principle of moral
psychology that the family is, for children, the first school of moral development
43
Rawls, Political Liberalism, p. 13.
Rawls, Political Liberalism, p. 175.
Rawls, Political Liberalism, p. 137.
Rawls, Political Liberalism, p. xix.
44
45
46
47
Susan Moller Okin, Justice, Gender, and the Family (New York: Basic Books, 1989),
pp. 93–101; ‘Humanistic Liberalism’ in Nancy L Rosenblum, ed., Liberalism and the Moral Life
(Cambridge, Mass.: Harvard University Press, 1989), pp. 39–53. Okin, it should be noted, broadly
endorses Rawls’s account of liberal justice, but believes that Rawls does not press his own argument
to its logical conclusions. See also Elizabeth Frazer and Nicola Lacey, ‘Politics and the Public in
Rawls’ Political Liberalism’, Political Studies, 43 (1995), 233–47, pp. 242–6.

302
APPERLEY
in which the sense of justice is planted and nurtured.48 The idea is that the sense
of justice will be transmitted from one generation to the next in part through the
socializing family unit. However, as Okin points out, this assumes that the
family is already a just institution and this assumption is unsubstantiated, or at
the very least idealistic.49 Her criticism of Rawls focuses primarily on his A
Theory of Justice, but as we have already noticed, Rawls’s reworking of his
theory as an explicitly political conception of justice also emphasizes that the
family is a non-political unit. Of course, Rawls acknowledges that our sense of
justice does not only develop within the family but also through our participation
as ‘fully cooperating members’ of a democratic society. Now if, as is often
claimed, many women are (for whatever reasons) prevented from co-operating
fully in the democratic societies in which they live, then it might follow that their
ability to develop the requisite sense of justice could be impaired. As another
commentator has put it: ‘Women who do not venture beyond the family or
participate in practices beyond mothering cannot attain an adequate understand-
ing of the way in which politics determines their lives’.50 This is an important
point in its own right, but becomes more poignant if it is also accepted that these
women, whether one likes it or not, will largely shoulder the responsibility of
socializing the next generation of citizens into a sense of justice.51
This is not to hold women responsible for the weakness or absence of a strong
sense of justice in the public culture; after all, alienation from the political
process in the United States is widespread and cuts across gender boundaries.
Rather, it suggests that there may be a structural problem bearing on the
development of the sense of justice that the liberal state would do well to
address.52 As we have seen, Rawls appears to be confident that ‘the work of free
practical reason within the framework of free institutions’ has spontaneously
created and entrenched the shared intuitions upon which the overlapping
consensus depends.53 But this rather sanguine portrait of the development of the
conditions for an overlapping consensus does not address the concern that
alienation from the political community may disrupt the process of socialization
which is a necessary element in the development of liberal values and the liberal
sense of justice, and may itself be a symptom of a wider malaise which must
be addressed beyond the realm of the political, as Rawls understands this. Okin,
for her part, believes that Rawls’s principles of justice could be pressed into
service as a means of achieving justice ‘both within the family and in society
atlarge’.54 ThiswouldrequireextendingtheboundariesofwhatRawlsatpresent
understands as ‘the political’.
48
Rawls, A Theory of Justice, p. 490; also Johnston, The Idea of a Liberal Theory, p. 86.
Okin, Justice, Gender, and the Family, p. 94.
Mary G. Dietz, ‘Citizenship with a Feminist Face: The Problem with Maternal Thinking’,
49
50
Political Theory, 13 (1985), 19–37, p. 32.
51
Okin, ‘Humanistic Liberalism’, p. 41.
Frazer and Lacey, ‘Politics and the Public’, p. 245.
Rawls, Political Liberalism, pp. 36–7, 129; cf. Larmore, ‘Political Liberalism’, pp. 347, 354,
52
53
356, 357.
54
Okin, Justice, Gender, and the Family, p. 109.

Liberalism, Autonomy and Stability
303
The Associational Realm
Writing in the shadow of absolutism, the fear of the all-powerful state drove
classical liberals to develop the distinction between the state – the sphere of law
– and civil society, the sphere of contract, exchange and of voluntary
association. Much intellectual energy was spent trying to establish precisely
where the boundary between state and civil society lay, but the greatest concern
was with how best to regulate the state in the interests of individual liberty. It
has, however, recently been suggested that the greatest obstacle to individual
liberty is no longer solely the political realm of the state, for civil society itself
has come to be ‘dominated by large-scale quasi-public and private institutions
possessed of powers that dwarf those of many pre-modern states’.55 These
‘quasi-public and private institutions’ along with the major institutions of the
state effectively comprise a hierarchical, unaccountable and, hence, undemo-
cratic ‘bureaucratic monoculture’ which threatens freedom and equality.56 Paul
Hirst, whose argument this is, does not believe that liberalism can remedy this
situation unless it overcomes its reluctance to accept that the state/civil society
distinction is anachronistic in the face of these developments. Yet were Rawls’s
conception of justice to be applied outside of the political realm (as Rawls
understands it) and to the associational sphere, then many of Hirst’s concerns
might well be addressed.57
Rawls, as we have seen, acknowledges that our sense of justice develops both
within the family and in and through our participation as ‘fully cooperating
members’ of a democratic society.58 The requisite sense of justice for a liberal
democratic regime then depends upon a process of socialization which is not
itself focused solely on the political realm, as Rawls understands this. In
contemporary liberal democratic societies, the vast majority of those citizens
who are lucky enough to work will be luckier still if they find themselves in
places of work that reinforce their sense of justice through either co-operation
(let alone ‘full’ co-operation) or democratic participation. This is not to say that
those who work do not experience satisfaction in the jobs they do, but
job-satisfaction is not the issue here. The issue is whether or not the environment
in which one spends a sizeable proportion of one’s waking life adequately
sustains, or contributes to the maintenance of, a liberal sense of justice.
If Rawls is right that moral autonomy, in the form of a liberal sense of justice,
must be in place as a prerequisite of our adopting the principles of justice his
theory provides, then the weakness or absence of this attribute must be of some
concern. Even if the requisite sense of justice is already established, it may need
to be sustained and reinforced positively, rather than negatively as a by-product
of enacting the principles of justice (for which this attribute is a prerequisite
55
Paul Q. Hirst, ‘Democracy and Civil Society’, in Paul Q. Hirst and Sunil Khilnani, eds,
Reinventing Democracy (Oxford: Blackwell, 1996), pp. 97–116, at p. 101.
56
Hirst, ‘Democracy and Civil Society’, p. 104.
Hirst himself is sceptical about this possibility.
Rawls, Political Liberalism, pp. 3, 9, 19, 20.
57
58

304
APPERLEY
anyway). Rawls acknowledges a tendency for justice to be eroded even
when individuals act fairly towards one another but he does not consider
the implications of this erosion for the development of the shared understand-
ings which implicitly embody the sense of justice in the first place.59 Given that
most of us will probably engage with the familial and the associational spheres
more fully and directly than the ‘political’ sphere (and after all it is liberals who,
by and large, view engagement in the political realm as a choice rather than a
duty), liberals cannot afford to be indifferent to the impact of these
extra-political spheres on the development of a sense of liberal justice. They
cannot be indifferent to the development of the requisite notion of moral
autonomy.
It is true that any blurring of the boundaries between state and civil society
(assuming that these boundaries have ever been clearly drawn) will cause
consternation amongst liberals who rightly fear the totalitarian consequences of
the politicization of the non-political. However, arguing for an extension of the
scope of the political is not the same as arguing for the total politicization of
society and it is important to retain a distinction between the personal and the
political. I shall return to these concerns in the final part of the article.
Whilst accepting Johnston’s and Rawls’s claim that autonomy (agency and
moral) is a central liberal value I have suggested that the claim that autonomy
is a neutral background assumption of liberal democratic regimes is unconvinc-
ing. I have also suggested (following Hirst and Okin) that liberals ought not to
treat it as such but instead ought to understand it as a partially comprehensive
conception of the good, one that reaches beyond the political realm as Rawls
defines this to the associational and familial realms also. In the next section I
want to turn to a consideration of the consequences of understanding autonomy
as a partially comprehensive conception of the good. I begin with an
examination of Johnston’s claim that to do so would open the way to a despotic
paternalism, before finally considering Rawls’s fear that this would destabilize
and fragment society.
STABILITY
Making autonomy one of the presuppositions of liberal democratic societies and
yoking it to a conception of reasonableness would seem to render those groups
who do reject autonomy unreasonable by definition. Since both Johnston and
Rawls bound their neutrality with appeals to reasonableness, those conceptions
of the good that fall outside of the stipulated realm of reasonableness are thus
liable to be met with illiberal responses. But it is not unreasonable to reject the
value of autonomy, and liberal democratic societies will always be characterized
in part by those who so reject it. I have suggested that there is good reason to
treat autonomy as a partially comprehensive conception of the good. Johnston
rejects this move on the grounds that it leads liberalism towards illiberalism;
59
Rawls, Political Liberalism, p. 267.

Liberalism, Autonomy and Stability
305
Rawls rejects this move on the grounds that it will destabilize the well-ordered
liberal democratic regime. In this section, I argue that the fears of these writers
can be allayed in part because of liberalism’s focus on the development and
maintenance of autonomy.
Autonomy and Self-Realization
We have seen that Johnston distinguishes between three kinds of autonomy of
which the first two – agency and moral autonomy – are, or so he believes,
compatible with a liberal society. The third form of autonomy – personal
autonomy – is not. Let us now examine Johnston’s case against personal
autonomy.
Personal autonomy, Johnston claims, is the capacity critically to appraise
one’s projects and values. Such a capacity is often justified in terms of its
contribution to measuring the value of things – one comes to understand what
is worthwhile and what is not through the process of critical appraisal; to social
progress – failure to critically appraise one’s values and projects inclines one
to social and political conservatism; and to making individuals morally better
people and societies as a whole more just.60 Moreover, personal autonomy,
understood as ‘critical self-appraisal or reflective self-evaluation’, may be
construed as a defining aspect of personhood or as a telos for human nature.61
For Johnston the defining characteristic of human nature is ‘the imaginative
capacity to formulate projects and values’ and this, he suggests, does not require
critical self-appraisal since one may formulate projects and values and create
a meaningful and worthwhile life for oneself in the absence of this capacity.62
Johnston does not deny that the capacity critically to appraise one’s own
projects and values is valuable. But, he argues, since this capacity is not easily
distinguishable conceptually and practically from the kind of rational self-
direction that Berlin eloquently warns us against in his essay Two Concepts of
Liberty, it should not feature explicitly as a goal of liberal society.63 Johnston
fears that the explicit promotion by the liberal state of personal autonomy,
understood as rational self-direction, would open the way for a ‘despotic
paternalism’ fuelled by the belief that ‘there is one unique pattern [of life] which
alone fulfils the claims of reason’.64 Again, personal autonomy is incompatible
with reasonable value pluralism.
As an opening shot, one might respond to Johnston by pointing out that the
60
Johnston, The Idea of a Liberal Theory, pp. 88–93.
Johnston, The Idea of a Liberal Theory, pp. 89–90.
Johnston, The Idea of a Liberal Theory, pp. 89, 93.
Johnston, The Idea of a Liberal Theory, p. 95; Isaiah Berlin, ‘Two Concepts of Liberty’ (1958),
61
62
63
reprinted in Berlin, Four Essays on Liberty (Oxford: Oxford University Press, 1969), pp. 118–72.
64
Johnston, The Idea of a Liberal Theory, p. 95. Berlin addresses his fears to positive liberty, but
many writers take this to be an account of personal autonomy, e.g. Richard Lindley, Autonomy
(London: Macmillan, 1986), p. 6; Gerald Dworkin, The Theory and Practice of Autonomy
(Cambridge: Cambridge University Press, 1988), p. 13; Steven Lukes, Individualism (Oxford:
Oxford University Press, 1973), p. 55.

306
APPERLEY
self’s relationship to its self-chosen ends does not, as a matter of fact, rule out
pluralism since nothing is said about the ends that the individual actually
chooses. If critical self-appraisal of one’s projects and values were in fact the
defining characteristic of personhood one need not find oneself restricted in the
projects or values that one ultimately settles upon. What is important is the way
that one comes to hold projects and values and this is a long-standing concern
of liberalism. However, let us consider more closely Johnston’s claim that a state
that promotes strong autonomy tends towards despotic paternalism. Since
Johnston himself invokes Berlin in defence of his position, it will be useful to
remind ourselves of this writer’s arguments.
What underlies Berlin’s suspicion of personal autonomy is in part his
rejection of essentialism, the idea that there is a ‘true self’ to be discovered. It
is, for Berlin, a short step from this assumption to ignoring the actual empirical
self in the cause of pursuing this ‘true self’. If I, having achieved self-realization,
recognize that you have not yet done so I might think myself justified, in my
elevated position, in bullying you, oppressing you, even torturing you in the
name of your ‘true self’.65 Clearly, no liberal society could live with such
illiberalism, yet precisely this tendency, says Berlin, ‘is at the heart of all
political theories of self-realization’.66 The question is: do those writers whom
Johnstoncitesasdefendersofautonomyinthestrongsense–namelyJohnStuart
Mill and Joseph Raz – really defend an essentialist account of the self in Berlin’s
sense?
In the case of Mill, it is true that he defends liberty of thought and expression
in terms of a search for objective truth, at one point even suggesting that the
progressive uncovering of such truths through the free play of ideas would result
in a ‘gradual narrowing of the bounds of diversity’.67 For defenders of pluralism,
this looks ominous and appears to play to Johnston’s fears. For Johnston, value
is not an objective fact somehow inhering in things, but rather resides in some
aspect of the agent and his or her conception of the good life and this is illustrated
by the fact that there is conflict over values. The idea that personal autonomy
might be important in discovering the ‘correct’ value of things is therefore
misguided. There are, however, problems with Johnston’s account.
First, it is clear that Mill is not just concerned with the discovery of objective
truth. He is also concerned with the way these truths come to be held by
individuals. Were such truths to be discovered and simply taught by rote from
generation to generation, then they would be deprived of their vitality and would
no longer actively inform one’s character and conduct. Even if objective truths
existed it would still be important that people develop the capacity to grasp
65
This claim can be resisted. One might hold the view that self-realization can only be achieved
through one’s own efforts. Berlin, of course, seeks to make a historical point rather than a
philosophical one.
66
Berlin, ‘Two Concepts of Liberty’, p. 134.
John Stuart Mill, ‘On Liberty’ (1859), reprinted in Harold B. Acton, ed., Utilitarianism, On
67
Liberty, and Considerations on Representative Government (London: Dent/Everyman, 1972),
pp. 63–170, at p. 103.

Liberalism, Autonomy and Stability
307
these truths for themselves, for only under these circumstances will these truths
retain their ‘vital effect on character and conduct’.68 It is reasonable to conclude
that Mill’s claim concerning the capacity to identify and evaluate truth is
independent of his claim that there may be objective truths to be discovered, and
that the former is of greater importance than the latter since it is this capacity
that defends the individual, and hence society at large, against the unwarranted
imposition of possibly erroneous truths. Despotic paternalism in the name of
objective truth would therefore be ruled out by this argument.
Secondly, Mill himself argues that in respect to society, politics, religion and
morals (those fields which Berlin refers to as ‘ideological’) ‘truth depends on
a balance to be struck between two sets of conflicting reasons’.69 ‘Truth’, on this
account, is always in question for it is always possible that a stronger argument
will overturn what we at present take to be truth. Recognizing this should
properly breed a suspicion towards the claims of any one person or group to
know once and for all what is ‘truly’ in the interests of everyone else. The
consequence of Mill’s position here, as Berlin himself recognises, is that in the
ideological fields there can be ‘no single, universally visible, truth’ applicable
to all human beings.70
Berlin also recognizes the role that personal experience plays in Mill’s
argument.71 This is important in defending Mill because it shows that he does
not place us under the despotism of reason at the expense of the affections. Mill
himself exemplifies this. In his response to his mental crisis and against the
excessive rationality of the Benthamite creed, ‘the cultivation of the feelings
became one of the cardinal points’ of his work.72 Reason is important, but it is
not everything and a life of reason alone would be one-sided.73 Thus Mill rejects
the Kantian notion of autonomy as duty in accordance with an abstract and
putatively rational moral law and does not, as Kant did, equate autonomy with
rationality. Raz, like Mill, is also careful to point out that persons who have
personal autonomy ‘are not merely rational agents’, nor does the autonomous
life ‘necessitate any high degree of self-awareness or rationality’.74
Johnston sees in personal autonomy a requirement that individuals constantly
subject their lives to critical appraisal, and quite rightly objects to this on the
68
Mill, ‘On Liberty’, p. 112.
Mill, ‘On Liberty’, p. 96.
Isaiah Berlin, ‘John Stuart Mill and the Ends of Life’ (1959), reprinted in Berlin, Four Essays
69
70
on Liberty (Oxford: Oxford University Press, 1969), pp. 173–206, at p. 188.
71
Berlin, ‘John Stuart Mill and the Ends of Life’, p. 188.
John Stuart Mill, ‘Autobiography’ (1873), in John M. Robson and Jack Stillinger, eds, The
72
Collected Works of John Stuart Mill, Vol.1 (Toronto and London: University of Toronto Press and
Routledge & Kegan Paul, 1981), pp. 1–290, at p. 147.
73
Mill implies this in ‘Utilitarianism’ (1862), where judging between higher and lower pleasures
requires that one has experienced both kinds (Acton, ed., Utilitarianism, On Liberty, and
Considerations on Representative Government, pp. 1–61, at pp. 8–9).
74
Raz, The Morality of Freedom, pp. 154, 381; see also Raz’s ‘The Duties of Well-Being’,
in Joseph Raz, Ethics in the Public Domain (Oxford: Oxford University Press, 1994), pp. 3–28, at
p. 5.

308
APPERLEY
grounds that ‘the lack of desire to be in the habit of critically appraising
alternative values and alternative paths of life may itself be integral to a person’s
project or conception of himself as a person’.75 But Raz agrees. To illustrate the
point, he tells the story of a woman who is trapped on a desert island and who
is hounded every hour of every day by a fierce animal.76 In order to remain alive,
the woman must constantly tax to the limit her mental and physical powers. The
story shows that a life of constant reflection and critical scrutiny is an
impoverished life all the same, precluding many of the goods that any life,
autonomous or otherwise, might properly contain, such as relaxation, going with
the flow, and so on. Some goods, such as love, might even be irreparably
damagedifsubjectedtoconstantscrutiny.77 AsRazputsit, ‘autonomouspersons
are those who can shape their life and determine its course’.78 It does not follow
that they must do so, or that they must do so constantly. Some people might find
living the autonomous life stressful and may retreat from making choices.79 But
provided that the decision is made in accordance with reasons that you endorse,
and not because I have made your life so difficult that you have retired hurt, then
this is compatible with the autonomous life.
Nor is it the case, as Johnston claims, that personal autonomy requires that
all one’s goals and projects be unified.80 On the contrary, Raz explicitly
acknowledges that the autonomous life ‘may consist of diverse and heteroge-
neous pursuits’.81 It is difficult then to sustain the claim that personal autonomy
invites despotism in the name of the imposition of a rational or unified self, or
even in the form of a self that is constantly and critically reflecting upon its goals
and values.
More importantly, a state that really does value personal autonomy and seeks
to promote it as a good will in fact have good reason not to act despotically
towards its citizens, for it would actually undermine autonomy if it did so. For
this reason, departures from neutrality in the direction of liberal autonomy are
not equivalent to departures from neutrality in the direction of, say, a dogmatic
theological good.82 If there is an Enlightenment project associated with the
history of liberalism it may be understood as manifesting itself in the belief that
it really is better that individuals be capable of critically assessing their plans,
projects, beliefs and so on rather than simply bending to the will of an authority
just because it is an authority. On this reading of liberalism, the counter-concept
75
Johnston, The Idea of a Liberal Theory, p. 93–4.
Raz, The Morality of Freedom, p. 374.
This adapts an argument by Susan Mendus, ‘Some Mistakes about Impartiality’, Political
76
77
Studies, 44 (1996), 319–27, p. 322.
78
Raz, The Morality of Freedom, p. 154, emphasis added.
James Griffin, Well-Being: Its Meaning, Measurement, and Moral Importance (Oxford:
79
Clarendon Press, 1986), pp. 58, 82, 114, 137, 145, 236; Hayek, The Constitution of Liberty, p. 18.
80
Johnston, The Idea of a Liberal Theory, p. 75.
Raz, The Morality of Freedom, p. 371.
As Brian Barry’s notes in his ‘How Not to Defend Liberal Institutions’ in Brian Barry, Liberty
81
82
and Justice: Essays in Political Theory, Vol.2 (Oxford: Clarendon Press, 1991), pp. 23–39, at p. 38.

Liberalism, Autonomy and Stability
309
to autonomy is not heteronomy but rather immaturity as Kant outlines this in
his important essay ‘What is Enlightenment?’:
If I have a book to have understanding in place of me, a spiritual adviser to have
a conscience for me, a doctor to judge my diet for me, and so on, I need not make
any efforts at all. I need not think so long as I can pay; others will soon enough
take the tiresome job over for me.83
This, Kant believed, is a recipe for despotism, for where immaturity prevails ‘it
is all too easy for others to set themselves up as [our] guardians’.84 On this point
Mill and Kant agree.85 For both of these writers, immaturity actually facilitates
despotism whilst personal autonomy is an important defence against it. It is this
thought that underlies the liberal defence of freedom of thought, speech and
association. Access to knowledge, the capacity to evaluate it, and the
opportunity where necessary to participate in its production – these are amongst
the cornerstones of liberal autonomy, and also of liberal democracy.
Autonomy and Stability
We have addressed Johnston’s fears; let us now consider Rawls’s. As we have
seen, Rawls now takes it to be the case that ‘the problem of stability is
fundamental to political philosophy’. Given this concern, he fears that a state
that pursues a conception of the good will have to use coercion to do so and this,
he believes, would destabilize and fragment society. Now, clearly, liberal
democratic states can and do use coercive means in order to implement policies
that have been agreed amongst democratically elected representatives. Even in
the most legitimate of democratic states there will be those who disagree with
certain policies and who are prepared (perhaps for personal gain, perhaps for
principled reasons) to ignore or transgress against them, and the state will
legitimately employ coercive means to secure compliance. This is a fact of
political life in liberal democratic states and would presumably be a fact of
political life even if the state were neutral amongst conceptions of the good. It
would be surprising to find that modern liberal democratic states characterized
by Rawls’s ‘fact of pluralism’ did not contain intolerant conceptions of the good
for whom even a rigorously observed neutrality on the part of the state would
be an affront.
But bearing Rawls’s comments concerning the burdens of judgement in mind,
we should acknowledge that the idea of coercion itself may be problematic, both
in terms of what it is and of when it is considered appropriate. For example,
coercion may be understood as a particular form of power-relation where
83
Immanuel Kant, ‘An Answer to the Question: “What is Enlightenment?”’ in Hans Reiss, ed,
Kant: Political Writings, 2nd edn (Cambridge: Cambridge University Press, 1991), pp. 54–60, at
p. 54.
84
Kant, ‘What is Enlightenment?’ p. 54. Notice that we cannot press Kant’s argument into service
on behalf of the mature against the immature in the way Berlin fears, for it is precisely this move
that Kant rules out.
85
See also Raz, The Morality of Freedom, p. 156.

310
APPERLEY
compliance on the part of the ‘victim’ is secured by structuring a given situation
so that the costs of non-compliance outweigh those of compliance. Now, it is
true that non-payment of taxes is, by law, a punishable offence and so it might
be concluded that payment of taxes is secured by means of coercion. But if, for
example, one believes that it is one’s duty to the community to pay one’s taxes,
or that paying taxes is a requirement of social justice, then one’s behaviour is
not contingent upon the threat of punishment and one is not therefore coerced.
The point is that what counts as coercion may itself be the subject of controversy,
and this controversy may itself indicate a lack of agreement over what
constitutes the proper scope or subject of justice itself. Such conflicts of value
are not restricted to the spaces between rival conceptions of the good either, but
may also be internal to them.86 It does not then follow that supporters of discrete
conceptions of the good will themselves always be agreed upon what constitutes
a coercive imposition.
Rawls claims that ‘political power is always coercive power’.87 But if, as Raz
insists, coercion is ‘evil’ precisely because it invades personal autonomy, then
the state that wishes to promote personal autonomy as a good will, contra Rawls,
have good reason to seek non-coercive means to achieve this goal.88 Whilst it
would be foolish to deny that states can and do act coercively, it is worth
remembering that there are many other ways in which states can pursue their
objectives, and legislation can be permissive rather than prescriptive.89 A state
that wishes to promote autonomy amongst its citizens may consider, for
example, devolving decision making to various sub-state associations where
this enhances and is compatible with the aims of promoting autonomy.90 If Raz
is right that coercion represents an invasion of autonomy then the pursuit of
autonomy as a good leads to a presumption against the use of coercion precisely
because of the kind of good it is. Such a presumption will not rule out, and may
even generate, controversy around what constitutes a coercive imposition and
when such an imposition would be appropriate, but this is to be expected in any
open, democratic polity. Indeed, given the burdens of judgement it is reasonable
to expect this to be so.
86
On this point, see Kymlicka, Liberalism, Community and Culture, pp. 168–9; Susan Mendus,
‘Human Rights and Political Theory’, Political Studies, 43 (1995), Special Issue on ‘Politics and
Human Rights’ edited by David Beetham, 10–24, pp. 18–24; Amy Gutman, ‘The Challenge of
Multiculturalism in Political Ethics’, Philosophy and Public Affairs, 22 (1993), 171–206, pp. 175–6;
Charles Taylor, The Ethics of Authenticity (Cambridge, Mass.: Harvard University Press, 1991),
p. 72.
87
Rawls, Political Liberalism, p. 68.
Raz, The Morality of Freedom, pp. 155–6, 377.
Raz, The Morality of Freedom, pp. 25–6, 161; Hayek, The Constitution of Liberty, p. 144;
88
89
Frazer and Lacey, ‘Politics and the Public’, p. 244; cf. Jeremy Waldron, ‘Autonomy and
Perfectionism in Raz’s Morality of Freedom’, Southern California Law Review, 62 (1989),
1098–152, pp. 1138–41.
90
Paul Q. Hirst, Associative Democracy (London: Polity Press, 1994), pp. 49–56, 58.

Liberalism, Autonomy and Stability
311
CONCLUSION
I have argued that autonomy ought to be understood, not as a neutral background
assumption of liberal theory, but as a partially comprehensive conception of the
goodinitsownright. Ihave alsoarguedthat protecting andpromotingautonomy
need not lead to either a despotic paternalism, as Johnston fears, or to the
destabilization and fragmentation of society, as Rawls fears. I realize that the
stance I have taken in this article is largely negative in that I have not explicitly
defended personal autonomy as a good against possible rivals and I accept that
many of my arguments could conceivably be enlisted on behalf of conceptions
of the good other than that of personal autonomy. All I will say here concerning
this issue is that I think it unavoidable that states pursue some conception of the
good and so the question is what kind of good is it that liberalism ought properly
to be defending? Once again, conceptions of the good are not all alike; if
autonomy is a good it is one that specifies a set of capacities without, for
example, bending the exercise of these capacities towards ‘a place that has
somehowbeenpreparedforhumanityinadvance’.91 Assuch, itisanopen-ended
conception of the good rather than a closed one.
Finally, the goods of liberalism are worth defending, but the assumption that
the good of autonomy is already embedded in the public culture of liberal
democratic regimes, and the subsequent attempt to neutralize it by representing
it as a background feature of these regimes, smacks of a complacency that has,
in the wake of the collapse of the communist states, come to be associated with
a misplaced and misguided liberal triumphalism. History has not ended but has
moved on, and liberalism faces many challenges both from without and within.
Yet liberalism – as radicals such as Chantal Mouffe and others have come to
recognize – contains much that is worth defending.92 Acknowledging and
publicly defending the good of the good of autonomy restores political
liberalism to the status of a political enterprise.
91
Richard Rorty, ‘Solidarity or Objectivity?’ in his Objectivity, Relativism, and Truth:
Philosophical Papers, Volume 1 (Cambridge: Cambridge University Press, 1991), pp. 21–34, at
p. 28.
92
Chantal Mouffe, ‘Democratic Politics Today’, in Chantal Mouffe, ed., Dimensions of Radical
Democracy (London: Verso, 1992), pp. 1–14. See also Stuart Hall and David Held, ‘Citizens and
Citizenship’ in Stuart Hall and Martin Jacques, eds, New Times: The Changing Face of Politics in
the 1990s (London: Lawrence & Wishart, 1989), pp. 173–88, at pp. 184–8. Of course, liberalism for
these writers also contains much that is to be rejected.