john gray mill’s liberalism and liberalism’s posterity

JOHN GRAY

MILL’S LIBERALISM AND LIBERALISM’S POSTERITY

(Received 15 October 1998; accepted in revised form 17 May 1999)

ABSTRACT. It is argued that the moral theory undergirding J.S. Mill’s argument inOnLiberty is a species of perfectionism rather than any kind of utilitarianism. The conceptionof human flourishing that it invokes is one in which the goods of personal autonomy andindividuality are central. If this conception is to be more than the expression of a particularcultural ideal it needs the support of an empirically plausible view of human nature and adefensible interpretation of history. Neither of these can be found in Mill. Six traditionalcriticisms of Mill’s argument are assessed. It is concluded that in addition to depending onimplausible claims about human nature and history Mill’s conception of the good containsdisabling incommensurabilities. It is argued that these difficulties and incommensurabili-ties plague later liberal thinkers such as Isaiah Berlin and Joseph Raz who have sought toground liberalism in a value-pluralist ethical theory. No thinker in Mill’s liberal posterityhas been able to demonstrate the universal authority of liberal ideals.

KEY WORDS: enlightenment, liberalism, perfectionism, progress, utilitarianism, value-pluralism

Joseph de Maistre once observed that, when Rousseau asked why it was that men who wereborn free were nevertheless everywhere in chains, this was like asking why it was thatsheep, who were born carnivorous, nevertheless everywhere nibbled grass. Similarly theRussian radical Alexander Herzen observed that we classify creatures by zoological types,according to the characteristics and habits that are most frequently found to be conjoined.Thus, one of the defining attributes of fish is their liability to live in water; hence, despitethe occurrence of flying fish, we do not say of fish in general that their nature or essence –the “true” end for which they were created – is to fly, since most fish fail to achieve this endand do not display the slightest tendency in this direction. Yet in the case of men, and menalone, we say that the nature of men is to seek freedom, even though only very few menin the long life of our race have in fact pursued it, while the vast majority at most timeshave shown little taste for it, and seem contented to be ruled by others, seeking to be wellgoverned by those who provide them with sufficient food, shelter, rules of life, but not to beself-governed. Why should man alone, Herzen asked, be classified in terms of what mostsmall minorities here or there have ever sought for its own sake, still less actively foughtfor? This skeptical reflection was uttered by a man whose entire life was dominated by asingle-minded passion – the pursuit of liberty, personal and political, of his own and othernations, to which he sacrificed his public career and his private happiness.1

1 Isaiah Berlin,Four Essays on Liberty(Oxford: Clarendon Press, 1969), footnotepp. lix–lx.

The Journal of Ethics4: 137–165, 2000.© 2000Kluwer Academic Publishers. Printed in the Netherlands.

138 JOHN GRAY

Like any other political philosophy, liberalism presupposes a philosophy ofhistory. To be sure, different liberalisms will go with somewhat differentviews of history. Yet, in all their many varieties, liberal political theoriesexpress common beliefs or assumptions about human history, on whichtheir claims in crucial respects depend. The dependency of John StuartMill’s liberalism on a philosophy of history is beyond any doubt. In theIntroduction toOn Libertyhe asserts that “I regard utility as the ultimateappeal on all ethical questions; but it must be utility in the largest sense,grounded on the permanent interests of man as a progressive being.”2 HereMill himself affirms the dependency of his moral and political theory ona conception of progress – a dependency which Mill’s liberalism shareswith every other kind of liberal political philosophy. This is a truth of nosmall importance. Over the past quarter of a century or more, the dominantschool of liberal thinkers has aspired to a pure philosophy of right, in whicha liberal theory of justice relies at no point on claims about human natureor history. For these thinkers – of whom John Rawls has been the mostinfluential – the central values of liberal political morality can be defendedwithout invoking any conception of historical progress, and indeed do notpresuppose any philosophy of history. They can be grounded in a concep-tion of the person, and a theory of the principles to which rational personswould assent. If such a foundationalist project regarding liberal politicalmorality could be carried off, then liberalism would not need support fromany interpretation of history.

If Mill is right in grounding liberal theory in a claim about progress,then this Kantian project of a pure philosophy of right founders at its verybeginning. Further, one may reasonably suspect that the independence ofany philosophy of history that is claimed for itself by the dominant schoolof contemporary liberal theory is delusive. In fact, in sharp contrast withMill’s liberalism, it is a notable feature of postwar liberal theory in theEnglish-speaking world – which has been dominated by the Americanliberalisms of John Rawls, Ronald Dworkin, and Richard Rorty, each ofwhich owes much of its inspiration to Kantian conceptions – that it isindeed reliant upon a particular philosophy of history in which the ideaof progressive cultural convergence on a universal civilization is central,but this dependency is tacit and unexpressed, where it is not altogetherrepressed or denied. In signalling unmistakably the reliance of his liberalpolitical theory upon an idea of progress, Mill is – characteristically – farmore self-critical and more candid than his liberal posterity in our own day.

2 John Stuart Mill,On Liberty and Other Essays, ed. John Gray (Oxford and New York:Oxford University Press, World’s Classics, 1991), p. 15.

MILL’S LIBERALISM AND LIBERALISM’S POSTERITY 139

Mill’s affirmation points to the central and fundamental weakness ofall liberalisms, which is that their claim to a universal authority rests on ananachronistic and parochial Eurocentric interpretation of history. It was notunreasonable for Mill to invoke this interpretation of history – in which it isassumed that the adoption by non-Occidental societies of forms of scienceand technology, of literacy and numeracy, and of industrialism and urbanlife which originated in Europe will inexorably be followed, if only after along cultural lag, by acceptance of Western institutions and moral beliefs –since it was shared by virtually all of his contemporaries, including Marx.This Eurocentric historical philosophy, which identified European hege-mony with the advance of the entire species and understood progress as theuniversal adoption of Western institutions, beliefs and values, was a centralelement in the Enlightenment project that Mill – in this, at least, at one withhis father, the author of the Eurocentric History of British India – endorsedunequivocally. It animates, naively but pervasively, the dominant Americanliberalisms of our own time, and underpin their – increasingly anomalous– claim to universal authority. However understandable it was for Mill toinvoke this scheme of historical interpretation – writingOn Libertyin themiddle of the nineteenth century – it is manifestly unreasonable for us tofollow him in it now.

In Mill on Liberty: A Defence,3 I show an awareness of the limitation,and indeed the indefensibility, of Mill’s philosophy of history in the lastchapter of the book, when he tries to detach Mill’s doctrine of libertyfrom the conception of progress that underpins his liberalism as a whole.Yet his manoeuvre – so I now think – leaves Mill’s doctrine of libertydefenseless and without foundation. Though a good deal more coherentand credible than the slight and shallow liberalisms that have dominatedpolitical thought in the postwar period, Mill’s liberalism is unavoidablyimplicated in the falsehood of his Eurocentric historical philosophy, andthe attempt to sever the doctrine of liberty from this larger liberal theorywas bound to fail. The falsity of Mill’s philosophy of history – in whichmodernization and Westernization are conflated and there is an unshakableexpectation of cultural convergence on a universal liberal civilization – hasprofoundly subversive consequences for all forms of liberalism. It suggeststhat the independence of any philosophy of history claimed for their polit-ical philosophies by Mill’s liberal posterity – such as Rawls, later or earlier– is entirely spurious. If Mill’s liberalism founders on the falsity of itsassociated philosophy of history, so does all subsequent liberalism, insofaras it continues to make any claim to universality for its values. If Mill’s

3 John Gray,Mill on Liberty: A Defence(London: Routledge, 1983).

140 JOHN GRAY

liberalism fails because it rests on a mistaken conception of progress, theliberal project itself fails. Or so at least I shall argue.

THE TRADITIONAL CRITICISMS OFMILL’ S LIBERALISM:A REASSESSMENT

The traditional criticisms of Mill’s liberalism have rarely, if ever, beencriticisms of its underlying view of history. They have been criticismsof Mill’s attempt to derive liberal maxims from utilitarian principles, orelse criticisms of his liberal principles themselves. They have focussed onthe supposed impossibility of deriving a liberal political morality froma utilitarian moral theory, a Principle of Liberty from the Principle ofUtility. Or else they have attacked the Principle of Liberty itself for itsindeterminacies and consequent inability to serve the action-guiding usesMill demanded of it.Mill on Liberty: A Defenceargued that Mill’s was aspecies of indirect utilitarianism, that Mill’s “doctrine of liberty,” was anapplication of the theory of the Art of Life set out in the System of Logic,and that the argument ofOn Liberty traded heavily on the account ofjustice, and of the higher pleasures, advanced in Utilitarianism, publishedwell afterOn Libertybut written more or less contemporaneously with it.All of these central interpretive claims remain eminently defensible. Whatis now at issue is the viability of Mill’s doctrine, not its character as asystematic moral and political theory applied by Mill throughout his majorwritings. The “revisionist” interpretation of Mill’s moral and politicalphilosophy has been accepted as valid, in its most central aspects, by themajority of scholars. My present concern is with the substance of Mill’sargument, and its implications for liberal political philosophy, not withminutiae in its exegesis. This reflects my conviction that Mill’s argumentin On Liberty sets standards of rigor, resourcefulness and imaginationthat have not been matched by his liberal posterity – except, perhaps,in the work of Isaiah Berlin and Joseph Raz, in which liberal politicalmorality is grounded in a value-pluralist ethical theory, rather than anykind of utilitarianism, but in which something akin to Millian liberalismis, in differing forms, nevertheless recognizably renewed. My concernhere is with the assessment of Millian liberalism, as we find it in Mill andin later thinkers, and with the viability of the liberal project itself. It isimportant that we know why Mill’s project failed, if it did, and what are theimplications of its failure for other versions of the liberal project. These are


questions which most of the critics of the first edition of my book fail toaddress.4

Contrary to much in traditional Mill criticism, and in contemporaryliberalism, Mill’s most fundamental failure arises not from his misguidedattempt to give liberal political morality a foundation in utilitarian moraltheory, but from the dependency of his doctrine of liberty on a philosophyof history he shares with all, or nearly all, other liberal thinkers. This isnot to say that the central traditional criticism ofOn Liberty – that itsattempted derivation of the priority of liberty from the claims of utility fails– is without force. On the contrary, in a number of subsequent writings,5

I have acknowledged that, whereas the traditional interpretation of Mill asan unsystematic and muddled thinker is wide of the mark, the traditionalcriticisms of his project inOn Liberty retain a force that is not met bythe revisionist account of it which was proposed inMill on Liberty: ADefence. Nevertheless, it is not these traditional criticisms that mount themost decisive challenge to Mill’s liberalism, but rather the conception ofprogress on which – along with every other species of liberalism – it rests.

4 Among the Mill scholars, and moral and political philosophers, who have commentedon the first edition ofMill on Liberty: A Defence, the most noteworthy are: Fred Berger,Happiness, Justice and Freedom: the Moral and Political Philosophy of John Stuart Mill(Berkeley: the University of California Press, 1984); John Skorupski,John Stuart Mill(London and New York: Routledge, 1989); Maurice Cowling, “Preface to the SecondEdition,” Mill and Liberalism, second edition (Cambridge: Cambridge University Press,1990); C.L. Ten, “Mill’s Defence of Liberty,” in John Gray and G.W. Smith (eds.),J.S.Mill on Liberty in Focus (London and New York: Routledge, 1991); Wendy Donner,The Liberal Self: John Stuart Mill’s Moral and Political Philosophy(Ithaca and London:Cornell University Press, 1991); David Lyons,Rights, Welfare, and Mill’s Moral Theory(Oxford and New York: Oxford University Press, 1994).

5 John Gray, “John Stuart Mill: the Crisis of Liberalism,” inPlato to Nato: Studies inPolitical Thought(London: Penguin Books/BBC Books, 1984 and 1995); “Indirect Utilityand Fundamental Rights” and “Mill’s and Other Liberalisms,” in John Gray,Liberalisms:Essays in Political Philosophy(London and New York: Routledge, 1989), Chapters 8and 12; “Introduction” in John Stuart Mill,On Liberty and Other Essays; “Introduc-tion” by John Gray and G.W. Smith in John Gray and G.W. Smith (eds.),J.S. Mill OnLiberty in Focus(London and New York: Routledge, 1991). Also concerned with, orclosely relevant to Millian liberalism are: “An Epitaph for Liberalism” in John Gray,Post-liberalism: Studies in Political Thought(London and New York: Routledge, 1993), Chapter16; “A Conservative Disposition: Individualism, the Free Market and the Common Life”in John Gray,Beyond the New Right: Markets, Government and the Common Environment(London and New York: Routledge, 1993), Chapter 2, especially pp. 51–55; John Gray,Berlin (London: Harper/Collins, 1994); John Gray,Liberalism, second edition (MiltonKeynes: Open University Press, 1995), Chapter 6 and Postscript; and “After the NewLiberalism” and “From Postliberalism to Pluralism” in John Gray,Enlightenment’s Wake:Politics and Culture at the Close of the Modern Age(London and New York: Routledge,1995), Chapters 8 and 9.

142 JOHN GRAY

The central thrust of the traditional criticisms of Mill’s project inOn Liberty – stated classically by James Fitzjames Stephen inLiberty,Equality, Fraternity6 and echoed by John Plamenatz7 – is that it was anexercise in squaring the circle – the project of grounding a strong prin-ciple about the protection of individual liberty in the utilitarian concern forcollective well-being. As such it was foredoomed to failure. Against thistraditional view, it is the argument ofMill on Liberty: A Defencethat thereis nothing incoherent or misconceived in Mill’s project of a liberal utilitari-anism, provided his species of indirect utilitarianism be accepted, and hisdetailed arguments about the character and content of human happiness –in particular his account of the place of individuality, and its prerequisite,autonomy, in human well-being – are judged plausible. Mill’s projectcannot be dismissed as a mere mistake, or an inadvertence, on his part;his arguments must be considered, and shown to be wanting. Nor can thepicture of Mill as a confused eclectic thinker, which figures so prominentlyin much of the earlier secondary literature about him, be taken on trust –even if it turns out that his attempt at a systematic reconstruction of moraland political theory in utilitarian terms that squares with his judgmentsabout the importance of liberty does in the end fail. Contrary to the thesisof Mill on Liberty: A Defence, Mill’s project does indeed run aground onthe reefs charted by the most astute of his traditional critics; but the deeperreasons for its failure undo not only Mill’s but all other liberalisms.

It may be worth while specifying what are the central traditional criti-cisms of Mill’s project inOn Liberty, before going on to consider thosefailings of Mill’s liberalism that are not peculiar to or distinctive of it,but undermine all subsequent liberalisms. There are six defects of Mill’sargument in the essay, noted by the most perceptive among his traditionalcritics, which remain valid and disabling to Mill’s project. These are, first,that the Principle of Liberty Mill defends cannot give to individual libertythe priority and the equal distribution that any liberal morality requires, andthat Mill himself plainly desired; secondly, the prohibition on paternalistrestraints of liberty entailed by the Principle of Liberty cannot be givenany compelling utilitarian justification; thirdly, no evaluatively uncontro-versial or morally neutral conception of harm can be formulated of the sortneeded by the Principle of Liberty; fourthly, the account of human well-

6 James Fitzjames Stephen,Liberty, Equality, Fraternity, ed. Stuart Warner (Indiana-polis: Liberty Fund, Liberty Classics, 1993). A very useful collection of critical reviewsof On Liberty by Mill’s contemporaries has been published under the title,Liberty:Contemporary Responses to John Stuart Mill, ed. Andrew Pyle (Bristol: Thoemmes Press,1994).

7 John Plamenatz,The English Utilitarians(Oxford: Oxford University Press, 1949),Introduction.


being required by the view of harm specified in the Principle of Libertyis not an application of any utilitarian theory but the expression of anideal of the good life whose underlying ethical theory is perfectionist;fifthly, the account of human flourishing contained in this perfectionisttheory is unrealistic and implausible in privileging the human interestin autonomy; and sixthly, that Mill’s inability to provide any decision-procedure for resolving conflict among vital human interests, such asautonomy and security, renders his theory practically indistinguishablefrom value-pluralism of the sort we find later in Berlin and Raz. Thesetraditional criticisms of Mill’s project hit the mark; but they suggest others,which effect the ruin of the liberal project itself. Let us consider them,before we go on to look at the disabilities which Mill’s liberalism shareswith its liberal posterity.

The first of the traditional criticisms of Mill maintains that, even if thePrinciple of Liberty could be given a utilitarian justification, it would stillfail to accord individual liberty the protection that liberal morality demandsfor it. In the central tradition of liberal thought to which Mill belongs, it isconstitutive of liberal political morality that liberty be accorded priorityover other goods, and that it be distributed equally. Mill’s project inOn Liberty is to give this liberal morality a foundation in utility. If theformal claim of his indirect utilitarianism is the paradoxical one that thesuccessful maximization of utility demands adoption of maxims which barits direct pursuit by imposing constraints on its maximization, the materialor substantive claim Mill makes is that it is a liberty-protecting maxim –the Principle of Liberty – that best serves this indirect utilitarian strategy.The questions that arise here are: Can the adoption of Mill’s Principle ofLiberty be justified in these terms? And, even if it can, does that principlesucceed in protecting the priority and equal distribution of liberty whichMill – along with most other liberal thinkers of his day – thought necessaryin any civilized society?

Mill’s difficulty is that, though the attempt to derive a liberal preceptgoverning the restraint of individual liberty by law and opinion from thePrinciple of Utility is far from being self-evidently absurd, it comes upagainst the truth – a logical truth within Mill’s doctrine of liberty – thatsuch a principle can supply only a necessary, and not a sufficient condi-tion for the just restraint of liberty. The sufficient condition is that therestraint in question be supported by the Principle of Utility, or, moreprecisely, that it be maximally expedient in utilitarian terms. The Principleof Liberty tells us when restraint of liberty may be justified, not when itis just. This truth has the consequence that any coincidence between thePrinciple of Liberty, as that must be applied by Mill, and the standard

144 JOHN GRAY

content of liberal political morality, must be partly accidental. It is truethat, in forbidding restraint of liberty save when harm to others is at issue,Mill’s Principle of Liberty rules out any number of illiberal limits onliberty such as are demanded by paternalist and moralist considerations.It rules these out because, unless and until harm to others is at issue, noother consideration – and, in particular, no utilitarian consideration – caneven count as a good reason in favor of restraint of liberty. The problemis that, once the trip-wire set by the Principle of Liberty has been crossed,even trivial harms to others could sanction substantial restraints of liberty.The protection afforded the priority of liberty by Mill’s principle, thoughapparently stringent, is for this reason in reality slight. At the same time,in requiring that restraint of liberty be utilitarianly maximally expedient,Mill’s doctrine of liberty allows, and indeed may sometimes necessitatepolicies which result in a distribution of liberty that is grossly inequitableaccording to accepted standards of liberal morality. In particular, nothingin Mill’s doctrine requires that liberty, and restraint of liberty, be allocatedaccording to principles of equality. A restraint of liberty may prevent harmor harmful conduct, as required by the Principle of Liberty, and resultin a social distribution of liberty that is highly unequal. It may have thisconsequence, if the utilitarian calculus mandates some harm-preventingrestraint of liberty but dictates letting some harms to others be – as it wellmight. (If the propensity to harmful criminal behavior is in certain societiesmuch commoner in some groups than in others, might not expediencydictate an application of the Liberty Principle that discriminates unfairlyagainst individual members of these groups?) This suggests that, whateverelse it may be, Mill’s Principle of Liberty is entirely distinct, in its contentand implications for policy, from liberal principles – such as Rawls’sGreatest Equal Liberty Principle – whose provenance is Kantian ethicsrather than utilitarian morality. This, in itself, may be no bad thing. Yet, inleaving open the possibility that expediency may dictate a highly unequaldistribution of restraints on liberty, Mill’s doctrine surely fails to matchliberal intuitions that are widely found compelling, and which he himselfundoubtedly possessed. It is true that Mill’s doctrine of liberty may encom-pass an unstated Principle of Equity, as is maintained inMill on Liberty:A Defence;8 but, if it does, it has no compelling justification in utilitarianterms. A circumstance in which unfair discrimination against individualmembers of a minority social group is consistent with acceptable termsof cooperation and with stability in the larger society is, unfortunately,not difficult to imagine; and in those circumstances expediency may wellmandate unfairness. This is an objection to utilitarianism in all of its vari-

8 Gray,Mill on Liberty: A Defence, p. 67.


eties, which Mill’s indirect theory is no more successful than any otherutilitarian theory in answering.

Secondly, the ban on paternalist restraints of liberty – which Mill isconcerned to stress as one of the implications of his Principle of Liberty– seems impossible to justify by any kind of utilitarian reasoning. Mill’sargument against paternalism is an application of his two-tier, or split-level, species of indirect utilitarianism, which allows that the adoption ofutility-barring secondary maxims may be defensible for, or even requiredby, utility-maximization. There are a good many problems with indirectutilitarian theories, most of which turn on the difficulty of insulating thepractical level of moral deliberation from the critical level of utilitarianevaluation. (These difficulties apply to all forms of indirect consequen-tialism, and not merely those in which it is joined to a welfarist theoryof value.9) Despite the elaborate apparatus of the theory of the Art ofLife developed in the System of Logic, and the revisionary conception ofmorality in terms of enforceable obligations and of Utility as an axiolog-ical principle which that theory incorporates, Mill’s indirect utilitarianismtends to disintegrate when confronted with the fact that an appeal tothe Principle of Utility is unavoidable where the maxims of the variousdepartments of the Art of Life come into competition with one another.At that point, the purely axiological character of the Principle of Utilitycannot be sustained, and Mill’s theory of right action collapses into themore familiar theory of sophisticated act-utilitarianism, with all of itswell-known difficulties.

At this stage in my argument, however, I do not wish to questionthe cogency of indirect utilitarian theories. I do not need to do so, sincethe ban on paternalist restraint on liberty would be indefensible evenif indirect utilitarianism were true. Even if we suspend our disbeliefregarding indirect utilitarianism, and allow that it is reasonable in indirectutilitarian terms to adopt maxims barring utility-maximizing policies, wedo not know which maxims we are to adopt; and Mill’s arguments in favorof adopting the Principle of Liberty as such a utility-barring maxim are farfrom compelling. His argument in favor of a prohibition on the restraintof liberty in regard to self-harming conduct appeals to human fallibility,invoking the epistemic difficulty we face in identifying circumstances inwhich paternalist intervention can be successful. This is the difficultyarising from our imperfect knowledge, in virtue of which we cannot be surethat a paternalist intervention is warranted in a particular case even whenwe know, or have good reason to believe, that it can be justified in a whole

9 Some of the difficulties of indirect consequentialism are illuminatingly discussed byAlexander in his “Pursuing the Good-Indirectly.”

146 JOHN GRAY

class of cases. Mill appeals also to the practical difficulty confrontingsociety – whether embodied in a legislature or in public opinion – inacquiring, and in using expediently, such knowledge as is available. Evengranting some force to these arguments, they do not show that a ban onall paternalist restraint of liberty is justifiable in indirect utilitarian terms.For it is reasonably clear that we know that some behaviors are nearlyalways seriously self-harming, and that their prohibition is demanded byutilitarian morality where its enforcement is feasible and not overly costly(prohibiting the use of crack cocaine falls into this category, very probably,in some countries). There seems no insuperable difficulty in our acquiringthe knowledge needed to make such judgments, nor in society – law andpublic opinion – implementing them. No doubt we need to be consciousof the costs and risks incurred in any paternalist policy; but they are not sogreat, in every case, as to warrant a blanket prohibition on paternalism.Note that the ban on paternalist restraint on liberty, which is a centralelement of Mill’s doctrine of liberty, is unreasonable in indirect utilitarianterms, even if – contrary to Stephen and a myriad other traditional critics ofMill since Stephen – a clear, workable and morally significant distinctioncan be made between self-regarding and other-regarding actions. Even ifself-harming conduct injures the interests of the agent and of no-one else,it is utilitarianly unreasonable to rule out of court restraint of the libertyto engage in it. Or, to put the same point in other terms, if banning allliberty-limiting paternalism is indeed a necessary ingredient in any liberalpolitical morality, it is one that cannot be given a utilitarian justification.

An implication of this second criticism of Mill’s project inOn Libertyis that it is not the indeterminacy of the self-regarding area – the sphere ofconduct in which what an agent does affects only his or her interests, andno-one else’s – that is the chief objection to the Principle of Liberty, as thatis intended by Mill to work in the context of his broader theory of liberty.It is that sticking to the Principle of Liberty would be unreasonable inutilitarian terms, even if the self-regarding sphere could be determinatelymarked out. Nevertheless – and this is the third weighty traditional criti-cism of the liberal project undertaken inOn Liberty– there are disablingindeterminacies in the Principle of Liberty, arising from the conceptionof harm that it incorporates. The central difficulty is that there is noconception of harm that is neutral between different moral outlooks. Mill’sargument inOn Liberty presupposes that a conception of harm can beformulated which enables a utilitarian calculus of harms to be operatedwhich depends at no point on controversial conceptions of human well-being. The conception of harm that Mill needs must be – so to speak– empirical. Mill’s difficulty – which is not resolved in Joel Feinberg’s


proposal, advanced in the context of his resourceful contemporary restate-ment of a Millian jurisprudence,10 that harm may be understood as set-backto interests – is that no such account of harm can be framed. Even if aconception of harm as set-backs to interests could be adequately formu-lated, judgments about the relative weight of different harms would stillvary with divergent moral outlooks and their associated ideals of life.For assignments of comparative weight to different harms will express, orembody, differing judgments as to the weights of various human interestsin human well-being, while differing judgments about the contributionmade to human well-being or flourishing of particular interests will expressdifferent conceptions of the human good. If this is so, it is a devastatingblow to the Principle of Liberty, since it deprives it of the chief use Millhoped for it – that of settling issues about restraint of liberty that arisebetween people of different moral outlooks. Reasoning about how libertyshould be limited cannot then be neutral between competing conceptionsof the good; it cannot avoid making substantive moral claims about thecontent of human well-being. This is a defeat, not only for the liberalproject which Mill undertakes inOn Liberty, but for any liberalism whichclaims for its principles that they occupy a space of neutrality between rivalideals of human life.

This third criticism implies that the “one very simple principle”11 Millaimed to state in the essay governing social control of individual libertycannot, in fact, be formulated. The Principle of Liberty looks definiteenough in proscribing restraint of liberty except when harm to others isat stake; but determining when restraint of liberty is justified turns out tobe an intractably difficult business, since it hinges on assessments of therelative severity of harms that, in their dependency on disputed concep-tions of the good life, are inherently controversial. This disability of Mill’sprinciple suggests that it borrows whatever determinacy it possesses in itsapplications from a particular view of human well-being. That this is thecase and that this view of well-being cannot be defended in even the modi-fied utilitarian terms that Mill invokes in its support, are the bases of thefourth criticism made of Millian liberalism by its traditional critics – thatit is not a derivation from any utilitarian morality but a free-standing idealof human life. Millian liberalism is, in other words, a political concep-tion whose undergirding moral theory, if it has one, is perfectionist rather

10 Joel Feinberg,The Moral Limits of the Criminal Law, Volume 1,Harm to Others(Oxford: Oxford University Press, 1984).

11 Mill, On Liberty and Other Essays, p. 13.

148 JOHN GRAY

than utilitarian.12 In Mill on Liberty: A Defence, it was argued that Mill’smoral theory remained throughout want-regarding: “Mill is committed tothe proposition that men who have tasted the advantages and pleasuresof liberty will not trade them away for other benefits . . . provided Mill’sprediction holds up in the generality of cases, there is nothing ideal-regarding in his conception of happiness.13 The difficulty – which in myjudgment is insoluble for Mill – is that we have no evidence which couldsupport such a strong claim. True, the author ofMill on Liberty: A Defencerecognizes that the evidence is lacking which could back up Mill’s claim:he notes that Mill’s doctrine “. . . can claim for itself only that it representsa not unreasonable wager.”14 In contrast, what is striking to me now isthe absence in Mill of any evidence whatsoever for this bold claim. Wecan make sense of this extraordinary omission, only if we accept thatMill did hold to “an ideal of personality independent of its contributionto want-satisfaction.”15

This result is corroborated when we note that under critical pressureMill’s theory of value, and thereby his liberalism, collapses ultimately,and against all his hopes for it, into a species of perfectionist ethics. ForMill’s qualitative hedonism, adumbrated in Utilitarianism in the doctrineof the higher pleasures, cannot in the end be sustained – a result that hasthe closest relevance to Mill’s liberal political philosophy. The theory ofthe higher pleasures has many difficulties, some of them fatal to it. Theimplication of the theory, that any amount of a higher pleasure, no matterhow small, is worth more than any amount of a lower pleasure, howeverlarge, cannot be squared with any utilitarian calculus. It confers on thehigher pleasures an infinite weight, or lexical priority, as against the lowerpleasures, that makes comparative judgments of different bundles of higherand lower pleasures impossible except in limiting and marginal cases.16

Moreover, it is unclear whether what an experienced judge chooses isevidence for what the higher pleasures are, or criterial for them. In eithercase, Mill’s evident assumption that experienced judges will converge onpleasures of the same kinds – intellectual, imaginative and moral pleas-

12 I use the term “perfectionism” here in the sense it has acquired in recent philosophy,in which it means any moral theory whose core is a conception of human flourishing. Itis roughly this idea, in which the good for human beings is given by the demands of theirnature (or natures), that is the subject matter of the best available study of this kind ofmoral theory: Thomas Hurka,Perfectionism(Oxford: Oxford University Press, 1993).

13 Gray,Mill on Liberty: A Defence, p. 89.14 Gray,Mill on Liberty: A Defence, p. 89.15 Gray,Mill on Liberty: A Defence, p. 89.16 For an ingenious, but unpersuasive, counter-argument, see Roderick T. Long, “Mill’s

Higher Pleasures and the Choice of Character,”Utilitas 3 (1992), pp. 284–287.


ures, rather than bodily pleasures – has no ground in common experience,whether his or ours, and it is dubiously consistent with his repeated asser-tion, inOn Libertyand elsewhere, of the diversity of individual natures andneeds.

The crucial difficulty for Mill’s theory of value – because it is the onethat best reveals its distance from any kind of utilitarianism – is that bychoosing to develop the abilities that are most distinctive of them as indi-viduals people may well forfeit on-balance personal well-being. (It maybe that my abilities as a novelist are what most distinguish me from myfellows; but circumstances, or the modesty of my novelistic gifts, maymake developing my other, and less distinctive, abilities the course mostcongenial to my well-being.) In neglecting this possibility, or reality, Millmay be expressing a belief in the harmony of values that ill accords withhis focus on the importance of diversity. Or else – as Berlin has maintained,in the most sympathetic and profound statement of the traditional criticismof Mill 17 – he may be revealing that he values self-development and indi-viduality independently of their contribution to personal well-being andeven – sometimes – in competition with it. It is this last point that containsthe fundamental criticism of Millian liberalism – that it is not a derivativeof any utilitarian morality, even of the revised sort that Mill developed, butinstead the defence of a specific ideal or way of life – the way of life ofa liberal culture, in which autonomy and individuality, making choices foroneself and trying out “experiments of living” are valued as intrinsicallyimportant goods. If any definite moral theory underlies this ideal in Mill, itis perfectionist or eudaemonist – a theory of human flourishing, in whichit is claimed that human nature is most completely expressed in a societyin which the freedoms of autonomy and individuality are respected andprized. But what claim does this ideal have on us? Does it square withwhat we know of the conditions of human well-being?

The fifth criticism of Millian liberalism is that the conception of happi-ness or flourishing which its underlying perfectionist theory comprehendsis implausible and unrealistic in privileging autonomy as a necessaryingredient of human well-being. InMill on Liberty: A Defence, Mill’sconception of happiness is distinguished from any found in Benthamite orclassical utilitarianism by its requirement that human well-being consistsin the successful pursuit of self-chosen projects or activities, valued as endsin themselves and not as means to pleasurable sensations. Like Aristotle’s,with which it has much in common, Mill’s conception of human happiness

17 Isaiah Berlin, “John Stuart Mill and the Ends of Life,” in Isaiah Berlin,Four Essayson Liberty(Oxford: Oxford University Press, 1969), pp. 173–206.

150 JOHN GRAY

is one of active self-development, not passive enjoyment or contentment.18

Further, Mill’s account of happiness requires that the activities and projectspursued must be ones in which generically human powers and abilities aredeveloped and exercised, and, at the same time, that these pursuits expressthe abilities and needs that are distinctive of, or peculiar to each individual.Finally, and as a consequence of the two previous requirements, Mill’sconception of happiness implies that individuals who have experienced thehigher pleasures that go with being autonomous and developing their indi-viduality will not trade these off for any lower pleasure: they will alwaysprefer activities in which their generic and individual human powers ofautonomy and individuality are exercised over ones in which they are not.There are many difficulties with this conception, some of them arising fromthe problems with Mill’s qualitative hedonism discussed earlier, but I willfocus here on those that are most relevant to his defence of liberalism.

It is far from clear that any plausible account of human well-beingshould privilege choice-making over that which is chosen in the way thatMill’s does by requiring that the higher pleasures must be autonomouslychosen. Consider the example of arranged marriages and marriages that arematters of choice for the marriage partners. Mill’s conception of happinesswould seem to have the implication that arranged marriages could not beas happy as chosen marriages. For arranged marriages do not originatein the autonomous choices of the partners, and, if they satisfy the needsof the partners that are most distinctive of them as individuals, it is byaccident, not because the terms of the marriage reflect the autonomouschoices of the partners. Yet, if the experience of cultures containing botharranged marriages and marriages by choice is anything to go by, thereis no discernible difference between marriages of these two types as theyaffect the happiness – at least as that is ordinarily conceived and assessed– of the partners. Of course, there are many degrees, and even kinds, ofarranged marriages, and these will affect the autonomy of the partnersto varying degrees; but this does not affect the point at issue, which isthat happiness in marriage seems in many cultures as common – and asmuch a matter of chance – in arranged as in chosen marriages. Again,consider those circumstances in which an agent’s interest in security is incompetition with that in autonomy – circumstances, say, in which the costof job security is a permanent or long-lasting restriction of future options.It is implausible to suggest that such conflicts will not sometimes occur,or that, when they do, reasonable people will always put their interest in

18 For a helpful discussion of this aspect of human well-being, see James Griffin,Well-Being: Its Meaning, Measurement and Moral Importance(Oxford: Clarendon Press, 1986),Chapter 2.


remaining, or becoming, an autonomous agent over their other interests,such as that in security.

The truth is that common experience does not support Mill’s belief thatpeople will not trade off their autonomy for the sake of their other interests.Nor does experience suggest that they are unreasonable in doing so. It isplain enough that human behavior exhibits no consistent, or even discern-able preference for “higher” pleasures – activities involving the exercise,through the making of autonomous choices, of one’s generic and individualabilities – over “lower” pleasures, even when the agents have experiencedboth. Nor does Mill’s idea of “experiments of living” – whose functionin On Liberty is analogous to that of the “higher pleasures” in Utilitari-anism in privileging both the activity of choosing autonomously and itsoutcomes over other ingredients in well-being – lend any support to theclaim that autonomous agents will generally act to protect their autonomy,or to the claim that autonomous persons do better – in terms of well-being,where this can be assessed in reasonable empirical terms – than agentswho are not autonomous. In fact, it is doubtful if much clear sense canbe attached to Mill’s conception of “experiments of living,” since manydecisions to change the course of one’s life so alter, and alter irreversibly,the choosing self that a judgement of the “success” of the “experiment”– assessed in terms, say, of the agent’s well-being – cannot be made: theforms of life under assessment may well be incommensurable. In thosecases where a comparative judgment is feasible, however, it will by nomeans uniformly favor modes of life that have been chosen.The connectionbetween a person’s autonomous choices and that person’s well-being orhappiness is, on any empiricist assessment, a chancy and exception-riddenaffair – and certainly not one that warrants the privileging of autonomy inhuman well-being that Mill’s qualitative hedonism requires.

In Mill on Liberty: A Defenceit was argued that there were goodindirect utilitarian reasons for elevating the “vital interests” in securityand autonomy over other, less “permanent” human interests. At the sametime, it was acknowledged that the belief that autonomous agents willchoose to protect their interest in remaining autonomous over their otherinterests, when a conflict arose among them, could be no more than adefeasible presumption for Mill, or, for that matter, for us: “That menaccustomed to making their own choices will prefer to go on makingthem for themselves can only be for Mill an inductive wager, groundedin social-psychological conjecture.”19 Accordingly, it was proposed thatMill’s doctrine of liberty be detached or severed from his liberalism, sothat it no longer had the support of Mill’s “conviction of the inherent

19 Gray,Mill on Liberty: A Defence, p. 120.

152 JOHN GRAY

progressive character of human history and his belief in the irreversibilityof the condition of liberty.”20 Once his doctrine of liberty has been disso-ciated in this way from his larger beliefs about human nature and progress,his commitment to liberty “will depend on certain social and psychologicalconditions and hold good only in a cultural milieu where these conditionsare satisfied.”21

This is to accept, however, that the perfectionist ethical theory which,rather than any kind of utilitarianism, undergirds Mill’s liberalism cannotbe sustained in any general, still less universal way. Its account of thecontent of human well-being will hold good – if at all – only in certainspecific cultural milieu. According to this theory, human beings flourishonly when they are autonomous and have developed their individualities:and this is held to be a universal truth. If, however, this theory lacks theunderpinning in human psychology that Mill himself requires of it, thenthe ideal that his perfectionist theory stipulates will have no universal,cross-cultural authority. In other words, contrary to the argument ofMill onLiberty: A Defence, Mill’s doctrine of liberty cannot be severed from thebroader commitments of his liberalism, without depriving it of all universalauthority, and revealing its culture-bound particularity. If it is to be morethan the distillation of a particular cultural ideal, Mill’s liberalism needsthe support of an account of human nature. Otherwise, his conception ofhuman well-being will be indefensible in the terms of any general accountof human flourishing. It will be an ideal that is appropriate to, and indeedderived from a specific cultural tradition, that of European individualistsocieties.

The sixth criticism of Mill’s liberalism focuses on Mill’s retreatfrom the full force of the value-pluralism which his account of happi-ness intimates. It was recognized inMill on Liberty: A DefencethatMillian utilitarianism is committed to denying the incommensurability ofdifferent ingredients of human happiness: “It cannot be gainsaid that Mill’sdoctrine does not cope with. . . ultimate conflicts of value or with the prac-tical dilemmas they generate . . . It remains unclear what utility demandsbecause it is unclear how we are to weigh its competing elements.”22

Within a single human life, genuine goods are often rivals: how does Mill’squalitative hedonism assist us in choosing between them? The dilemmaposed for Mill’s ethical theory arises yet more starkly when genuine goodscannot be combined, or realized fully, within a single society or culture.Different mixtures of freedom and restraint, which is to say different liber-

20 Gray,Mill on Liberty: A Defence, p. 120.21 Gray,Mill on Liberty: A Defence, p. 120.22 Gray,Mill on Liberty: A Defence, pp. 126–127.


ties, will allow for different combinations of goods that are combinable,and different privations of those that are not. There can be no doubt thatMill wanted (and needed) a rational decision-procedure for the resolutionof such dilemmas: indeed, the whole project ofOn Liberty is to providea principle whereby conflicts among liberties can be subject to rationalarbitration. InMill on Liberty: A Defence, the divergence between Millianutilitarianism and value-pluralism was accepted, but the advantage was leftwith Mill’s doctrine of liberty. “Mill’s theory remains a utilitarian theory,distinguishable from value-pluralism, if only because of the crucial claimMill defends in utilitarian terms that the vital interests are in the circum-stances with which he is concerned always to be ranked over men’s otherinterests.”23

If, however, Mill’s ethical theory contains nothing which guides us inmaking choices among conflicting ingredients of happiness; if the generalpriority it accords, within the vital interests, to autonomy over securitycannot be given a rational justification; and if Mill’s doctrine cannot attachweights to different harms and rival liberties – then his revised utilitari-anism does, after all, collapse into an unwitting and incomplete form ofvalue-pluralism. The absence in Mill’s writings of any compelling accountof how conflicts between ingredients of human well-being that are uncom-binable, whether in one person’s life or in any given society, are to beresolved, pries open his ethical theory, undoes his revision of utilitarianismand amounts to a tacit admission of the truth of incommensurabilitiesamong the elements or ingredients of human happiness. Such an admis-sion is the ruin, not only of Mill’s project inOn Liberty, but also ofhis overarching project of a rational reconstruction, undertaken in revisedutilitarian terms, of moral and political life.

This is the conclusion reached by Berlin, as the result of his magisterialrestatement of the traditional criticism of Millian liberalism. Despite whatwas argued inMill on Liberty: A Defence, Berlin’s conclusion seems to menow to be unassailable, and its consequences for the ultimate unravellingof Mill’s liberalism unavoidable.24 As was emphasized in the first editionof this book, severing Mill’s Doctrine of Liberty from his liberalism hasinexorably the result that “the commitment to liberty which . . . his theoryembodies in respect of the future of mankind in general – can only havethe character of a wager.”25 However, far from the commitment to libertycontained in Mill’s doctrine representing what the author of the first edition

23 Gray,Mill on Liberty: A Defence, p. 127.24 For a more detailed and systematic consideration of Berlin’s argument, see myBerlin.25 Gray,Mill on Liberty: A Defence, p. 122.

154 JOHN GRAY

of this book termed “a not unreasonable wager,”26 it seems to me now thatMill’s wager is closer to Blaise Pascal’s famously bad bet than to any kindof empiricist hypothesis about the future of the species. Rather, it is an actof faith, expressing the religion of humanity Mill shared with the FrenchPositivists, which – though it continues to be made by the many lesserliberal thinkers among Mill’s posterity who claim universal authority forthe values of liberal culture – would for us be thoroughly unreasonableto make. If Mill’s doctrine of liberty fails, it is not only because many ofthe traditional criticisms of his argument retain considerable force evengranted the revisionary interpretation of his thought set out in the book butalso – and more importantly – because the liberalism that underpinned thatdoctrine, along with the Eurocentric perspective on human history whichit incorporates, is no longer a position that can be reasonably sustained.This can be demonstrated by considering those later liberalisms, whoseaffinities with Mill’s are not in doubt, which do not attempt to ground thevalues of a liberal culture in the requirements of utility but which followMill in elevating autonomy and individuality to the status of central andindispensable conditions of human flourishing.

MILL’ S AND LATER LIBERALISMS

Postwar liberal political philosophy has had a decidedly apologetic idiom,which distinguishes it sharply from Mill’s thought. In faithful conformitywith his deep-seated opposition to intuitionism in ethical theory, Mill neveradopted the conventional intuitions of his day or his culture as fixed pointsin his theorizing. On the contrary, he was at pains to mark the points atwhich his thought diverged from accepted opinion, noting inOn Libertythat his doctrine restricted individual liberty in areas where it was in hisday regarded as sacrosanct – such as freedoms in respect of procreation andthe education of children27 – even as it sought to protect it from control bypublic opinion in other areas of conduct. Much subsequent liberal philos-ophy, particularly in the past twenty years or so, has disregarded Mill’sexample, and has treated as fixed data of moral and political theory theintuitions of the liberal academy. This intuitionistic and conventionaliststance in liberal political philosophy has, indeed, been elevated to the statusof an explicit methodology in Rawls’s massively influential conceptions ofreflective equilibrium and overlapping consensus. Rawls’s work has set thetone for the dominant strand of postwar liberal philosophy, in its uncritical

26 Gray,Mill on Liberty: A Defence, p. 89.27 Mill, On Liberty and Other Essays, p. 117.


adoption as fixed points of theorizing the intuitions of liberal culture –more particularly, of American academic liberal culture, and in the claimto universal authority for liberal principles made by Rawls at least in hisearlier work.28 It is only in the writings of contemporary liberal thinkersworking outside this dominant tradition – such as Berlin and Raz – that wefind a contemporary restatement of liberal political philosophy that hasthe critical and historical self-consciousness which characterized Mill’sthought.

Since I am skeptical of the value and enduring significance of thisdominant strand of postwar liberal philosophy, I do not intend to give it anyextended discussion. I will instead confine myself to making two relatedobservations about its relations with Millian liberalism. In the first place,though Mill’s liberalism claims universal authority for its central prin-ciples, these are not principles which specify any set of fundamental rights,or structure of basic liberties, as being authoritative for all human societies.Mill’s Principle of Liberty is meant by him as a maxim for the guidance ofan ideal legislator, not as an exercise in constitution-framing, and it is clearfrom everything Mill wrote on political questions that he expected it toprotect different liberties in different historical and cultural circumstances.This is only a consequence of the logic of Mill’s doctrine of liberty itself,in which the freedom of self-regarding action is absolutely protected, butrestraint of liberty in other-regarding conduct is governed by utilitarianassessments which will, in their very nature, have different outcomes ascircumstances and the balance of utilitarian advantage change. For thisreason, there is no possibility, within Mill’s doctrine of liberty, of a listof basic liberties such as we find in the work of Rawls. Applying Mill’sdoctrine of liberty will yield very different results in different countries,and in the same country at different times, and entirely properly. In the vari-ability of the liberties it will protect, and in its refraining from the task ofspecifying with legalistic determinacy and finality the structure of libertiessupposedly demanded by justice, Mill’s liberalism has a clear advantageover that of Rawls, and of all other recent liberalisms whose unarticulatedassumptions are the local conventions of American constitutionalism.

28 See John Rawls,A Theory of Justice(Harvard: Harvard University Press, 1971) andPolitical Liberalism(New York: Columbia University Press, 1993). Other examples of thisapologetic mode in recent liberal political philosophy, in which an intuitionist approachto morality is married to an uncritical endorsement of the norms and values of Amer-ican liberalism, are to be found in Ronald Dworkin,Taking Rights Seriously(London:Duckworth, 1977);A Matter of Principle(Oxford: Oxford University Press, 1985);Law’sEmpire (London: Fontana, 1991);Life’s Dominion(New York: Alfred A. Knopf, 1993);B. Ackerman,Justice and the Liberal State(New Haven: Yale University Press, 1980); R.Nozick,Anarchy, State, and Utopia(New York: Basic Books, 1974).

156 JOHN GRAY

Secondly, Mill’s Principle of Liberty comes into play only when acertain level of cultural and economic development has been met, andthis is specified with explicit reference to his conception of progress.“Liberty, as a principle, has no application” Mill tells his readers to anystate of things anterior to the time when mankind have become capable ofbeing improved by free and equal discussion.29 Mill amplifies this point,later in the essay, by reference to China, of whose people he says “theyhave become stationary” and “if they are ever to be farther improved, itmust be by foreigners.”30 Again, in his “Considerations on RepresentativeGovernment,” Mill affirms that “Conduciveness to Progress. . . includesthe whole excellence of government.”31 By contrast with later liberalisms,Mill makes entirely manifest the dependency of his thought on a concep-tion of progress in which it consists in the universalization of Europeaninstitutions. Whatever the weaknesses of Mill’s liberalism – and they areweaknesses of all forms of political thought, liberal or otherwise, whichadopt the philosophy of history embodied in the Enlightenment project – ithas at least the virtue of seeking to justify itself by reference to an accountof human historical development which, because it is set out explicitly,can be the subject of critical evaluation. This degree of historical andself-critical consciousness cannot be claimed of any of the liberalismsto be found within the dominant strand of Mill’s posterity in our owntime.

In the work of Berlin and Raz we find a restatement of liberalism fullyas self-critical as Mill’s, in which – perhaps against their intentions –its dependency on a particular philosophy of history is little less evidentthan in Millian liberalism. Central to both Berlin’s and Raz’s work are anethical theory of value-pluralism and the assertion of freedom as the centralconstitutive value of liberal political morality. The idea of freedom isunderstood differently by the two thinkers, to be sure, with Raz followingMill, and joining many recent liberal theorists in embedding autonomy inthe very center of political morality, while Berlin reserves that place for hisconception of negative freedom.32 Raz situates his defence of autonomy

29 Mill, On Liberty and Other Essays, p. 15.30 Mill, On Liberty and Other Essays, p. 80.31 Mill, On Liberty and Other Essays, p. 223.32 For Berlin’s work, see especially hisFour Essays on LibertyandThe Crooked Timber

of Humanity(London: John Murray, 1990), and Claude J. Galipeau’sIsaiah Berlin’s Liber-alism (Oxford: Clarendon Press, 1994). For an illuminating discussion of morality that ishighly relevant to Berlin’s views, see Charles Taylor, “A Most Peculiar Institution,” inJ.E.J. Altham (ed.),World, Mind, and Ethics: Essays on the Ethical Philosophy of BernardWilliams (Cambridge: Cambridge University Press, 1995). I have discussed Berlin’s rela-tions with Mill in my book,Berlin. For the work of Joseph Raz, see particularly hisThe


in an explicitly perfectionist liberal moral and political theory,33 while inBerlin the underlying perfectionist view is less systematic in character,being expressed in a rejection of utilitarianism and of rights theories andan explicit and repeated endorsement of value-pluralism.34 Both writersacknowledge their debt to the Millian liberal tradition, with Raz developinghis perfectionist liberalism by way of a radically innovative reinterpreta-tion of Mill’s Principle of Liberty35 and Berlin recognizing many affinitiesbetween his own views and the half-conscious value-pluralism he finds inMill. 36

Both writers confront a difficulty in reconciling their strong pluralistaffirmation of the irreducible diversity, rivalry and incommensurabilityof human values with their assertion of the priority of liberty over othersocial goods. If goods and excellences are many, if some unavoidably andperhaps necessarily crowd out others, and there is no overarching principlewhereby these conflicts can be arbitrated, what can justify according toliberty – however conceived – a general priority over other social goodswith which it competes? If value-pluralism is true, is not liberty properlyviewed as one value among many, having no special privileges? True,liberty may be an essential and central ingredient in ways of life whichprize choice-making, which celebrate individuals as authors of their ownlives. Such ways of life express and embody particular human ideals; butwhat could warrant according such ideals – ideals of autonomy and indi-viduality, which animate in almost equal measure the different liberalismsof Mill, Berlin and Raz – the status of vital ingredients in human well-being? Both negative liberty and autonomy are defended, by Berlin andRaz, as expressing the ideal of human self-creation, in which persons areat least part-authors of their lives; and, particularly in Berlin, the valueof freedom is derived from the idea of man as essentially a chooser. Thecentrality of choice in the good life, and its associated image of man as a

Morality of Freedom(Oxford: Clarendon Press, 1986), andEthics in the Public Domain(Oxford: Clarendon Press, 1994). Raz’s work is discussed in S. Mulhall and A. Swift,Liberals and Communitarians(Blackwell: Oxford, 1992), Chapter 8. I have discussedRaz’s work in my bookEnlightenment’s Wake(London: Routledge, 1995), Chapter 6.

33 For contemporary studies of autonomy as it figures in recent liberal philosophy, see L.Crocker,Positive Freedom(the Hague: Martinus Nijhoff, 1980); John Christman (ed.),TheInner Citadel: Essays on Individual Autonomy(Oxford: Oxford University Press, 1989);Richard Lindley,Autonomy(London: Macmillan, 1986); Lawrence Haworth,Autonomy:an Essay in Philosophical Psychology and Ethics(New Haven: Yale University Press,1986); S.I. Benn,A Theory of Freedom(Cambridge: Cambridge University Press, 1989);and David Miller (ed.),Liberty (Oxford: Oxford University Press, 1991).

34 Isaiah Berlin, “The Pursuit of the Ideal,” inThe Crooked Timber of Humanity.35 Joseph Raz,The Morality of Freedom, pp. 412–420.36 Isaiah Berlin, “John Stuart Mill and the Ends of Life,” inFour Essays on Liberty.

158 JOHN GRAY

choice-making species, are, however, patently culture-bound conceptions.They are not features of the ethical life that is captured in theIliad or theBhagavad-Gitaor in cultures whose moral inheritance is Confucian. Givenits evident cultural particularity, why should this ideal of human life befavored over others – particularly by a value-pluralist who acknowledgesthe diversity of forms of genuine human flourishing? And why should thepolitical institutions in which this ideal finds embodiment be privilegedover others that express different – and, for a value-pluralist, sometimesno less legitimate – ideals?

The problem for a perfectionist liberalism, of the sort that Mill’s worktacitly expresses, and which the liberalisms of Berlin and Raz explicitlyadvocate, is that of accounting for the authority of the ideal of humancharacter and of life which they contain. This ideal cannot be defended inwant-regarding, utilitarian terms, and the very nature of perfectionism asan ethical theory in which a conception of the good is primordial precludesany foundational role for rights. Given the highly developed historicalconsciousness of both of these writers, it would be incongruous in theextreme for either of them to follow the dominant tradition of contem-porary Kantian liberalism in assuming that only an autonomous agent canlive a good human life – and, in fact, neither of them does so. If they do not,however, how do they avoid the relativistic position of Rorty,37 in whichliberalism is represented as only one form of life among others – if, as weshall see, a form of life with unique historical privileges?

It is in the work of Raz that the most systematic and comprehensiveattempt is made to answer these questions.38 For Raz, the ethical theoryunderlying liberal political morality is a perfectionist one, inasmuch as itadvances a particular ideal of human character, and is neither rights-basednor want-regarding in content. As he puts it:

Perfectionist liberalism has firm moral foundations. On the one hand, on this conceptiongovernments’ function is to protect and promote, within the bounds of their competence,the well-being of people. On the other hand, people prosper through a life of self-definitionconsisting of free choices among a plurality of incompatible but valuable activities,pursuits and relationships, i.e., a plurality of valuable and incompatible styles and forms oflife.39

Within the framework of this broadly Aristotelian account of ethicaltheory, in which ideas of human flourishing are fundamental, and the

37 Richard Rorty,Contingency, Irony and Solidarity(Cambridge: Cambridge UniversityPress, 1989).

38 I have argued that Berlin necessarily fails to derive liberal political morality fromvalue-pluralist ethical theory in myBerlin, Chapter 6.

39 Raz,Ethics in the Public Domain, p. 105.


priority of the right over the good that is affirmed in Kantian ethics isrejected, Raz advances a perfectionist liberalism, in which what distin-guishes liberal culture is its promotion of autonomy. What is it in Raz’sAristotelian account, though, that justifies the elevation to central place inpolitical morality of the good of autonomy? How does Raz defend thisparticular perfectionist conception? In Raz’s writings we find not one, buttwo distinct arguments for the centrality of autonomy in political morality.The first is a functional argument – the argument that skills of autonomouschoice are functionally indispensable to personal well-being in a societymarked by mobility in occupations and abode, innovation in technologyand forms of work, and more or less incessant change in beliefs and mores.The second is a cultural argument – the argument that autonomous choiceis indispensable to the well-being of persons whose cultural traditionhas inculcated a particular self-conception. These arguments are not fullydistinguished in Raz’s writings. Indeed, to some extent they overlap, in thatRaz argues that even the well-being of those who lack the understanding ofthemselves as autonomous agents requires the skills of autonomy, if theylive in a society in which most other people have that self-understanding.This is a variation on the functional argument, but one that presumably hasforce only in certain cultures.

It is unclear on which of the two arguments Raz’s perfectionist concep-tion, and the valuation of autonomy it encompasses, leans most heavily.He tells us that

In western industrial societies a particular conception of individual well-being has acquiredconsiderable popularity. It is the ideal of personal autonomy. . . It is an ideal particularlysuited to the conditions of the industrial age and its aftermath with their fast changingtechnologies and free movement of labor. They call for an ability to cope with changingtechnological, economic and social conditions, for an ability to adjust, to acquire new skills,to move from one subculture to another, to come to terms with new scientific and moralviews.

This sounds very much like a functional argument, but Raz goes on at onceto say

. . . it would be wrong to identify the ideal with the ability to cope with the shifting dunesof modern society. Autonomy is an ideal of self-creation. There were autonomous peoplein many past periods, whether or not they themselves or others around them thought of thisas an ideal way of being.40

Later in the same book, Raz asserts:

The value of personal autonomy is a fact of life. Since we live in a society whose socialforms are to a considerable extent based on individual choice, and since our options

40 Raz,The Morality of Freedom, pp. 369–370.

160 JOHN GRAY

are limited by what is available in our society, we can prosper in it only if we can besuccessfully autonomous. . . ultimately those who live in an autonomy-enhancing culturecan prosper only by being autonomous.41

In a subsequent book, he states:

We value autonomy to the extent that it adds to the well-being of the autonomous person.We regard the fact that a life was autonomous as adding value to it. We think of our ownlives and the lives of others as better for having been developed autonomously. But wevalue autonomous choices only if they are choices of what is valuable and worthy ofchoice.42

And, in a reply to his critics, he makes clear that he does not regardautonomy as a necessary condition of the good life, or even of the bestlife:

I think that there are, and there can be, non-repressive societies, and ones in which peoplespend their lives in worthwhile pursuits, even though their pursuits and the options open tothem are not subject to individual choice. Careers may be determined by custom, marriagesarranged by parents, childbearing and childrearing controlled only by sexual passion andtradition, part-time activities few and traditional, and engagement in them required ratherthan optional. In such societies, with little mobility, even friends are not chosen. Thereare few people one ever comes in contact with, they remain there from birth to death, andone just has to get on with them. I do not see that the absence of choice diminishes thevalue of human relations or the display of excellence in technical skills, physical ability,spirit and enterprise, leadership, scholarship, creativity or imaginativeness, which can allbe accomplished in such lives.43

What are we to make of Raz’s arguments? The functional argument isvulnerable to, and indeed falsified by, the evidences of diverse contem-porary Asian cultures. These have absorbed Western technologies andforms of scientific knowledge, and have achieved high levels of indus-trialism and urban life and of adaptation to continuous processes oftechnological and economic change, without accepting Western values ofautonomy and individuality. Such societies – of which Japan is only themost striking example, because it was historically the first, and remainsthe most successful – have modernized in that they have absorbed Westerntechnologies and adopted some Western institutions, but they have doneso without significantly compromising their own indigenous cultural tradi-tions, within which individualist values are not prized. They – Singapore,South Korea, and Malaysia, Taiwan, for example – have thereby over-thrown the Eurocentric historical philosophy, taken for granted by both

41 Raz,The Morality of Freedom, p. 394.42 Raz,Ethics in the Public Domain, p. 105.43 Joseph Raz, “Facing Up: a Reply,”University of California Law Review62 (1989),

pp. 1153–1235.


Karl Marx and Mill, according to which flourishing market institutionspresuppose an individualist moral culture. The functional argument forthe value is overturned by these examples, but it is also defeated by theexamples of Asian immigrants in Western liberal cultures, many of whomhave done better than their host populations by any standard apart from thatwhich invokes peculiarly Western ideals of autonomy and individuality. AsBhikhu Parekh has rightly argued:

As the cases of Japan, South Korea, Singapore and other countries show, some forms ofindustrialization do not require and are even best achieved without personal autonomy.They do, of course, require mobility of capital, labor and so on, but that has little to do withself-creation and self-ownership . . . The argument that autonomy is a functional require-ment of the modern society fares no better, for it treats autonomy as if it were no differentfrom such socially necessary amoral skills as literacy and numeracy, and denies it the statusof a moral value that Raz claims for it. The argument is also empirically false . . . In Raz’sterms, the Asian immigrants to Britain do not value autonomy. Yet their material successis remarkable and widely acknowledged. Indeed, they have prospered precisely becausethey do not set much store by autonomy and draw on the ample resources of a flourishingcommunal life and a readily available network of social support. As for personal well-being, the Asians do have their share of suffering and unhappiness, but no more or, somemight say, even less than their allegedly autonomous citizens.44

The functional argument for the value of autonomy is empirically falsified,not only for non-Occidental cultures of various kinds, but also for membersof Asian subcultures in Western liberal societies in which autonomy isprized. Indeed, if Parekh is right, members of such subcultures will dobetter, insofar as they do not adopt the ruling ideal of autonomy of theliberal cultures in which they live. To adopt it would not enhance butwould rather diminish their well-being. Far from it being the case thatthose who live in an autonomy-enhancing culture can only prosper bybeing autonomous, as Raz claims, it will be true of some of them thatthey will prosper better than the members of such a culture to the extentthat they remain immune to it. For such people, there is no functional orinstrumental argument to the value of autonomy, and therefore none whichcan appeal to their well-being.

Even if Parekh’s empirical claims are not accepted, his argumentreveals the very slender empirical base on which Raz founds his functionalargument for the value of autonomy, and how tenuous are the links – in thereal world even of liberal cultures – between personal autonomy and indi-vidual well-being. It seems anomalous that the central ideal of perfectionistliberalism should derive its value from so exception-ridden, defeasibleand often implausible or false empirical claims. The cultural argument for

44 Bhikhu Parekh, “Superior People: The Narrowness of Liberalism from Mill to Rawls,”Times Literary Supplement, February 25, 1994, p. 12.

162 JOHN GRAY

autonomy fares little better. It is that autonomy is a vital and indispensablecondition of individual well-being for persons who have a self-conceptionas authors of their lives – for people whose liberal cultures have (one mightsay) habituated them to making their own choices. Now it is true that thepersonal well-being of such persons is unlikely to be promoted by socialforms which injure their autonomy, since they conceive their well-beingin terms that are inseparably connected with their remaining autonomousagents. Yet this does not establish the value of autonomy even for suchpersons, since they may well be mistaken in their beliefs about the contri-bution made to their well-being by autonomous choice-making. This mustbe so, for anyone who is not a complete cultural relativist, and it is certainlyso for Raz, who is a moral realist who insists that well-being is – at leastpartly – objective. In his writings on multiculturalism,45 Raz has rightlyobserved that cultures must never be taken at their own estimation – in, forexample, the claims they make about the contribution of their distinctivesocial forms to individual well-being. This is surely no less true of liberalcultures, whose claims about the role of autonomy in the well-being oftheir members we are wise not to take at their face value. Accordingly,though it is true that a deep-seated belief in the worth of autonomy – likeany other deep-seated moral belief – must be taken into account whenevaluating the well-being of those who hold it and express it in their lives,the “fact of life”46 that autonomy is highly valued in liberal cultures in noway demonstrates its contribution to personal well-being on balance – noteven the well-being of those who hold it most dear.

In truth, Raz’s argument has most force if we attribute to him thebelief that autonomy is an historical fate, forced on us by other historicalforces – of modernization, for example – that are irresistible. However,though it is true that Western liberal cultures are pervasively animatedby the ideal of autonomy and the image – or illusion – of self-creationthat goes with it, this does not mean that modernization and autonomynecessarily go together, or that human flourishing and individual well-being have autonomy as a prerequisite in all, or even most modern cultures.It will be a prerequisite of well-being, if at all, only in cultures – suchas contemporary Western liberal cultures – whose social forms alreadyembody it, and which themselves enhance well-being. The proposition –which in Mill is explicit, and in Raz a tacit assumption which, if true,would make his argument more forceful – that all modern, or “civilized,”societies are bound to promote autonomy as a condition of the individualwell-being of their members has now been falsified by the history of some

45 Raz,Ethics in the Public Domain, Chapter 7.46 Raz,The Morality of Freedom, p. 394.


non-Occidental peoples and by the experience of Asian immigrants inliberal Western societies. As far as I can see, it is only this Eurocentrichistorical philosophy, according to which modernization entails the accept-ance of Western individualist values, that underpins – in Mill or in Raz –the valuation of individuality and autonomy as other than, and more than,ideals animating local cultural forms.

It is, again, only such a Eurocentric philosophy of history that explainsthe larger relevance claimed for themselves by more recent liberalisms,such as those of Rawls and Rorty. In Rawls’s more recent, Deweyanphase, no doubt, the explicit universalist claims of the earlier theory aremuted; but what supports the claim that the account of justice developed inRawls’s later writings has any enduring significance? Rawls is at pains tostress that, unlike Mill’s, his liberalism does not privilege comprehensivemoral ideals of autonomy and individuality;47 but the model of the personwith which the later, as well as the earlier writings work is patently adistillate of the individualist form of life of contemporary Western liberalcultures, and, most particularly, of the United States. Moreover, it is onlyby invoking this conception of the person – a human subject disembeddedfrom any constitutive communal attachment and emptied of any distinctivecultural or historical identity – that Rawls is able to move as he does fromthe historical fact of diversity in worldviews and conceptions of the good toa liberal state. For, taken by itself, the historical fact of pluralism supports,most naturally and reasonably, the Hobbesian project of seeking a peacefulmodus vivendirather than any Kantian project of framing a liberal constitu-tion to which all autonomous agents can give their rational assent. Withoutthe undergirding support of a Deweyan philosophy of history in whichWestern, and especially American, individualist social forms are presumedto be the historical fate of the species, Rawls’s account of justice has onlya local interest, as an articulation in systematic terms of the intuitions andself-conception of certain strata within American liberal culture. It has noauthority, and little interest, for anyone else.

The same is true for Rorty’s far more profoundly deliberated projectof a liberalism without foundations in which the contingency of liberalselfhood, discourse and community is fully acknowledged.48 This projectis conjoined in Rorty’s work with the idea of a “pragmatist utopia” –a “cosmopolitan world-society” that “embodies the same sort of utopiawhich the Christian, Enlightenment, and Marxist meta-narratives of eman-

47 Rawls,Political Liberalism, p. 78.48 Rorty,Contingency, Irony and Solidarity.

164 JOHN GRAY

cipation ended”49 – in a way that makes sense only if a Deweyan – orMillian – theory of progress is presupposed in which a liberal cultureis judged to be the best guarantor of the interests of the entire species.Otherwise, as with Rawls, Rorty’s project – the project of “an ideal liberalsociety (which) has no purpose except freedom”50 – would have only theinterest attaching to an articulation of a distinctively Western, and indeedparticularly American, individualist ideal. Rorty perceives that all formsof foundationalist liberalism elevate the local cultural forms of liberalsocieties to the status of universal demands of reason, utility or justice– and that such foundationalist projects in liberal political philosophy haveall failed. At the same time, he makes universal claims for American liberalculture that are defensible only in virtue of his trading on a Deweyanhistorical philosophy in which local American individualist cultural formsare conceived as the germ, or exemplar, of a universal or cosmopolitancivilization. If his Deweyan conception of progress is abandoned as confer-ring unique and unwarranted historical privileges on one particular culturalform, then Rorty’s liberal ideal ceases to be uniquely authoritative even forAmerica, in which liberal culture, though powerful, is – at least outsideacademic institutions – far from hegemonic. If this is so, then liberalcultural forms embody only one way of life among many, and have noauthority to dictate the terms of peaceful coexistence with other forms oflife.

The task of political philosophy is not then the apologetic task offinding bad reasons for what liberal academia believes by instinct butinstead the task of framing reasonable terms of coexistence amongdifferent communities and ways of life. The diversity of values that setsthe intellectual agenda for this task is not the anemic pluralism of life-plans celebrated in Rawls and in the American liberalism which Rawlsexpresses. It is rather the incommensurabilities and conflicts found in thereal world between whole ways of life and the conceptions of the goodwhich are embodied in the lives of historic communities. The project of apostliberal and pluralist political philosophy is that of theorizing conflictand the pursuit of peace among diverse cultures, communities and ways oflife. It is this pluralist project that I believe to be the historic successor ofthe liberal project, and liberalism’s true posterity. In this pluralist projectliberal cultures and forms of life enjoy no privileges.

The conclusion is inescapable that Mill’s liberal posterity, like Millhimself, depends upon a conception of progress that history has over-

49 Richard Rorty,Objectivity, Relativism and Truth: Philosophical Papers1 (Cambridge:Cambridge University Press, 1991), p. 209.

50 Rorty,Contingency, Irony and Solidarity, p. 60.


turned. For only a philosophy of history in which the universalization ofWestern values of individuality and progress is equated with the progressof the species can sustain the universal claims of liberalism when thefoundationalist projects that once engaged liberal thinkers – thinkers suchas Mill – have been abandoned. Perhaps other conceptions of progress arepossible, which do not privilege liberalism or Western values; or perhapsthe very idea of progress is rightly suspect as a barbarous relic of Enlight-enment. Either way, the empirical refutation of this Eurocentric philosophyof history by history itself means the ruin not only of Mill’s but of all liberalpolitical philosophies. As a great liberal thinker of our time has written,

It may be that the ideal of freedom to choose ends without claiming eternal validityfor them, and the pluralism of values connected with this, is only the late fruit of ourdeclining capitalist civilization: an ideal which remote ages and primitive societies havenot recognized, and one which posterity will regard with curiosity, even sympathy, butlittle comprehension.51

In understanding and accepting this truth – that liberal values are in nosense underwritten by history and have no claim to embody the permanentinterests of the species – we acknowledge that we belong not to Mill’sliberal posterity but instead to the posterity of liberalism.

The European InstituteLondon School of EconomicsHoughton StreetLondon, WC2A 2AEUK

51 Berlin, Four Essays on Liberty, p. 172.

john gray mill’s liberalism and liberalism’s posterity

Documents

nature of men

history mills conception

toground liberalism

john graymills liberalism

plausible view of human

isaiah berlin

liberal thinkers

political philosophy