The Diversity of Comprehensive Liberalisms

Gerald F Gaus. Handbook of Political Theory. Editor: Gerald F Gaus & Chandran Kukathas. Sage Publication. 2004.

Comprehensive Liberalisms

The distinction between ‘comprehensive’ and ‘political’ liberalisms, explored in the previous chapter, has become central to contemporary political theory. My aim in this chapter is to examine various ‘comprehensive’ liberalisms, with particular care to identifying in what sense they are comprehensive. As I have argued elsewhere (Gaus, 2003: ch. 7), the distinction between political and comprehensive liberalisms is elusive. Rawls repeatedly describes as ‘comprehensive’ ‘philosophical,’ ‘moral’ and ‘religious’ ‘doctrines’ (1996: xxv, 4, 36, 38, 160) or ‘beliefs’ (1996: 63). Indeed, so often does Rawls characterize comprehensiveness in terms of moral, religious and philosophical doctrines or beliefs that a reader may be tempted to conclude that a doctrine is comprehensive if and only if it is moral, religious or philosophical. But though it is tempting to understand ‘comprehensive conceptions’ in this way, it would be wrong. Rawls is clear that ‘the distinction between the political conception and other moral conceptions is a matter of scope; that is, the range of subjects to which a conception applies and the content a wider range requires’ (1996: 13). Comprehensive and general doctrines cover a wide range of topics, values and ideals applicable to various areas of life. Even given the terms of Rawls’s own analyses, rather than conceiving of comprehensive liberalisms as all relying on a fully comprehensive doctrine, it is better to conceive of them in terms of a spectrum of theories, from those that rely on something like a fully comprehensive view to those that rely on, say, only a general theory of the right. In this chapter I shall focus on the following versions of comprehensive liberalism:

  • Liberalism as a secular philosophy;
  • Liberalism as a philosophy of the good life;
  • Liberalism as a political theory derived from a specific moral theory;
  • Liberalism as itself a distinctive theory of the right or justice.

Liberalism as a secular philosophy is a distinctly radical conception, which in some ways is the paradigmatic ‘fully comprehensive’ liberalism. On this view, human reason leads to convergence on a theory of human life in society, which includes a metaphysics, an epistemology, as well as theories of morality and politics. On the other hand, liberalism as a theory of right is much more cautious about the extent that human reason converges; its more modest versions shade off into Rawlsian political liberalism. Thus I shall argue that the ‘comprehensive’ liberalism of A Theory of Justice (1971) was a distinctly ‘partial’ comprehensive view, which was not as comprehensive as many other varieties of liberalism.

Liberalism as a Philosophical System

John W. Chapman (1965) argues that all political theories are inherently comprehensive as they combine an account of social reality, epistemology, psychology and ethics to provide political diagnoses and prescriptions. Liberalism certainly has been understood as a political theory in this sense, a truly comprehensive liberalism—an overall theory of inquiry, social life, as well as the good life and political justice. I have argued elsewhere (Gaus, 2000b) that liberal theory over the last hundred years has been characterized by recurring debates about the psychologies, value theories, epistemologies and theories of self and society, as well as principles of justice that must or may form part of a truly liberal comprehensive philosophy. I shall focus here on two core aspects of this debate about liberalism as a comprehensive philosophy: whether there is a distinctively liberal epistemology or social metaphysics.

Liberal Epistemology

We can observe a split between two liberal epistemologies. The rationalistic camp is associated with ‘Enlightenment liberalism’; indeed, as Stephen Holmes points out, liberalism’s critics often associate it with a ‘hyperrationalism’ (1993: 247). Thus understood, liberalism not only manifests a faith in reason and science, it is an attack on superstition, custom and, importantly, religion. Thus the secular and anti-religious character of much liberal thought. Here ‘liberalism’ is ‘secular humanism.’ This sort of militant, confident rationalism is also associated with great confidence in the ability of humans to understand nature and control their social world.

Liberalism as secular humanism remains important today, though liberalism as a self-confident rationalism has been under attack by pluralists, relativists, postmodernists and pragmatists (see Gaus, 2003: ch. 1) However, in an interesting and surprising sense, the pragmatist liberalism of Richard Rorty (1991) and others, although it depicts itself as rejecting Enlightenment rationalism and epistemology, is nevertheless an inheritor of this conception of liberalism as an overall method for arriving at the truth. To be sure, pragmatism is a reaction to rationalism and representational views of the mind and knowledge; as Rorty stresses, our minds do not mirror nature, and truth is not a correct representation of nature (1979: 176-9). Nonetheless, truth is still the result of convergence in individual reasoning: what is true is what a certain sort of community of inquirers would converge on (Misak, 2000). So, while rejecting the specific view of reason and truth that characterized much Enlightenment thinking, twentieth-century pragmatist liberalism continues to identify liberal democracy with a certain mode of inquiry, and one which, when properly implemented, leads to a convergence of rational belief. Moreover, in the hands of liberals such as Dewey this mode of inquiry allowed society to obtain ‘conscious control’—for example, in the form of economic planning—over its collective life (1980: 87). Thus liberalism is understood as a doctrine about the convergence of rational inquiry that provides for a rationally ordered society.

According to F. A. Hayek, the flaw at the heart of such liberalisms is their faith in the ability of reason to understand and control complex social processes. It is, insists Hayek, ‘our ignorance’ that makes social rules necessary (1976: 20). Karl Popper (1945) made a similar charge against Plato, Hegel and Marx, namely that they failed to appreciate the limits of knowledge. Hayek and Popper, then, represent the other strain of liberal epistemology: an insistence that reason is limited, and our basic position is one of ignorance. In contrast to liberal rationalism, this cautious, fallibilistic liberalism is less apt to be militantly secular than tolerant of religion; it is more likely to stress the incremental and experimental nature of social policy than to advocate grand social reconstructions. And it is more likely to appreciate the market, as a device for coping with our constitutional ignorance, and less likely to be enamoured with state planning.

The Metaphysics of Liberalism

Throughout the last century, liberalism has been beset by controversies between, on the one hand, those broadly identified as ‘individualists’ and, on the other, ‘collectivists,’ ‘communitarians’ or ‘organicists’ (for scepticism about this, though, see Bird, 1999). These vague and sweeping designations have been applied to a wide array of disputes; I focus here on controversies concerning (1) the nature of society, and (2) the nature of the self.

Liberalism is, of course, usually associated with individualist analyses of society. ‘Human beings in society,’ Mill claimed, ‘have no properties but those which are derived from, and which may be resolved into, the laws of the nature of individual men’ (1963b: 879; see also Bentham, 1987: ch. I, s. 4). Spencer agreed: ‘the properties of the mass are dependent upon the attributes of its component parts’ (1995: 1). In the last years of the nineteenth century this individualist view was increasingly subject to attack, especially by those who were influenced by idealist philosophy. D. G. Ritchie, criticizing Spencer’s philosophy in 1891, explicitly rejected the idea that society is simply a ‘heap’ of individuals, insisting that it is more akin to an organism, with a complex internal life (1902: 13). Liberals such as L. T. Hobhouse and Dewey refused to adopt radically collectivist views such as those advocated by Bernard Bosanquet (2001), but they too rejected the radical individualism of Bentham, Mill and Spencer. Throughout most of the first half of the twentieth century such ‘organic’ analyses of society held sway in liberal theory, even in economics (see A. F. Mummery and J. A. Hobson, 1956: 106; J. M. Keynes, 1972: 275).

During and after the Second World War the idea that liberalism was inherently individualist arose again. Karl Popper’s The Open Society and its Enemies (1945) presented a sustained critique of Hegelian and Marxist theory and its collectivist and historicist, and to Popper inherently illiberal, understanding of society. The re-emergence of economic analysis in liberal theory brought to the fore a thoroughgoing methodological individualism. Writing in the early 1960s, James Buchanan and Gordon Tullock adamantly defended the ‘individualistic postulate’ against all forms of ‘organicism’: ‘This [organicist] approach or theory of the collectivity… is essentially opposed to the Western philosophical tradition in which the human individual is the primary philosophical entity’ (1965: 11-12). Human beings, insisted Buchanan and Tullock, are the only real choosers and decision-makers, and their preferences determine both public and private actions. The renascent individualism of late-twentieth-century liberalism was closely bound up with the induction of Hobbes as a member of the liberal pantheon. Hobbes’s relentlessly individualistic account of society, and the manner in which his analysis of the state of nature lent itself to game-theoretic modelling, yielded a highly individualist, formal analysis of the liberal state and liberal morality (see Buchanan, 1975; Hampton, 1986).

Of course, as is widely known, the last 20 years have witnessed a renewed interest in collectivist analyses of liberal society—though the term ‘collectivist’ is abjured in favour of ‘communitarian.’ Writing in 1985, Amy Gutmann observed that ‘[w]e are witnessing a revival of communitarian criticisms of liberal political theory. Like the critics of the 1960s, those of the 1980s fault liberalism for being mistakenly and irreparably individualistic’ (1985: 308, emphasis added). Starting with Michael Sandel’s famous (1982) criticism of Rawls, a number of critics charged that liberalism was necessarily premised on an abstract conception of individual selves as pure choosers, whose commitments, values and concerns are possessions of the self, but never constitute the self. Although the now famous, not to say infamous, ‘liberal-communitarian’ debate ultimately involved wide-ranging moral, political and sociological disputes about the nature of communities, and the rights and responsibilities of their members, the heart of the debate was about the nature of liberal selves. For Sandel the flaw at the heart of Rawls’s liberalism was its implausibly abstract theory of the self, the pure autonomous chooser. Rawls, he charges, ultimately assumes that it makes sense to identify us with a pure capacity for choice, and that such pure choosers might reject any or all of their attachments and values and yet retain their identity.

Throughout the 1990s various liberals sought to show how liberalism may consistently advocate a theory of the self which finds room for cultural membership and other non-chosen attachments and commitments which at least partially constitute the self (Kymlicka, 1991). Much of liberal theory has become focused on the issue of how we can be social creatures, members of cultures and raised in various traditions, while also being autonomous choosers who employ our liberty to construct lives of our own. What is important for our purposes is that these debates focus on whether liberalism entails an individualist theory of humans in society, or whether its political and moral commitments can be conjoined with various conceptions of the self and the social order; it is thus a debate about just how ‘comprehensive’ liberalism really is.

Liberalism as a General Theory of the Good Life

The Ideal of the Developed Individual

Over the last century and a half—say, roughly since John Stuart Mill—a great deal of liberal philosophy has been built on a particular view of human excellence. What might be called a perfectionist theory of the good life, or one devoted to self-realization as the end, can be found in Mill, T. H. Green, Bernard Bosanquet, L. T. Hobhouse, John Dewey and even, I would venture, in the third part of John Rawls’s Theory of Justice the most distinctly ‘comprehensive’ element of the book (Gaus, 1983a). The crux of this theory is presented in the third chapter of On Liberty, ‘Of Individuality, as One of the Elements of Well-Being,’ where human nature is compared to ‘a tree, which requires to grow and develop itself on all sides, according to the tendency of the inward forces that make it a living thing’ (Mill, 1963a: ch. 3). Mill closely ties individuality to this growth or development of human nature: ‘Individuality is the same thing with development’ (1963a: ch. 3). Mill believes that reason reveals our nature and its needs; human nature possesses impulses or energies that try to manifest themselves. Not only do we naturally possess different capacities, but these capacities are sources of energy that seek to express themselves. Consequently, to block a person from developing her capacities is to deenergize her—to make her passive and lethargic (1963a: ch. 3; Gaus, 1983a: ch. 4).

This perfectionist theory of the good life—the good life involves the perfection of human beings in society—has wide appeal in contemporary political theory. It was at the heart of William Galston’s earlier (1980; 1991) work and has been employed by Douglas B. Rasmuessen and Douglas J. Den Uyl (1991) as a foundation for a defence of classical liberalism. Although readers are often confused by Ayn Rand’s description of her position as ‘egoism,’ some idea of human perfection also seems foundational to Randian-inspired liberalism (Machan, 1989; Smith, 1995: 62ff).

Such perfectionist accounts of the good life are distinctly liberal in two ways. First, and most obviously, they provide the grounds for an argument for liberty. People need room to grow, room to find out which ways of living suit their unique natures and which do not. As Mill puts it, people need freedom to engage in ‘experiments in living.’ The lack of freedom will constrain growth, thus blocking human impulses and producing passive personalities. Second, such theories tend to place the individual and her choices at the centre of ethical life: liberalism is understood as a theory of ethical individualism. This is not to say that such theories see development as asocial; indeed, they often put stress on the way social life is necessary for complete development (Gaus, 1983a, chs 2 and 3; Kymlicka, 1991). Still, it is the individual and her self-realization or flourishing that has ultimate value, and individuals are not so deeply embedded in society as to make their choices a reflection of social history or culture (Sher, 1997: ch. 7).

Loosening the Telos: Theories of Personal Autonomy and Project Pursuit

Thus interpreted, Mill advances a quintessential Enlightenment argument: we can know human nature, and the knowledge of human nature provides truths about how we ought to live (Gaus, 2003; ch. 1; cf. Shapiro, 2003). Liberalism becomes identified with the promotion of a certain sort of self-realizing individual, one who develops her nature, is rational and suspicious of custom, experiments with different ways of living and is not prone to conformism. Two worries have been advanced about this as a conception of liberalism. First, its picture of the good life seems too specific and controversial to serve as a basis for liberal politics. Many in liberal societies are not devoted to the cultivation of individual perfection; in the face of this, liberalism seems to be a theory of the elite, which must struggle against the mass, uninterested in perfection. The mass of society, according to Mill, is a ‘collective mediocrity’: they tend to conform and are not interested in new ideas. The few who do think and invent are ‘the salt of the earth: without them, human life would become a stagnant pool’ (1963a: ch. 3, para. 10). Second, and following from this, such perfectionist theories raise the spectre of widespread paternalism. Although Mill argued for a strongly anti-paternalistic morality, it seems that the ideal is so specific and demanding as to open the gates to interferences with liberty, seeking to prod the mediocre mass towards a richer personality. It also becomes less than obvious why they should be granted liberty equal to that of the perfecting elite.

Many have argued that a defence of freedom based on personal autonomy is not subject to these objections. According to Joseph Raz, whereas Mill’s ideal of ‘[s]elf-realization consists in the development to their fullest extent of all, or all the valuable capacities a person possesses… [t]he autonomous person is one who makes his own life and he may choose the path of self-realization or reject it’ (1986: 325). The basic thought is that, according to the ideal of autonomy, it is not crucial that a person decides to develop her capacities, but that she decides whether to develop her capacities and, more generally, how to live her life. The fully autonomous person leads a life of her own choosing she makes decisions about her life on the basis of the things to which she is committed. The importance of liberty, argue tthe advocates of personal autonomy, is that it makes such a life possible.

The ideal of personal autonomy fractures into a variety of more specific doctrines (Lindley, 1986). Personal autonomy has been understood in terms of project pursuit, self-rule, self-creation and critical reflection on one’s projects and values, or consistency between first- and second-order volitions (on this last, see Gill, 2001: 20ff). Most conceptions of personal autonomy draw on several of these ideas.

According to Steven Wall, for example, ‘autonomous people need (a) the capacity to choose projects and sustain commitments, (b) the independence necessary to chart their own course though life and to develop their own understanding of what is valuable and worth doing, (c) the self-consciousness and vigor to take control of their affairs’ (1998: 132; see also Raz, 1986). And to Gerald Dworkin, ‘[w]hat makes an individual the particular person he is is his life plan, his projects. In pursuing autonomy, one shapes one’s life, one constructs its meaning. The autonomous person gives meaning to his life’ (1988: 31). Such visions of autonomy retain much of the structure of nineteenth-century self-realization perfectionism, while making less of the idea of a rich development of one’s capacities. The notion of a coherent plan of life was central to nineteenth-century self-realization theory (Gaus, 1983a: 34-44); the idea of a project or a plan points to a coherent and integrated set of ends. To the extent that a conception of personal autonomy presupposes a certain rational structure of ends, or a rationally constructed plan, it invites the elitist and paternalist objections raised against nineteenth-century liberal perfectionism.

These problems are mitigated by conceptions of autonomy according to which ‘the fundamental idea in autonomy is that of authoring one’s own world without being subject to the will of others’ (Young, 1986: 19). An autonomous person employs her critical faculties to evaluate and choose her aims and projects in such a way that they are truly hers, rather than simply imposed by, or unreflectively taken over from, others. Autonomy is thus understood as ‘an ideal of self-creation… Autonomy is opposed to a life of coerced choices. It contrasts with a life of no choices, or of drifting through life without ever exercising one’s capacity to choose’ (Raz, 1986: 370, 371). This conception of autonomy is thus a much more open-ended, and so less controversial, ideal than the ideals of either self-realization or project pursuit. Autonomy does not tell us what to choose; it only insists on the value of a chosen life. The worry, though, is that nobody really creates himself. Our personalities and choices are deeply influenced by our natural talents and propensities, our culture and our upbringing. What options we consider attractive are strongly affected by our upbringing and culture. As John Stuart Mill pointed out, it is ‘mere accident’ that decides the traditions into which one is inducted: ‘the same causes which make him a Churchman in London, would have made him a Buddhist or a Confucian [or, we might add, a Maoist] in Pekin’ (1963a: ch. 2, para. 4).

Given that we necessarily come to adulthood with values and commitments that we did not choose, Stanley Benn argues that an autonomous person is one who is engaged in an ongoing process of ‘critical adjustment within a system of beliefs in which it is possible to appraise one sector by canons drawn from another’ (1988: 32; but cf Wall, 1998: 128-9). On this view, a person who leads a self-chosen life is not really one who creates herself, but one who continually evaluates all her commitments and values to ensure that they are ones that she can continue to affirm in light of the other things she accepts. She cannot evaluate everything at once, but she can always be prepared to look critically at her values and projects to ask whether they are really things she is prepared to continue to affirm. Thus understood, a person’s life is not a freely chosen, autonomous life, if there are some parts of it she refuses to examine—if she has some commitments that she will not, or cannot, critically reflect upon. If she possesses some such commitments, they are ones that she cannot freely affirm, for her refusal to evaluate them indicates that she can only continue to affirm them so long as she is not free to revise or reject them. Benn recognizes, though, that this renders personal autonomy a character ideal that can be achieved to various degrees, and that many people fall far short of. Thus, in contrast to most liberal autonomists, Benn refuses to base liberal freedoms on autonomy, seeing it as a personal ideal, but not a foundation for basic liberal justice (1988: ch. 9).

What Makes Personal Autonomy a Liberal Theory of the Good Life?

The theory of personal autonomy, interpreted widely to include Millian self-development, is not simply a view of the good life that has been held by liberals, or even a view of the good life that justifies liberal political institutions. It is a distinctively liberal conception of the good life: the good life is a freely chosen life, and so the good life is a free life. It is, as Raz (1986) says, a morality of freedom; it puts a certain conception of a free life at the centre of morality. This is not to say that the autonomist project succeeds; as I have stressed, freedom qua autonomy seems to teeter on the verge of justifying elitism and paternalism, and so invites the sort of critique famously advanced by Berlin in ‘Two concepts of liberty.’ As Berlin quotes Kant, ‘paternalism is the greatest despotism imaginable’(1969: 157). Horacio Spector (1992) provides perhaps the most sophisticated attempt to show that a grounding in autonomy need not lead to such policies, but can justify strong classical liberal rights.

It is a mistake to try to define liberalism; liberal theories are complex clusters of conceptual and value commitments. But surely a crucial criterion for describing a view as ‘liberal’ is whether freedom is the core conceptual commitment (Freeden, 1996; Gaus, 2000a). Theories of personal autonomy, in insisting both that traditional liberal freedoms are necessary for personal autonomy, and that achieving autonomy is itself a type of freedom, thus have a strong claim to be advancing distinctively liberal moralities. Contrast this to Ronald Dworkin’s famous claim that liberalism rests on a basic commitment to equality, not liberty (1978: 115; see also 2000: Part I). According to his egalitarian liberalism, liberties such as freedom of speech and association are ways to achieve equal concern and respect. Their equal distribution is an instance of the general case for an equal distribution of resources and opportunities. The special status attributed to these basic liberties in liberal thought does not derive from the unique importance of freedom, but arises as a way to express equal concern and respect. It must be wondered whether liberalism is rendered more plausible by, first, almost entirely removing its traditional core of liberty and, second, replacing it with equality, a value that has traditionally had a complex and difficult place in liberal theory (Freeden, 1996: 241; Gaus, 2000a: 166-8).

Liberalism Derived from Moral Theories

A moral theory such as that of personal autonomy, I have argued, has a good claim to be a liberal conception of morality as it articulates a view of the good life that has at its core a notion of a free life. Reason should lead us to converge on a liberal understanding of the good life, and in that sense it is a ‘comprehensive,’ though not ‘fully comprehensive,’ view. Now a liberal theory of the good life and morality must be distinguished from a commitment to liberalism built on a moral theory; these two distinct conceptions of liberalism are often lumped together as ‘comprehensive’ liberalism. Liberal political principles can be derived from moral theories that themselves are not intrinsically liberal. I consider three such theories: utilitarianism, Hobbesian contractualism and value scepticism.

Utilitarian Liberalism: Reason and the Principle of Utility

Utilitarian moral theories hold that we can possess knowledge of both the good and the right; pace Rawls, these are not matters of ‘reasonable pluralism.’ The most straightforward versions of utilitarianism maintain that the good is either pleasure, happiness or preference satisfaction, and the right is the overall maximization of the good. Bentham, interestingly, did not think that the principle of utility could be proven; he did, though, contend that it could not reasonably be denied (1987: ch. 11, s. 11). Any reasonable person would see that pleasure is the ultimate end: consequently the principle of utility was beyond reasonable dispute. Whether or not the principle of utility could be established by reason was and is, though, a matter of dispute. Mill, famously, advanced a proof (1963c: ch. 4). Sidgwick, in contrast, insists that basic intuitions must be drawn upon in any argument for utilitarianism; in the end, Sidgwick appeared to accept that one could be an egoist and yet not irrational (1962: 418-22). It does seem, though, that if one accepts that (1) I value my own happiness, (2) because happiness is good, and (3) indeed is the only good and (4) more good is better than less good, then (5) assuming that we can interpersonally compare the happiness of different people, (6) one ought to seek the greatest happiness. Each step of this argument is controversial (and remains so regardless of whether ‘pleasure’ or ‘preference satisfaction’ is substituted for ‘happiness’); that, however, is a problem in utilitarian ethical theory that need not occupy us here.

Utility and Liberty: The Orthodox Critique

For present purposes the important question is whether, if accepted in ethics, the principle of utility justifies liberal political principles. Rawls (1971), of course, argued that it was, at best, uncertain whether a principle that aims at maximizing the aggregate amount of utility (happiness, pleasure, etc.) would yield an equal distribution of liberty. If greater happiness for many could be achieved by granting a few a lesser liberty, then the principle of utility would apparently justify illiberal policies. According to this common criticism, the problem is that the principle of utility takes the decision rule for choice by an individual—how he can best maximize his good—as a choice rule for society. While it is perfectly consistent with liberalism for an individual to sacrifice some parts of his life for more good in other parts, this seems illiberal in political decisions about the distribution of benefits between people. Thus the complaint that the principle of utility treats society ‘as a sort of single great person’ (Chapman, 1964: 163).

This criticism is so common as to be an orthodoxy in contemporary political theory. Utilitarian liberals, though, have advanced a number of replies.

The Theory of the Market

Rawls’s criticism of utilitarianism is so widely accepted that we are apt to forget that liberalism and utilitarianism marched hand-in-hand throughout the nineteenth century. Most of the great classical political economists were utilitarians of some sort (Gaus, 1983b), as probably are most economists today. The theory of the market is, in effect, a sophisticated argument showing that, under certain conditions, the best way to maximize aggregate utility is for each person to act to promote her own welfare. J. R. McCulloch stated the doctrine thus:

When individuals are left to be guided by their sense of what is best for themselves in the employment of their stock and industry, their interests are identified with those of the public; and those who are most successful in increasing their own wealth, necessarily, also contribute most effectually to increase the wealth of the state to which they belong. (1964: 125)

This supposes, of course, that wealth is a proxy for happiness or utility, an assumption often explicitly made by the liberal political economists (Gaus, 1983b) and indeed by much of economics today (which employs money as an objective utility function). We thus find today that, while most philosophers and political theorists are apt to see a conflict between liberalism and utilitarianism, most economists do not; neoclassical economics reconciles what philosophy draws asunder.

To be sure, it can be pointed out that the conditions under which market transactions maximize preference satisfaction are idealized, and so are not met in the actual world. Thus state action to correct market failures can be justified. Moreover, given the assumption of decreasing marginal utility (or decreasing rates of substitution between goods), utilitarianism can justify an egalitarian redistribution of incomes; P. J. Kelly (1990) argues that even before marginalism, Bentham’s utilitarianism endorsed moderate egalitarianism. Given the extent of market failure, utilitarianism can also support a more extensive state. On the other hand, the theory of public choice advances a theory of state failure (Mueller, 2003). If the government action itself suffers from serious failures to promote optimal outcomes, then the market’s failure to do so does not establish a case for intervention. Thus even in the face of serious market failures, a utilitarian may endorse relatively free markets.

Whether utilitarianism underwrites liberal politics and economics thus turns on economic theory, public choice, theories of institutional design (Goodin, 1996), and so on. In that sense liberal utilitarianism is indeed a partially comprehensive theory, with various theories of economics and politics being part of the case for liberal utilitarianism. Many philosophers are apt to reject liberal utilitarianism just because it turns on empirical claims; these anti-utilitarians often advance fanciful what if examples, showing that under strange circumstances, utilitarianism might lead to strange results. In contrast, utilitarians typically have high confidence in these theories, and see no reason to suppose that our theory of political right should be independent of our best empirical theories of economics and politics (Goodin, 1982).

Paretian Liberalism

Many standard economic utilitarian cases for liberalism presuppose interpersonal comparisons of utility. Although economists are sensitive to problems of comparing the utilities of different people, a great deal of political economy does so (Mueller, 2003: 566ff). If, instead, we suppose that the utilities (or happiness, or pleasure) of different people are incommensurable, we confront the problems of pluralism and social incommensurabilities. Even given incommensurable personal utilities, we still can make some minimal overall welfarist judgements. According to the Pareto criterion, (1) social state S1 is Pareto-superior to S2 if and only if at least one person is better off in S1 than in S2 and no one is worse off in S1 than in S2; and (2) if no state is Pareto-superior to S 1, then S1 is in the set of Pareto-optimal social states. These criteria, of course, may identify a large set of Pareto-optimal states, and so might often be indecisive among the choices open to us.

Even though it may often be indecisive, at least the Pareto criterion avoids the problem identified by Rawls: sacrificing the welfare of the few to benefit the many is excluded. Thus Paretian welfarism would seem at least consistent with liberalism. Amartya Sen, however, has proven this is not necessarily the case. Sen (1970) shows that if rights are understood as conferring individual authority to decide between at least two social states, then if there are two rights holders, liberal rights may conflict with unrestricted Paretian welfarism. Some have contested Sen’s characterization of rights as jurisdictions over the selection of social states (for a discussion, see Mueller, 2003: 650-51); recent work on the theory of rights, however, has shown the independent plausibility of such a jurisdictional theory (Mack, 2000; Gaus, 1996: 199-204). Sen’s result is important as it shows that in principle even a very minimal form of welfarism may be inconsistent with liberalism. More generally, Louis Kaplow and Steven Shavell (2002) have argued recently that almost any non-welfarist principle can conflict with the Pareto criterion.

Utilitarianism and Rights

Utilitarians, or more broadly, consequentialists, have spent a good deal of effort investigating in what ways personal rights might enter into a utilitarian system. Sen (1990) offers a version of consequentialism that takes rights satisfaction as part of the utility of a state of affairs (cf Scanlon, 1977; Nozick, 1974: 166). Mill’s complicated utilitarianism which seems to integrate rules into the concept of a morality—has often been used as a model for utilitarian rights (Lyons, 1978; Frey, 1984). Russell Hardin (1988; 1993) has advocated an ‘institutional utilitarianism’ that takes account of knowledge problems in designing utilitarian institutions, which he offers as an alternative to both act and rule utilitarianism. According to Hardin, ‘[w]e need an institutional structure of rights or protections because not everyone is utilitarian or otherwise moral and because there are severe limits to our knowledge of others, whose interests are therefore likely to be best fulfilled in many ways if they have substantial control over the fulfillment.’ That, he adds, ‘is how traditional rights should be understood’ (1988: 78).

L. W. Sumner (1987) presents an especially influential consequentialist case for rights. Sumner recognizes the paradoxical air of a thoroughly consequentialist argument for rights: in so far as the consequentialist seeks to maximize achievement of a certain goal, and rights are a constraint on the ways goals are achieved, it looks as if the consequentialist must argue that the best way to achieve the goal is to constrain our efforts to achieve it. The key to resolving this paradox, says Sumner, is to distinguish consequentialism as a theory of moral justification from the preferred theory of moral decision-making (1987: 179) or, we might say, consequentialism as a theory of evaluation from a theory of deliberation. This argument for rights consequentialism (or, more generally, rule consequentialism) argues that there is no easy transition from the claim that the right action is that which maximizes good consequences to the claim that the best decision procedure is to perform that action which one thinks has the best consequences. Thus, the best decision procedure from a consequential perspective may be to limit the extent to which one engages in consequentialist reasoning by constraining the pursuit of goals through the recognition of rights.

This type of argument was advanced by Sidgwick (1962: 489), who accepted that utilitarianism may be self-effacing, in the sense that it could instruct us not to encourage its use as a theory for making decisions. It may be better, he argued, if many people are guided by common sense morality. Two problems confront such a view. First, it is often not realized that rule utilitarianism puts more, not less, computational burdens on those devising the system of rules. To employ such a rule utilitarian approach, we have to know that following rule R is a good way to maximize, or at least promote, utility. But to do so, we need to forecast the utility of large sets of actions, as well as the expected rates at which people will ignore R or misapply it, the costs of teaching R and the costs of punishments. The problem is highlighted in Sidgwick’s own treatment of common sense morality, which makes Herculean assumptions about the tendency of common sense morality to promote the general happiness. Second, by divorcing utilitarianism as a standard of evaluation from its role as a standard of deliberation, we invite the sort of moral elitism that attracted Sidgwick: perhaps hoi polloi should be restricted to non-utilitarian reasoning, but the class of excellent calculators may be able to better promote utility by employing utilitarianism as a method of deliberation (1962: 489ff). Drawing inspiration from Sidgwick, Robert E. Goodin (1995: ch. 4) has recently defended ‘government house’ utilitarianism, which casts utilitarianism as a ‘public philosophy’ to be employed by policy-makers, rather than a guide to individual conduct.

Hobbesian Contractualism

As I noted above, the last 20 years has witnessed the induction of Hobbes as a core member of the liberal pantheon. In addition to his relentless individualist analysis on humans in society, the liberalization of Hobbes has been driven by his contractualism, and the way in which it lends itself to game-theoretic modelling, most importantly in the work of Jean Hampton (1986; for a discussion see Kraus, 1993). At first blush one might think that Hobbes did not offer a moral contractualist theory at all: the laws of nature are pre-contractual moral norms, and the contract concerns the institution of a political sovereign, not agreement on moral norms. However, as David Gauthier (1995) has stressed, the Hobbesian contract involves an authorization of the sovereign’s use of reason as right reason, including his reasoning about what morality requires; it is thus a political contract that subsumes morality. In any event, recent analyses inspired by Hobbes—most importantly Gauthier (1986)—have converted the Hobbesian approach into an account of justified morality which, in turn, endorses liberal arrangements (for doubts about the Hobbesian pedigree of Gauthier, see Lloyd, 1998).

As is well known, the starting point of neo-Hobbesian theory is a hypothetical analysis of the nature of unstructured interaction among rational individuals who are devoted to maximizing their preferences (Kavka, 1986: 123-4). There is some debate as to whether Hobbesian accounts suppose predominant egoism (Kavka, 1986: 64) or simply nontuistic preference maximization (Gaus, 1999: 12, 74; Wicksteed, 1946: vol. 1, 180). In such a situation, individuals will confront prisoner’s dilemmas, games of chicken and assurance games (Mueller, 2003: ch. 2; Skyrms, 1996); they will generally be unable to secure co-operation, and so the interests of all will suffer. Thus Hobbesian agents have reasons, based on the satisfaction of their own preferences, to agree to rules that structure their interactions and, in particular, that allow them to escape the prisoner’s dilemmas in which they find themselves.

Hobbesian contracts confront three main problems (Gaus, 1999; ch. 5; see also Kraus, 1993). First, given that life in the state of nature the condition of unstructured interaction—is ‘nasty, brutish and short’ (Hobbes, 1948: ch. 13), it will be relatively easy to show that some agreement is better than no agreement. However, there are likely to be a large number of social contracts that are Pareto-optimal: all are better than the state of nature, but none is Pareto-superior to the others. Some contracts will be preferred by some parties, different ones by others. Thus the Hobbesian contract may be indeterminate; if so, then it is doubtful whether it can be seen as a strong justification of substantively liberal principles. Second, given that Hobbesian individuals are single-minded rational maximizers, it seems that they would cheat on any social contract if they would benefit by cheating and their cheating will go undetected. But knowing this is the case, each knows the social contract would not bind anyone to sacrifice their interests, so there is no point in agreeing to it. Lastly, suppose, following James Buchanan (1975), one proposes the following contract: each keeps the holdings that each has in the state of nature, and agrees to call off the war of each against all. This would clearly benefit everyone, since each avoids the cost of protecting her holdings in the state of war. But it also seems unfair in the sense that it reflects the bargaining power of parties based on how well they did in the war that characterizes the state of nature. Such a bargain may be a modus vivendi a compromise among competing interests that produces peace—but it hardly seems the basis of morality (for a defence of the Hobbesian contract as a modus vivendi see Gray, 2000).

Sophisticated analyses such as David Gauthier’s (1986) contractualism seek to solve these problems (for general discussions, see Vallentyne, 1991). Gauthier argues that, in order to best pursue their goals, rational maximizers would agree to stop making maximizing choices. If individuals could adopt a disposition to obey the social contract the second problem, that of compliance, would be solved; once they have this disposition—this tendency to act—they no longer make choices by calculating what would best advance their goals, but on the basis of what would advance their goals in ways allowed for by the contract. If people adopted this disposition, then, somewhat paradoxically, they would do better at maximization, as they could honour the agreement that benefits all. Gauthier calls this ‘constrained maximization’ (1986: 158).

The obvious problem for Gauthier is that, at least on the face of it, the truly rational thing to do is to appear to turn yourself into a constrained maximizer while others really turn themselves into constrained maximizers. Gauthier has a two-part response. (1) Constrained maximizers do not adopt an unconditional disposition to constrain themselves no matter with whom they interact. They are only disposed to act in a constrained manner with those who are also constrained maximizers. (2) Gauthier insists that we are not totally opaque to each other; to some extent we can see into others and know their dispositions. As he puts it, we are ‘translucent.’ People can see through us to some extent. If one is translucent, the conditional co-operators will not act in a constrained way towards one if one is a cheater, and thus one will not achieve the benefits of social co-operation. Thus, concludes Gauthier, a rational agent would not seek to remain an unconstrained maximizer when others turn themselves into constrained maximizers.

It might seem that Gauthier would think that it is always rational to become an unconstrained maximizer when others do so. Not so. In Morals by Agreement Gauthier distinguished two dispositions to comply: broad and narrow (1986: 177ff, 225ff). Someone who is broadly compliant will comply with any agreement that benefits her; someone who is narrowly compliant will only comply with a fair and non-coercive agreement. As Gauthier sees it, it is not rational to become a broadly compliant person: you are asking people to take advantage of you, as you will honour any agreement you make regardless of how little you benefit. Thus we can see that Gauthier is addressing the worry about unfair contracts. It is only rational to turn yourself into a narrowly compliant person—someone who will comply with beneficial, fair and non-coercive arrangements.

Rational contractors will only comply with fair and rational bargains, but, as I have said, it is hard to know what is the uniquely best contract, the uniquely rational bargain. Gauthier originally defended a ‘minimax relative concession’ solution (1986: 226), but has since altered his views (1993: 178ff). Hobbesian accounts of justice thus seem dependent on some bargaining theory, and the complexities such theories involve (see Barry, 1989: Part I).

Value Scepticism

The idea that, somehow, liberalism could be intimately associated with scepticism about values, or some form of subjectivism, is controversial today: many important liberals such as Sher (1997) dispute it. Moreover, opponents of liberalism such as Alasdair MacIntyre (1981) have sought to make just this link, making contemporary defenders of liberalism suspicious about accepting some sort of tie. Nevertheless, scepticism about the interpersonal status of values has long been a part of liberalism. The sceptical camp includes all those liberalisms premised on the supposition that the powers of human reason are insufficient to provide public, definitive answers to the enduring questions concerning what makes life worth living, and to what ends we should devote ourselves. This line of liberal thinking can trace itself back to Hobbes and Locke. According to Locke:

The Mind has a different relish, as well as the Palate; and you will as fruitlessly endeavour to delight all Men with Riches or Glory, (which yet some Men place their Happiness in,) as you would satisfy all Men’s Hunger with Cheese or Lobsters; which, though very agreeable and delicious fare to some, are to others extremely nauseous and offensive: And many People would with Reason preferr the griping of a hungry Belly, to those Dishes, which are a feast to others. Hence it was, I think that the Philosophers of old did in vain enquire, whether the Summum bonumconsisted in Riches, or bodily delights, or Virtue, or Contemplation: And they might have as reasonably disputed, whether the best relish were to be found in Apples, Plumbs or Nuts; and have divided themselves into Sects upon it. (1975: 299)

Such subjectivist theories of value—which equate values with tastes or preferences—have a prominent place in twentieth-century liberal theory. A subjective conception of value was integral to the Austrian school; Carl Menger (1994: ch. 3) and his followers such as Ludwig von Mises (1966) explicitly endeavoured to integrate a subjectivist theory of value into economics. Of course in so far as economic liberalism is based on the supposition that the satisfaction of preferences alone determines value, then it too is subjectivist (for a criticism see Sunstein, 1997: 15ff). Subjectivist accounts of value have been defended by philosophers as well as economists: indeed it may well be that some form of subjectivism—that locates value either in the desires or in the feelings of agents—is the dominant account of value in twentieth-century philosophy (see Gaus, 1990: Part I). The upshot of these subjective accounts is that, by relativizing value to the desires, feelings or preferences of the individual agent, they undermine the proposal that the state should devote itself to pursuing the summum bonum. Liberal politics, on this view, cannot be reasonably grounded on pursuit of what is truly valuable, for value is a matter of taste, and our tastes differ. Of course such theories do not themselves lead to liberalism; although they act as ‘defeaters’ to arguments that seek to establish the justifiability of constructing politics around pursuit of the summum bonum, we still require positive arguments for liberal justice, to which we will now turn.

Liberal Theories of Justice

The line between a Hobbesian justification of liberal principles and what I shall call a ‘liberal theory of justice’ is fuzzy and open to challenge. The rationale for the distinction is this: utilitarian, Hobbesian and value subjectivist moralities may be employed to justify liberal arrangements, but depending on the details and assumptions, they can also justify distinctively illiberal policies. They thus require additional premises (say, the theory of the market) to ground liberal political principles. After all, Hobbes’s own theory was distinctively illiberal. In contrast, what I shall call ‘liberal theories of justice’ tie the very idea of justice and moral reasoning to basic liberal principles. On such views, although there may be no liberal theory of the good life or value, there is indeed a liberal theory of right.

Basic Liberal Rights Theories

I note in passing that Lockean natural right views (Nozick, 1974; Simmons, 1992) are certainly liberal accounts of justice. In general, rights-based theories of justice that give pride of place to individual liberty rights (Lomasky, 1987; Steiner, 1994) are distinctly liberal understandings of justice.

Harm and Liberty

In On Liberty J. S. Mill advances his ‘one very simple principle… that the sole end for which mankind are warranted, individually or collectively, in interfering with the liberty of action of any of their number… is to prevent harm to others’ (1963a: ch. 1, para. 9). Mill advances a radical liberal theory of political right: coercion—which includes social pressure intended to discourage any act A must be justified on the grounds that A constitutes a harm to others, and the coercion is intended to prevent that harm.

A good deal of recent liberal theory has been devoted to explicating this harm principle, and whether it really can serve as the sole ground for justified coercion. One dispute concerns whether Mill intends the principle to identify a set of acts those that do not directly harm others—that are immune from social coercion (see Riley, 1998: 93ff), or whether the principle is best interpreted as identifying a set of reasons harm to others—that can justify coercion (Ten, 1980: 50-7; Gaus, 1999: 106-13).

The classic work on the harm principle, and more generally on this Millian approach to political justice, is Joel Feinberg’s masterful four-volume The Moral Limits of the Criminal Law (1984-90). Feinberg’s four volumes carefully analyse the main issues in Millian morality. (1) Precisely what is a harm (1984)? (2) Does Millian morality allow coercion to prevent acts that, while not harmful to others, are offensive to some (1985)? (3) When individuals are unable to make fully voluntary choices, can coercion then be employed to stop them from harming themselves (1986)? And (4) are there any conditions under which liberals justify coercion that do not fall into one of the above categories (1990)? Feinberg convincingly shows that, when carefully examined, Mill’s radical proposal—that only harm to others can justify social interference—is implausible, but nevertheless is plausibly construed as the core of a liberal social morality (see further Gaus, 1999: Part II).

As Feinberg points out, moralities based on the harm principle are liberal in so far as there is a presumption of liberty: if a person’s action does not constitute a harm to others, then she has the right to act as she sees fit (1984: 9). Moreover, fundamental to the harm principle is the principle that where there is consent, there is no harm: thus one may consent to acts that set back one’s interests (such as taking drugs); not only does one have the right to harm oneself, but the dealer does not harm you if you have given informed consent to the purchase. However, critics of the harm principle (e.g. de Jasay, 1991) have argued that it is a poor grounding for liberal principles as the concept of harm is so malleable: it can be interpreted to encompass the prevention of psychological and culture harms (see e.g. Kernohan, 1997), thus justifying extensive and intrusive coercive interventions. Moreover, the requirement that the agent give ‘informed consent’ and that her self-harming acts are ‘voluntary’ opens the way to paternalistic interventions (Kleinig, 1983). Lastly, although the harm principle might be understood as a free-standing principle of morality that can be endorsed by those with very different views of the good life (Gaus, 1999: ch. 6), Mill and many of his followers have tied it to perfectionist theories, thus making it part of a more comprehensive package.

Kantian Liberalism I: From Respect to Liberal Rights

What Feinberg called the ‘presumption in favour of liberty’ has been defended by Benn in terms of a principle of non-interference based on respect for persons. Benn (1988) tells the story of Alan the pebble splitter, who is happily splitting pebbles on a public beach, when Betty comes along and demands that he justify himself to her. Benn agrees with Feinberg; he has no burden to justify himself to her. Now suppose she seeks to stop him, and he demands justification from her. Benn insists that a ‘tu quoque reply from her that he, on his side, had not offered her a justification for splitting pebbles, would not meet the case, for Alan’s pebble splitting had done nothing to interfere with Betty’s actions’ (1988: 87). There is, argues Benn, a basic asymmetry between you acting and you interfering with the actions of another. Alan does not have to justify his pebble splitting to Betty: he is under no standing requirement to show Betty that he has good reasons for what he is doing. On the other hand, it is required of Betty that she justify to Alan interfering with his actions, or stopping him for what he is doing.

Benn argues that Betty’s recognition of Alan’s right to act is required if she is to respect his person: ‘One may believe the other’s project quite worthless in itself. Its claim to respect rests not on its being valuable and worthy of one’s concern… but simply in its being a person’s project’ (1988: 107). Because one claims a right to non-interference for oneself, respect for others requires that one grants it to others. In a similar vein, Jeffrey Reiman concludes that ‘it is rationally required that we each limit our actions at that point at which all can pursue their sovereign interests to the maximum compatible with the same for everyone. We recognize the truth of the moral imperative of respect’ (1990: 141-142, emphasis in original). Despite Benn’s reference to ‘projects,’ the argument does not presuppose a robust notion of autonomy or a plan of life; as Benn makes clear, the presupposition is that agents possess what he call ‘autarchy,’ the capacity for genuine choice, a condition that is consistent with heteronomy (1988: chs. 8, 9). Thus Benn is advancing a liberal theory of the right that does not presuppose a liberal conception of the good or valuable life: even pebble splitters have a claim to non-interference.

A similar sort of argument was advanced by Alan Gewirth in his important and recently neglected book Reason and Morality (1981). Gewirth’s argument, like Benn’s, starts from a broad conception of agency and holds that, given this conception, individuals are committed to claiming for themselves, and honouring the claims of others to, basic rights to freedom and well-being. Benn and Gewirth thus share the common project of deriving basic liberal rights from the very idea of moral agency. Gewirth’s aim, though, is more radical: he argues that the nature of rational agency impels one to make a certain prudential claim, which gives rise to moral claims on others, which reason requires that one generalize. Moral agency is entailed by rationality. (For a general criticism of this type of argument, see Williams, 1985: 55-64.) In contrast, Benn distinguishes rational agency from moral agency. It is possible, he argues, to be a purely ‘natural person’ who makes no moral claims: psychopaths, he suggests, may possess rational natural personality—rational agents devoted to securing their goals—but be devoid of moral personality (1988: 101-2). Moral persons are those who see others and themselves in terms of moral relations: it is how we do in fact see ourselves, and it is moral persons who would recognize the basic right of non-interference.

Kantian Liberalism II: From Respect to the Moral Contract and Then to Liberal Rights

Despite their differences, Benn and Gewirth both seek a direct route from agency to liberal rights: if we understand the type of agents we are, we see that we must claim certain liberal rights and grant them to others. In contrast, what is often called ‘Kantian liberalism’ seeks to establish liberal rights via a hypothetical contract, which then generates basic rights. In the words of Sandel, its most famous critic, according to ‘deontological’ or ‘Kantian liberalism,’ ‘society, being composed of a plurality of persons, each with his own aims, interests, and conceptions of the good, is best arranged when it is governed by principles that do not themselves presuppose any particular conception of the good’ (1982: 1-7). Because, on this view, each is a chooser of her own ends in life, respect for the person of others demands that we refrain from imposing our view of the good life on her. Only principles that can be justified to all respect the personhood of each. Respect, then, requires a certain mode of justification, according to which moral principles are acceptable to all free moral persons in a fair choice situation. Liberal principles are then generated via this mode of justification. ‘According to contractualism,’ says Thomas Scanlon, ‘when we address our minds to a question of right and wrong, what we are trying to decide is, first and foremost, whether certain principles are ones that no one, if suitably motivated, could reasonably reject’ (1998: 189; see also Barry, 1995).

Notice that Scanlon’s test concerns ‘reasonable’ rejection, not what could be rationally rejected (1998: 191). For Scanlon, ‘[t]he distinction between what it would be reasonable to do and what it would be rational to do is not a technical one, but a familiar one in ordinary language’ (1998: 192). A reasonable person does not make claims that others cannot be expected to live with, or are grossly unfair. Rawls has a similar idea: parties to his original position are ‘rational and reasonable,’ not simply rational: ‘Persons are reasonable… when they are ready to propose principles and standards as fair terms of co-operation and to abide by them willingly, given the assurance that others will likewise do so’ (1996: 48). In contrast to Hobbesian contractors, Rawlsian contractors seek to respect each other’s status as free and equal moral beings (Larmore, 1996: ch. 6).

Kantian contractualism must build into the account some constraint that limits consideration to only justifications that all reasonable people would accept, or that none would reject. One way to do this is, à laRawls, to constrain the choice situation in such a way that the rational parties are forced to advance only reasonable considerations. The nature of Rawls’s argument behind the veil of ignorance (which excludes specific knowledge about a contractor’s post-contract life and personality) is such that given the constraints on choice, the most rational choice for a contractor will model a reasonable choice for you and me. Instead, though, of building into the framework of the choice situation our understanding of the demands of reasonableness, we might, as Scanlon suggests, appeal directly to our intuitions about reasonableness in the contactarian analysis (1998: ch. 5). Alternatively, one could seek to minimize the appeal to intuitions about reasonableness by developing a more systematic theory of justification and reasonable rejection (see Gaus, 1996).

A fruitful project for Kantian liberalism is to integrate the more direct version described above with the contractual argument (Reiman, 1990; Gaus, 1990: Part II). Some basic moral principles may be directly derived from our conceptions of ourselves as moral persons (i.e. a basic right to noninterference), while other moral principles (say, concerning specific schemes of property rights and distributive justice) may be justified via a contractual argument. The rights justified by the direct argument from moral agency would structure and constrain the contractual stage of the justification. In so far as Kantian contractualism supposes that some norms (i.e. concerning what principles are reasonable) are prior to the contract, admitting these additional pre-contractual norms is entirely consistent with the approach.

Kantian contractualism leads to the analysis of ‘public reason’ (see D’sAgostino, 1996), the key element of so-called ‘political liberalism.’ The line between Kantian contractualism and political liberalism is as thin and uncertain as the line between A Theory of Justice and Political Liberalism.


It should be clear that the label ‘comprehensive’ liberalism is misleading: it includes everything from truly comprehensive liberalisms as wide-ranging secular philosophies to Kantian liberal theories of political justice that seem consistent with a wide range of notions about value, social knowledge, and selfhood. To be sure, just how wide a range of conceptions of the self are consistent, say, with Kantian liberalism has been one of the hotly disputed issues over the last 20 years; thus just how comprehensive such liberalism must be has been the crux of the debate. For the most part, it seems mistaken to treat Millian and Kantian liberalism as equally comprehensive. Whereas Kantian liberalism makes some claims about the nature of agency and interpersonal rights, Millian liberalism qua a theory of individual development seems committed to a rich theory of the good life, while qua a utilitarian theory is committed to an overall theory of the good and the right. On the other hand, understood simply as an account of liberalism resting on the harm principle, Millian liberalism may not be any more comprehensive than Kantian liberalism; indeed it may be less comprehensive, adopting a moral principle that can be subject to what Rawls calls overlapping consensus (1996: Lecture 4). All told, it would probably be better to abandon the contrast between ‘comprehensive’ and ‘political’ liberalisms in favour of more fine-grained distinctions.