ANYONE TRYING TO GIVE a brief account of liberalism is immediately faced with an embarrassing question: are we dealing with liberalism or with liberalisms? It is easy to list famous liberals; it is harder to say what they have in common. John Locke, Adam Smith, Montesquieu, Thomas Jefferson, John Stuart Mill, Lord Acton, T. H. Green, John Dewey, and contemporaries such as Isaiah Berlin and John Rawls are certainly liberals—but they do not agree about the boundaries of toleration, the legitimacy of the welfare state, and the virtues of democracy, to take three rather central political issues. They do not even agree on the nature of the liberty they think liberals ought to seek (Berlin 1969, 122–34).

It is a familiar complaint in writing about politics generally that key terms are undefined or indefinable; the boundaries between “political” and “nonpolitical” behavior and institutions are disputed, the defining characteristics of statehood, along with the necessary and sufficient conditions of legitimacy, are incessantly debated. Liberalism may be no worse off than its ideological competitors, of course. In everyday political practice, all the isms seem to be in the same condition; liberals, conservatives, and socialists can be identified only issue by issue, and their stand on one issue offers little clue to their stand on another. The conservative who opposes railway nationalization supports government subsidies of defense contractors, while the liberal who applauds the establishment of an ethics committee to investigate the financial dealings of politicians will deplore the establishment of a committee to investigate the ethics of schoolteachers.

But even if conservatism and socialism are in the same plight, one is still inclined to ask, is liberalism one thing or many? Is liberalism determinately describable at all (Dworkin 1985, 183–203)? The observation that the terms of political discourse are not easily brought to an agreed definition is not new. More than three hundred years ago, Thomas Hobbes remarked that if anyone had stood to profit from a similar confusion in geometry, mankind would still be waiting for Euclid. While Hobbes’s remark suggests that it is the self-interest of priests, intellectuals, and politicians that explains this lack of precise definitions, twentieth-century writers have suggested another (p.22) reason, that political concepts are “essentially contested” (Gallie 1956, 167–98; Gray 1983, 75–101). A third explanation, and one more relevant to liberalism in particular, is that liberals’ political concerns have altered over the past three centuries. All three kinds of explanation suggest, however, that we should be seeking to understand liberalisms rather than liberalism.

One reason for the indefinability of political terms, or the systematic slipperiness of our concepts of the state, the political, or, as here, liberalism, is the use of these terms as terms of praise or obloquy in the political struggle; this is a modern version of Hobbes’s view that disputed definitions are the result of competing interests. Since the 1970s, for instance, there has been an intellectual and political movement known as “communitarianism” whose main defining feature is hostility to liberalism (Sandel 1982). Communitarians emphasize the innumerable ways in which individuals are indebted to the societies in which they are reared; liberals, they say, write as if human beings come into the world with no social ties, owing no allegiances, and one way or another entirely detached from the societies they in some fashion inhabit. So described, liberalism is unattractive, built on sociological falsehoods and moral autism. Self-described liberals have naturally said that this is a parody of their views (Rawls 1985, 233; Rorty 1991, 179ff.).

Liberals themselves have sometimes tried to define liberalism in such a way that only the very deluded or the very wicked could fail to be liberals. At the height of the Cold War, it was easy to present the alternatives as liberal democracy on the one hand and assorted forms of one-party totalitarianism on the other. This attempt to narrow the range of political options was itself resisted. Social democrats, who opposed both the one-party state and uncontrolled capitalism, believed that their disbelief in the legitimacy of private property in the means of production distinguished them from liberal democrats. American conservatives distinguished themselves from liberals by according state and central governments a greater role in preserving national identity and some form of traditional moral consensus than liberals accept, or else by advocating a more laissez-faire economy and a reduced role for government (Rossiter 1982, 235ff.). Their critics retorted that they were nonetheless doomed by American history to remain liberals (Hartz 1955, 145–49).

The attempt to produce a clear-cut definition of a political stance is not always part of a hostile campaign to present the doctrines in question as incoherent or malign. Many political movements have devoted much effort to establishing a creed to which members must swear allegiance. Lenin spent as much time denouncing his Marxist allies for their misunderstanding of scientific socialism as attacking the czarist regime. He thought a revolutionary movement must know exactly what it thought and hoped to achieve. If the fainthearted or intellectually unorganized were driven out, so be it; as the title of one essay proclaimed: “Better Fewer but Better.” Of all political creeds, liberalism is the least likely to behave like this. Whatever liberalism involves, it certainly includes toleration and an antipathy to closing ranks (p.23) around any system of beliefs. All the same, liberals have often asked themselves what they have in common, where the boundaries lie between themselves and, say, socialists on the one side and conservatives on the other.

Another explanation of the difficulty of defining political terms is that they are essentially contested terms, terms whose meaning and reference are perennially open to debate. If we define liberalism as the belief that the freedom of the individual is the highest political value, and that institutions and practices are to be judged by their success in promoting it—perhaps the most plausible brief definition—this only invites further argument. What is liberty? Is it positive or negative? How does the liberty of a whole nation relate to the liberty of its members? Nor is liberty the only concept to invite such scrutiny. Who are the individuals in question? Do they include children? Do they include the senile and the mentally ill? Do they include resident aliens or the inhabitants of colonial dependencies? This might be thought to be unsurprising; any definition opens up discussion of the terms in which the definition is proffered. The sting in the claim that these are essentially contested concepts is the thought that any elaboration will provoke further argument (Gallie 1956, 175ff.).

There is a clear direction in which any elaboration of the definition of a chair, say, must go, and a clear line beyond which discussion is merely captious. This seems not to be true of the discussion of political doctrines. Whether the view that there are essentially contested concepts is entirely coherent is another question. Unless some substantial portion of the meaning of a concept is uncontested, it is hard to see how the concept could be identified in the first place. There must be a central uncontested core of meaning to terms like “liberty” if arguments about the contested penumbra are to make sense. A man in jail is paradigmatically not free; a man threatened with punishment if he writes a book is paradigmatically less free to write it than a man not so threatened (Berlin 1969, 122ff.). Even so, we may agree that political terms are constantly being endowed with new meanings, in much the way the terms of the law are endowed with new meanings in the course of legal argument. Even if liberalism is distinct enough to be identifiable, it still changes over time.

Varieties of Liberalism: Classical versus Modern

To agree that liberalism may have a variety of institutional manifestations while resting on one moral basis—Locke’s claim that men are born “in a state of perfect freedom, to order their actions and dispose of their possessions, and persons, as they see fit. . . a state also of equality” (Locke [1690] 1967, 287), for instance—does not mean that all doubts about the porosity of liberalism have been laid to rest. One argument that has taken on the status of a commonplace is that there have been two kinds of liberalism: one “classical,” limited in its aims, cautious about its metaphysical basis, and political in its orientation; the other “modern,” unlimited, incautious, global (p.24) in its aims, and a threat to the achievements of “classical liberalism.” Classical liberalism is associated with John Locke ([1690] 1967), Adam Smith ([1775] 1976), Alexis de Tocqueville ([1835] 1964), and Friedrich von Hayek (1973–79). It focuses on the idea of limited government, the maintenance of the rule of law, the avoidance of arbitrary and discretionary power, the sanctity of private property and freely made contracts, and the responsibility of individuals for their own fates.

It is not necessarily a democratic doctrine, for there is nothing in the bare idea of majority rule to show that majorities will always respect the rights of property or maintain the rule of law (Madison, Hamilton, and Jay 1987, Federalist 10, 122–25); it is not always a progressive doctrine, for many classical liberals are skeptical about the average human being’s ability to make useful advances in morality and culture, for instance. It is hostile to the welfare state; welfare states violate the principle that each individual ought to look to their own welfare, and frequently couch their claims in terms of the achievement of social justice, an ideal to which classical liberals attach little meaning (Hayek 1976). More importantly, perhaps, welfare states confer large discretionary powers upon their politicians and bureaucrats, and thus reduce to dependency their clients and those who depend upon the state for their prosperity.

Modern adherents of classical liberalism often ground their defense of minimal government on what they take to be a minimal moral basis. Minimal government may, for instance, be justified by the prosperity that economies deliver when they are not interfered with by governments; this argument has been current from Adam Smith’s Wealth of Nations defense of “the simple system of natural liberty” (Smith [1775] 1976, 687) down to Hayek’s in our own time. It is not morally contentious to claim that prosperity is better than misery, and it has been given greater credibility than ever by the collapse of the communist regimes of Eastern Europe and the discrediting of military and authoritarian governments elsewhere.

An equally minimalist defense of liberalism as minimal government is provided by pointing to the nastiness of governmental coercion and to the contrast between the negative effects of mere brute force and prohibition compared with the benign effects of uncoerced cooperation. No classical liberal denies the need for law; coercive law represses force and fraud, and the noncoercive civil law allows people to make contracts and engage in any kind of economic activity. Still, every classical liberal holds that all the forces that make for imagination, invention, and growth come from the voluntary sector of the social order.

Classical liberals are not unanimous about the relationship between minimal government and the cultural and moral order, and this is perhaps the most important point about their moral views. Unlike modern liberals, they do not display any particular attachment to the ideal of moral and cultural progress. David Hume was more a political conservative than Adam Smith, (p.25) but was more inclined than Smith to admire the “brisk march of the spirits” typical of a flourishing commercial society. Tocqueville was doubtful whether liberty could survive in the absence of strong religious sentiment, think- ing that the self-reliance and self-restraint that he admired was not natural to modern man (Tocqueville 1964, 310–25), and Hayek was inclined to think that political liberalism rests upon cultural conservatism (Gray 1984, 129–31).

Contemporary defenders of classical liberalism think it threatened by modern liberalism. Modern liberalism, in this view, reverses the ambitions and restraints of classical liberalism, and in the process threatens the gains that classical liberals achieved when they replaced the tyranny of kings and courtiers with constitutional regimes. Modern liberalism is exemplified by John Stuart Mill’s On Liberty, with its appeal to “man as a progressive being” and its romantic appeal to an individuality that should be allowed to develop itself in all its “manifold diversity” (Mill [1861] 1974, 120–22). Philosophically, it is exemplified equally by the liberalism of the English Idealists and “new liberals” such as L. T. Hobhouse ([1911] 1964).

In practice, it is exemplified by the assault on freedom of contract and on the sanctity of property rights represented by the welfare legislation of the Liberal government in the UK before World War I, by Franklin Roosevelt’s New Deal between the wars, and by the explosion of welfare-state activity after World War II. Modern liberalism is usually (but not always) agreed, even by its critics, to be a form of liberalism, for its underlying moral basis is couched in terms of freedom. Negatively, the aim is to emancipate individuals from the fear of hunger, unemployment, ill health, and a miserable old age, and, positively, to attempt to help members of modern industrial societies flourish in the way Mill and Wilhelm von Humboldt wanted them to.

It is liberal, too, because it does not share the antipathies and hopes of a socialist defense of the modern welfare state. Although some defenders of the rights of property claim that almost any restriction on the absolute liberty of owners to dispose of their own as they choose amounts to confiscation (Epstein 1985), modern liberalism has no confiscatory ambitions. Inasmuch as the ideals of the welfare state cannot be achieved without a good deal of governmental control of the economy, modern liberalism cannot treat property as sacrosanct and cannot limit government to the repression of force and fraud; but distinguished modern liberals such as John Rawls argue that personal property is a necessary element of individual self-expression, especially by means of freedom of choice in careers, even if vast shareholdings are not (Rawls 1971, 272–74). Critics of modern liberalism usually insist that it is liberalism, but a dangerous variety.

The fear that modern liberalism is inimical to the spirit of classical liberalism and will, in practice, threaten the latter’s gains rests on two things. The first is the thought that modern liberalism is ideologically or metaphysically overcommitted. Mill’s vision of man as a progressive being, with its demand (p.26) that everyone should constantly rethink her opinions on every conceivable subject, is one with at best a minority appeal. To found one’s politics on a view of human nature that most people find implausible is to found one’s politics on quicksand. There is no need to appeal to such a vision of human nature to support classical liberalism; conversely, it is not clear that the kind of independent and imaginative personalities by which Mill set such store are best produced in a liberal society. History suggests that many of them have flourished by resisting an illiberal and conservative environment (Berlin 1969, 172).

The second is the thought that modern liberalism makes everyone an unrealizable promise of a degree of personal fulfillment that the welfare state cannot deliver, and that its efforts to deliver it will inevitably lead to frustration. For one thing, people resent being forced to part with their hard-earned income to provide the resources that supply jobs, education, and the various social services that modern liberalism employs to create its conception of individual freedom for other people. This creates a hostility between more and less favored groups of citizens that is wholly at odds with what modern liberals desire.

Moreover, the welfare state must employ an extensive bureaucracy whose members are granted discretionary powers and charged by law to use those powers for the welfare of their clients. This means that classical liberals’ concern for the rule of law and the curtailing of arbitrary discretion is ignored: bureaucrats are given resources to disburse to their clients, and meanwhile the allegiance of the citizenry is undermined when the state fails to produce the good things it has been asked to provide. The liberation the welfare state promises—liberation from anxiety, poverty, and the cramped circumstances of working-class existence—is easily obtained by the educated middle class and is impossible to achieve for most others. There is thus a grave risk of disillusionment with liberalism in general as a result of its failure when it overextends itself. Some writers suppose that the worldwide popularity of conservative governments during the 1980s is explained by this consideration.

Varieties of Liberalism: Libertarianism and Liberalism

There is a closely related but not identical divide within liberal theory between liberalism and libertarianism. Just as in the conflict between classical and modern forms of liberalism, there is a tendency for the partisans of one side or the other to claim that their version of liberalism is true liberalism and the alternative something else entirely. Contemporary libertarians often claim that they are classical liberals. This is not wholly true. There is at least one strand of libertarian thought, represented by Robert Nozick’s Anarchy, State, and Utopia, that advocates the decriminalization of “victimless crimes” such as prostitution, drug taking, and unorthodox sexual activities (p.27) (Nozick 1974, 58–59). There is nothing of that in John Locke or Adam Smith.

The line between liberal and libertarian theories is not easy to draw. Both are committed to the promotion of individual liberty; both rest most happily on a theory of human rights according to which individuals enter the world with a right to the free disposal of themselves and their resources. The line of cleavage lies between the libertarian view that government is not a necessary evil but a largely (and for so-called anarchocapitalists, a wholly) unnecessary evil, and the liberal view that governmental power is to be treated with caution but, like any other instrument, may be used to achieve good ends. Perhaps the most important point of difference is that libertarians see our rights as a form of private property, what Nozick has called “entitlements” (Nozick 1974, 150ff.). The individual is the owner of his or her person and abilities; so viewed, our rights have two sources only—our initial ownership of our own selves and capacities, and the claims on whatever resources and abilities other people have freely agreed to transfer to us. The state, if legitimate at all, may do no more than secure these rights. It has no resources of its own and cannot engage either in the redistributive activities of modern welfare states or in the quasi-charitable activities of such states. Nobody has the right to deprive anyone else of their property by force—if they have committed no crime—and neither does the state.

This is in sharp contrast to the most famous recent account of welfare state liberalism, John Rawls’s Theory of Justice. In Rawls’s account, we arrive at an understanding of what rights we possess, and of how far our liberty extends, by asking ourselves a hypothetical question—“What rights would we all demand for ourselves and acknowledge in others if we were to establish a social and political system de novo, knowing nothing about our particular abilities and tastes, and therefore being forced to strike a fair bargain with everyone else?” (Rawls 1971, 11–17). Rawls’s claim is that we should acknowledge two rights: the right to the most extensive liberty consistent with the same liberty for everyone, and a right to just treatment, enshrined in the thought that inequalities are justified only to the extent that they improve the situation of the least advantaged (60–61).

This second principle is often called the maximin theory of justice, since it explains social justice as maximizing the size of the minimum holding of social resources. This principle is clearly inimical to any account of the state that restricts it to the defense of property rights. The introduction of a conception of social justice into the defense of a liberal political theory rests on the idea that individuals have a right to self-development, and therefore on the kind of theory of individual development that underpins Mill’s On Liberty and alienates defenders of classical liberalism.

All dualisms ride roughshod over a complicated world. There are forms of liberalism that are nonlibertarian, but also more nearly “classical” than modern liberalism. Locke’s Two Treatises is, on the face of it, more friendly (p.28) to private property than the views of Rawls or Mill, and yet Locke shows none of the hostility to the state that libertarians do. The state is obliged to act according to the rule “salus populi suprema lex”—“the good of the people is the highest law”—and there is no suggestion that this is only a matter of repressing force and fraud (Locke [1690] 1967, 391). On the other hand, there is also no suggestion that the least advantaged members of society have a right to do as well as possible. Locke suggests that they have to do well enough to make membership in civil society a good bargain—otherwise, they might as well immigrate to some unoccupied part of the world and start again—but he does not suggest that they have any claim beyond that (314–15). Certainly, Locke’s individualism treats each person as responsible for his or her own welfare, but Locke’s concern with our moral welfare rather than economic well-being means that he was more concerned with religious toleration than with “health and human services.”

Because we are tempted to acknowledge that we are faced with liberalisms rather than liberalism, and are inclined also to say that they are all versions of one liberalism, it is tempting to suggest that liberalism is best understood in terms of what it rejects. Nor would it be surprising to come to such a conclusion. Conservatism is no easier to define than liberalism, and it is not infrequently observed that what conservatives believe is a matter of what they want to conserve and who threatens it. Indeed, Louis Hartz’s The Liberal Tradition in America argued that conservatives in the United States, as opposed to their counterparts in Britain and Europe, were in a bad way because the society and political system they want to conserve has always been a liberal one; temperamental conservatives are thus forced to be ideological liberals (Hartz 1955, 145–54). However that may be, it is not implausible to argue that liberalism is well defined in negative terms. Its central commitment, liberty, is in general a negative notion—to be free is to be not in jail, not bound to a particular occupation, not excluded from the franchise, and so on—and the history of liberalism is a history of opposition to assorted tyrannies.

Antiabsolutism

One way of understanding the continuity of liberal history in this light is to see liberalism as a perennial protest against all forms of absolute authority. It is notoriously difficult to suggest a starting date for liberal political theory, or, rather, it is notoriously easy to suggest all sorts of starting dates, running from the pre-Socratics onward, but notoriously difficult to find any kind of consensus on one of them. In British politics, for instance, it was only in the 1860s that the more radical members of the Whigs called themselves the (p.29) Liberal Party. Yet it would be odd not to count Locke among early liberals, just as it would be absurd to call Hobbes a liberal even while one might want to acknowledge that he supplied many of the ingredients for a liberal theory of politics in the course of defending absolute and arbitrary authority as the only alternative to the anarchy of the state of nature and the war of all against all.

Whatever liberalism has been concerned with, it has been concerned with avoiding absolute and arbitrary power. It is not alone in this. English constitutional theory had for several centuries an aversion to anything that smacked of confiding absolute power to anyone whatever. Neither Parliament, nor the judiciary, nor the king was entitled to a monopoly of political authority. The imagery of the body politic was called upon to suggest that the elements in the political system had to cooperate with one another for the body to function coherently. What makes liberal hostility to absolute rule liberal rather than merely constitutionalist is the liberal claim that absolute rule violates the personality or the rights of those over whom it is exercised (Locke [1690] 1967, 342–48).

This argument connects Locke’s Second Treatise, with its claim that absolute and arbitrary authority was so inconsistent with civil society that it could not be considered a form of government at all, with the twentieth-century liberal’s contempt for the totalitarian regimes of Nazi Germany and Stalinist Russia. Liberals have disagreed about just which sorts of absolute authority are intolerable. Locke agreed that a general needed absolute authority over his soldiers in battle, and might shoot deserters out of hand. But this was not arbitrary authority—generals might shoot deserters, but not take sixpence from their pockets (Locke [1690] 1967, 379–80).

Mill thought the principles of On Liberty did not apply to people who could not benefit from rational discussion (Mill [1861] 1974, 69–70). Elizabeth I and Peter the Great had rightly exercised unaccountable power over sixteenth-century Britain and eighteenth-century Russia, respectively, and the despotic power of the East India Company over its Indian subjects was legitimate. The nineteenth-century British working class, on the other hand, was entitled to full civil and political rights, and women of all classes as much as men. Other liberals have been rather less ready to describe entire populations as “childish,” and have thought absolute authority over colonial possessions as indefensible as any other absolute power.

The thought behind liberal opposition to absolute power is not complex, although it has several strands. One is the idea that political authority exists for purely secular ends, toward which we should adopt a rational, scientific attitude, adjusting our political institutions and our policies in an instrumentally efficient way. Negatively, this means that liberals do not see authority as conferred either by the voice of God, as in theories of divine right or charismatic authority, or by the dictates of history, as in Marxist theory, or by racial destiny, as in Nazi theory. Authority exists only to enable a society to achieve those limited goals that a political order enables us to (p.30) achieve—the security of life, property, and the pursuit of happiness (Locke [1689] 1956, 128–29).

It follows that nobody can claim absolute power, since their title to exercise power rests on their ability to pursue these limited goals efficiently. A second idea that reinforces the first is that the content of these limited goals can be set only by attending to the opinions of all the people under that authority, or at least all those who have not shown themselves to be antisocial or a menace to the political order. To exclude anyone’s views is to devalue them; it is also to deny what liberalism relies on for its effect as a moral argument, the claim that we are born free and equal (Dworkin 1985, 191ff.). As free, we must be persuaded to give our allegiance, and as equal, we must be obliged on the same terms as everyone else. This means that government must listen to the people and cannot therefore take to itself any kind of absolute power (Rawls 1971, 221–23).

A third element provides much of the antitotalitarian energy of modern liberalism. Free and equal individuals must be so recognized in the legal system as well as in the political system narrowly conceived. They must be free to form associations for their own purposes and to engage in varied social, commercial, and intellectual activities. Absolute authority is inimical to, and unwilling to share, control over the lives of the citizenry with the leaders of other, secondary groups. The history of twentieth-century totalitarian states indeed shows that such states have always destroyed the independent authority of all other associations they could lay hands on. Liberals believe that the energy and liveliness of a society comes from these secondary allegiances, and therefore that absolute power is both an affront to the moral personality of individuals and destructive of the life of society at large (Dworkin 1985, 193–200).

Antitheocracy

The opposition to absolutism, which links Locke to Mill and both of them to Rawls, Dworkin, and contemporary liberal thinkers, had its origins in another issue. This was the liberal hostility to the confusion of secular and religious authority, and the liberal obsession with the rights of conscience. It has often been pointed out that the first usage of the term “liberal” in a political context was in the context of European anticlerical politics in the nineteenth century. For many Roman Catholics, the term “liberal” was, except when used to qualify “education,” a term of abuse. Voltaire was not a wholehearted liberal, but the cry of “écrasez l’infame,” with which he attacked the repressive and brutal power of the Catholic Church in eighteenth-century France, became a rallying cry of anticlerical liberals all over Europe.

Liberalism was associated with the nineteenth-century movement of European ideas that was concerned to drive a wedge between church and state and to make the Catholic Church no more influential in the politics of Catholic (p.31) countries than the various Protestant churches were in the countries where they flourished. In essence, the argument was an argument in favor of religious toleration and against any kind of religious monopoly.

It is sometimes thought that toleration arises when people are convinced that there is no way of knowing what the truth is in matters of religion, and that toleration is the fruit of skepticism. But this is quite wrong. Hobbes was a skeptic, but he was also deeply hostile to supposed rights to toleration. It is this that marks him as a nonliberal. The advocacy or denial of toleration as a matter of right divides the liberal and the nonliberal more sharply than anything else. For Hobbes, religious doctrines were too important to be left to private men to pick and choose; even if those doctrines were intellectually quite absurd, they stirred up the passions and so threatened the peace. It was thus the task of the sovereign to regulate what might and might not be said in public on all such matters; if the sovereign failed in this duty, the peace would be broken, exactly the outcome that the sovereign existed to prevent (Hobbes [1651] 1991, 124–25).

Locke put forward the modern doctrine of toleration some thirty years after Hobbes. In Locke’s eyes, there were two distinct realms, the sacred and the secular. Locke thought the first much more important than the second, but he also thought that secular authority was quite impotent to achieve anything useful in that realm. The political realm dealt with what Locke termed bona civilia, the goods of earthly peace and security, which he otherwise characterized as life, liberty, property, and physical well-being (Locke [1689] 1956, 128). A sovereign who tried to dictate how we practiced our religion was overstepping the proper bounds of his authority. Conversely, a church that tried to dictate the secular law was overstepping the bounds of its authority. The state was essentially a nonvoluntary organization, and one to which we owed obedience willy-nilly; churches were essentially voluntary, and probably plural.

Locke was, as Hobbes was not, a devout Christian, one who thought a great deal about religion as religion rather than from a sociological perspective. It was this that made Locke a passionate defender of toleration. One of the arguments in favor of toleration and against the mingling of church and state was precisely that human beings—especially late seventeenth-century human beings of a Protestant persuasion—were extremely tender about matters of conscience. To force someone to assert a belief he did not really hold was to outrage his deepest nature.

Where Hobbes had suggested that men quarrelled over matters of conscience because there was next to nothing to be known about religion by the light of reason alone, and therefore men ought to be made to assert something in common, simply for the sake of peace, Locke was committed to the view that God required a willing assent and a real faith, so that whatever kind of forced assent the state might induce us to make was an insult to God as well as an outrage upon the individual (Hobbes [1651] 1991, 260ff.; Locke [1689] 1956, 132–33).

(p.32) Conversely, true religion can make no demands upon the state. This is a view that modern readers find harder to accept. Locke thought it impossible that there might be a valid religious reason for a group to do anything that might come into conflict with the ordinary criminal law. Thus he would have differed with most liberals of today over the case in 1990 in which the U.S. Supreme Court found that the First Amendment guarantees of religious liberty did not entitle Native Americans to use the hallucinogenic drug peyote in their religious rituals once the State of Oregon had banned the consumption of peyote.

Locke would have sided with the Court, but many contemporary liberals thought the demands of any religion should weigh more heavily than that. Locke also confined toleration to opinions that did not threaten the political order; modern readers are often shocked to find that neither Roman Catholics nor atheists would be tolerated in any society that followed Locke’s prescriptions. In both cases, the argument was that they were politically dangerous; atheists lacked motives to keep their promises and behave decently, while Catholics professed earthly loyalty to the pope and so could not be relied on by the rulers of whatever state they happened to belong to (Locke [1689] 1956, 157–58).

This reflected Locke’s sharp distinction between those matters over which secular authority might be exercised and those over which it must not. Locke argued that earthly governments existed for certain simple tasks and no others, an argument heavily dependent on the idea that it is obvious what the function of earthly government is, and that it does not include saving men’s souls. Mill’s On Liberty took a different route to much the same conclusion, not by arguing that it was obvious what the function of government was, but by showing that a consistent utilitarian who believed in the importance of individuality and moral progress must agree that coercion, especially the organized coercion exerted by governments, was legitimate only to defend certain this-worldly interests—our own liberty and security above all else (Mill [1861] 1974, 119ff.; Gray 1983).

Mill’s argument is no more conclusive than Locke’s. Enthusiasts for the mixing of church and state may set no value on individuality for its own sake and believe that an enlarged freedom would lead to depravity rather than moral progress. It is on this basis that they demand the union of spiritual and secular authority. In the second half of the twentieth century, liberals have generally taken a less rhetorical, more practical line than Locke and Mill. Totalitarian regimes, the lineal descendants of confessional states, have two great drawbacks. The first is that they employ a distasteful amount of force in securing their goals. Because it is so difficult to tell whether one’s subjects are really saved or really loyal to the Nazi Party, or whatever, the temptation is to pile on the penalties for dissent and to engage in acts of exemplary brutal punishment, which does little to secure a real loyalty to the regime and much to make its rulers insecure when they contemplate the hatred of the population they have intimidated (Arendt 1968).

(p.33) The second is that such regimes are inefficient; they may be effective when fighting a real, all-out war, but they are economically less efficient than liberal societies in which the division of labor between the sacred and the profane is respected in approximately the form Locke laid down. Whether this practical argument captures liberals’ deepest beliefs is doubtful. It is hard not to suspect that liberals feel more passionately than that about the wickedness of totalitarian regimes and, for that matter, about the wickedness of authoritarian clerical regimes of the kind typified by the Spain of General Franco. When they feel passionately about such regimes, it is in much the same way as Locke, for modern notions of the violation of personality reflect, in a secularized fashion, Locke’s view that the imposition of belief on any individual was an affront to that individual and God their creator (Rawls 1971, 205–11).

Anticapitalism

The history of hatred for despotism, theocracy, and the modern union of the two that is reflected in totalitarianism is a long history. The third of liberalism’s antipathies has a shorter history. From the middle of the nineteenth century until today, one strand of liberalism has regarded capitalism as an enemy of liberty (Mill [1848] 1965, 766–69; Dewey [1931] 1984). This marked a great reversal in the history of liberalism. It is not a large oversimplification to say that until the early nineteenth century there was no question of opposing liberalism to capitalism. The movement of ideas and institutions that emancipated individuals from tradition, insisted on their natural rights, and demanded that “careers should be open to talent” rather than birth was a seamless whole.

Just as a man had to think for himself, so he had to work for himself; just as society would progress only if each person took responsibility for their own ideas and moral convictions, so it would flourish economically only if everyone stood on their own two feet. How far this was an articulate defense of capitalism as such is debatable; the term “capitalism” itself did not come into general use until the late nineteenth century, and it is difficult to decide how appropriate it is to characterize as capitalist those societies that possessed nothing one could call a proletariat, whose populations still lived largely in the countryside and worked the land, and that thought of themselves as “commercial societies” rather than “capitalist economies” (Smith [1775] 1976, 399–403).

Moreover, many of the rights to dispose of property just as one wished, to work for anyone willing to employ one, and to contract with anyone for any purpose not obviously damaging to the security and good morals of the commonwealth had been established by successive decisions made by judges appealing to the English common law rather than by legislation of a self-consciously liberal kind. Still, there is an obvious affinity between liberalism on the one hand and the rule of private property and freedom of contract on (p.34) the other. The liberal view that the individual is, by natural right or by something tantamount to it, sovereign over himself, his talents, and his property is at once the basis of limited government, the rule of law, individual liberty, and a capitalist economy.

But it was apparent from the beginning that property might be employed oppressively as well as harmlessly or beneficially. Apart from the conflict between the rights of property owners and the traditional claims of rural workers—such as customary claims to gather wood or to glean in the fields or to take small game—there was a more general conflict between the liberty of the large property owner to do what he chose with his property and the impossibility of his workers or competitors striking anything like a fair bargain with him. Throughout the nineteenth century, the sentiment grew that if it had once been necessary to liberate the entrepreneur from misguided or oppressive government, it was now necessary to liberate the worker and consumer from the tyranny of the capitalist (Hobhouse [1911] 1964, 22–24, 82–84; Green 1892, 366–70).

Mill observed that the modern wage laborer had as little real choice of occupation as a slave had in antiquity. In that spirit, he defended the right of working people to organize into trade unions to redress the balance of power a little. T. H. Green and L. T. Hobhouse went further, suggesting that capitalism exerted a kind of moral tyranny over the ordinary person, as exemplified by the spread of drinking establishments that destroyed both the health and the self-respect of their victims (Green 1892, 380–85). The “New Liberalism,” exemplified in Britain by the social policy of the Asquith government of 1908–16, and in the United States by the demands of the Progressives and the practice of the Democratic Party after the election of Franklin D. Roosevelt in 1932, had many positive ambitions, but one negative assumption was that the workingman needed to be freed from the power of the capitalist. It is this that explains the seeming paradox that late twentieth-century conservatives are often characterized as “neoliberals.” The contemporary defense of property rights is not, as it was two centuries ago, the defense of landed property against commercial and industrial capital, but the defense of nineteenth-century laissez-faire and the property rights of commercial and industrial capital against modern reformers.

The tidiness of a definition of liberalism couched in terms of its oppositions is only apparent. Certainly, liberalism is antidespotic, anticlerical, and hostile to twentieth-century manifestations of those evils, including the perverted manifestations of totalitarianism. But just as there is a tension between classical and modern liberalisms, the same tension reappears between pro- and anticapitalist liberalisms. And just as most liberals would not wish to pursue the goals of the welfare state to the lengths of threatening the survival of (p.35) limited, lawful government, so they would not wish to restrain the operations of a capitalist economy to the point at which it turned into a command economy. Whether we start from liberal enthusiasms or liberal antipathies, we find the same controversies.

The wish to find a position that is intellectually attractive and politically responsible exposes liberals to accusations of not knowing their own minds or of being wishy-washy. Liberals have retorted that it is not their fault that the world is a complicated place that requires nuanced handling. One way of underpinning that reply is to provide the positive liberal theory that explains both why liberalism is hostile to the threats to freedom that it encounters and why these threats have varied over time.

A Theory for Individuals

In spite of the suggestion that liberalism should confine its attention to political institutions, liberalism is best understood as a theory of the good life for individuals that is linked to a theory of the social, economic, and political arrangements within which they may lead that life. John Rawls’s Theory of Justice provides some persuasive arguments for the view that we should build a liberal theory for institutional design without committing ourselves to any particular view of “the good life,” and those arguments’ eventual failure to convince tells us a lot about why a broader theory is needed.

Rawls argues that the search for a consensus in favor of liberal political and economic institutions will go more smoothly if we seek foundations that are neutral with respect to the great, but sharply contested, issues of religion and personal ethics (Rawls 1989, 233–38). Critics have noted, however, that Rawls’s minimalist assumptions about “the good life” remain decidedly liberal—he takes it for granted that slavery is an unspeakable evil, that the suppression of conscientious belief is so intolerable that no rational person could trade the chance of being in command of the Inquisition for the risk of being one of its victims, and that freedom of choice in career and lifestyle is essential for life to have any meaning.

The same critics have also pointed out that the principles of justice proposed by Rawls are not suited to absolutely anyone but especially to persons holding a late twentieth-century conception of themselves and the meaning of their lives. The thinness of the premises about human nature and the human good that Rawls builds on do not reflect skepticism or a lack of moral conviction so much as the eminently liberal thought that each person is in command of his or her own moral destiny and that it is not for others to dictate it, as Rawls has subsequently tended to agree (Rawls 1985).

At all events, liberalism viewed as a doctrine for individuals can be understood in terms one might borrow from Immanuel Kant, Humboldt, Mill, Bertrand Russell, or Dewey, since a variety of formulations seize on the same points. The essence is that individuals are self-creating, that no single good defines successful self-creation, and that taking responsibility for one’s own (p.36) life and making of it what one can is itself part of the good life as understood by liberals. Dewey labeled this experimentalism; Kant defined it as the spirit of the Enlightenment; Mill borrowed from von Humboldt to argue that the fundamental aim is to develop human nature in all its diversity (Mill [1861] 1974, 121–22; Dewey [1931] 1984, 114–20; Kant 1991, 53–54).

Its positive attractions become clearer when they are contrasted with preliberal or antiliberal views. Self-discipline is a great good because nobody can conduct “experiments in living” without the self-control that allows them to stand back and assess their success or failure; submission to discipline, as praised by many Christian writers, and before them by Plato, is not a good in itself (Plato 1941, 127–40). Attachment to one’s country and fellow citizens is a great good because few human virtues flourish except against a background of loyalty and strong fellow feeling; “my country right or wrong” is an illiberal sentiment, suggesting an immersion in patriotic sentiment inconsistent with the ideals of individual autonomy.

Plato condemned democratic Athens for its attachment to diversity and variety; liberals condemn Athens for being insufficiently hospitable to diversity and variety as good in themselves. Pericles’s famous funeral oration praises the Athenians for their willingness to allow others to live as they pleased, but suggests no positive enthusiasm for variety as a human good, denies that women have any place in public life, and ranks politics higher than any private good. Liberals generally praise public spirit, and most, at any rate, would agree that in time of crisis we are obliged to put aside our private concerns and do what we can for our country, but they would also see this as a sacrifice of one good for another, while Pericles was true to the classical ideal in ranking the goods of private life much lower than the goods of public life (Thucydides 1972, 143–51).

It is true that liberalism has no single positive picture of “the good life for man.” It is true because liberals have commonly been empiricists and inclined to believe that only experience can reveal what really conduces to individual flourishing, and also because liberals have often been pluralists and have thought that autonomous individuals might choose a great variety of very different but equally good lives (Berlin 1969, 172–74). It is not, as critics often maintain, that liberals elevate choice to the only absolute good; no liberal would applaud a life of crime merely because the criminal had chosen it. It is, however, true that most liberals have thought that the kind of autonomous individual they have admired can become a fully autonomous being only by exercising their powers of choice. Some people may get lucky and find what suits them without very much exploration of alternatives; others may need to search much longer. But a person incapable of making a choice and sticking to it will have little chance of leading a happy life.

This vision is not uncontroversial, and it is unattractive to many critics. It is unsympathetic to a vision of an orderly universe in which the best lay down the rules of life for the rest of us; it is antiaristocratic, at odds with a (p.37) belief in Platonic guardians, Aristotelian aristocrats, and the Catholic Christian tradition’s claim to know what we must do to be saved. Conversely, it is too strenuous for anyone who thinks most people do well enough by thoughtlessly following the habits and customs of their fellows. It is too optimistic for anyone who believes in the essential depravity of the human race. Liberals look for improvement, not merely to prevent our worse natures from getting out of hand. Writers like Joseph de Maistre and Georges Sorel have not unpersuasively ridiculed this outlook.

Looked at from the other side, it can be criticized as insufficiently serious about its own premises. Friedrich Nietzsche claimed that liberals did not take choice seriously, since they assumed that everyone would share their ideas about what constituted good choices and good reasons for choosing one path rather than another. His successors in the existentialist tradition made essentially the same point. As observed before, liberals are uncomfortably aware that they can seem equivocal at worst or wishy-washy at best in their attempts to steer a tidy course between the critics who complain that they overestimate the value of autonomy and critics who complain that they have not understood that human freedom is a curse and a source of anguish rather than an achievement. It is too late in the day to rely on Aristotle’s claim that the truth of these matters is to be found in the mean between extremes, but the liberal can at any rate reply that there is no more reason to suppose that it lies in the extremes than in the ground between them that liberalism occupies.

A Theory for Society

It is a common complaint against liberalism that it undervalues the role of community. Since the late 1970s, this has been a constant refrain, but it replicates the complaints made by critics of philosophical radicalism in the early 1800s and by philosophical idealists in the late 1800s (Sandel 1982). One response to the complaint might be to list those liberals who took the role of the community entirely seriously—they include Tocqueville, Mill, Green, Hobhouse, Émile Durkheim, William James, and Dewey. This is only the starting point for an answer to the question whether liberalism has or even can have a liberal theory of society. The answer is plainly that it can and indeed that it does. In fact, one might argue that it is only because liberals are so impressed by the ways in which society molds and shapes the lives of its members that liberals are so eager to ensure that society does not also cramp and distort those lives.

Sociologists used to claim that their opponents were attached to a contractual account of society, and by this they meant that their opponents believed that society literally had its origins in some kind of agreement. Although it is plain that no contemporary liberal would think anything of the sort, it is true that liberals find it illuminating to think of society as if it involved a sort of contract. The authority of the group over the individual is (p.38) not absolute, but extends only to the hypothetical terms of a bargain by which individuals agree to accept that authority (Nagel 1991, 33ff.). The terms of the bargain are what remain in dispute. In his essay On Liberty, Mill essentially treated it as a compact for self-protection. Society was, as it were, a device for lending individuals the force of the whole group in fending off attacks on their persons and property (Mill [1861] 1974, 119–22).

This covered only the coercive authority of society. A more elusive topic was what a liberal society would look like, going beyond the question of what rules it might properly enforce on its members. Just as in the case of its account of the values that give point to an individual existence, liberalism is to some extent hampered in giving a very rich account by its attachment to the value of choice. Once we have said that a society full of liberals would be replete with voluntary associations devoted to enhancing the existences of all their members, there is little more to say. We may agree that a liberal would think it desirable that stamp collectors should get together and discuss their enthusiasms, exchange stamps, circulate journals about their hobby, and all the rest, but it defies the imagination to offer a liberal theory of philately.

Liberals would object strongly to any regime that made philately difficult—it would be a pointless interference with liberty—and would divide on the question whether a government might properly assist philatelic societies to get started by a temporary subsidy, as liberals have always been divided in their attitudes to governmental assistance for art, education, and high culture. Beyond that, the liberal answer to the question of what a society attached to liberal principles would actually look like is that the answer is a matter for the society in question. It might have many churches or none, a multitude of different schooling systems or one, an effective public transport system or not; what would matter would be that the individual liberty or human rights of its members were respected in the process of reaching these outcomes. In particular, liberalism is agnostic about what the implementation of the vision of a society of free individuals entails for the economic arrangements it embraces. Certainly, too many state controls threaten liberty, just as a state monopoly of employment threatens liberty (Rawls 1971, 377ff.). So does a capitalism that allows rich men to buy politicians. Where the best feasible regime lies is a matter for experiment.

A Theory for the State

What applies to society does not apply in the same way to the state. Society is the realm of both informal and formal associations, a realm in which public opinion plays some coercive role, but there is much scope for voluntary association; in a manner of speaking, society is a plurality of smaller societies. The state is essentially the realm of coercively sanctioned coordination, and its essence is that it has no competitors or alternatives. That a liberal state must operate according to the rule of law goes without saying; that it must employ as little coercion as possible in its dealings with its citizens (p.39) also goes without saying. What is more hotly contested is whether liberalism dictates any particular form of government.

Historically, liberals have thought at one time that liberalism was threatened by democracy, and at another that liberalism entailed democracy. What liberalism is always committed to is constitutional government. Except in emergencies, when the preservation of a liberal regime may force governments to take powers that would otherwise be intolerable, the requirements of the rule of law extend to the ways in which governments acquire power and exercise it. How this is best achieved has no fixed answer. It is an ongoing argument whether the British view that governments are kept liberal by public opinion and fear of the voter is more or less plausible than the American view that a written constitution and a formal Bill of Rights are uniquely effective. It is more than plausible that such institutional devices as an independent judiciary, a diverse and free press, and a great variety of watchdog organizations such as the U.S. Council for Civil Liberties are all of them useful, and that one needs both the formal protections of American constitutionalism and a liberal-minded citzenry that makes them more than parchment barriers to oppression (Madison, Hamilton, and Jay 1987, Federalist 48, 309).

This leaves the connection between liberalism and democracy for further analysis. If democracy is just a matter of majority rule, it is a contingent matter whether the majority will generally subscribe to liberal views. If it does, there will be a liberal democracy; if not, not. Various devices may be set up to restrain the majority, such as an entrenched Bill of Rights, but all such devices favor liberty by restricting democracy. They are intrinsically undemocratic insofar as they restrict the authority of the majority. On the whole, this view was the view of Jefferson, Tocqueville, and Mill, who were correspondingly anxious to educate the fledgling democracy of their day in order that democracy should not be majority tyranny (Mill [1861] 1974, 62ff.; Tocqueville [1835] 1964, 269ff.).

The alternative view is that liberalism is committed to democracy and that illiberal democracy is not democracy at all. Each individual has a right to take part in the decisions that affect their society. Nobody ought to be governed without their voice being heard, for that is a violation of their human rights or of their right to be treated as a free and equal member of their society (Dworkin 1985, 193ff.). To the objection that majority rule may be inconsistent with liberty, the sophisticated reply is essentially that the authority, as distinct from the power, of the majority is intrinsically self-limiting. We cannot claim the right to vote, for instance, on terms that violate others’ rights. In this view, the Bill of Rights does not limit the majority’s authority so much as spell out what its authority is. Liberal democracy is not something one may realize if one is lucky; the only legitimate democracy is liberal democracy.

However we decide between these two conceptualizations, liberal government must be limited government. Freedom of conscience, freedom of occupational choice, privacy, and family rights all place limits on what governments (p.40) may do. Limited government may nonetheless be active government; securing these rights will keep government busy. More to the point, liberal governments will inherit many illiberal arrangements from their predecessors. Abolishing racial and sexual discrimination in the United States has been neither quick nor easy. Reducing the effects of inherited privilege in the United Kingdom has hardly begun. A government that takes liberalism seriously will be a busy government, especially since it will also have to be ingenious in pursuing its goals through lawful channels.

On this point, defenders of classical and modern forms of liberalism can agree. Both deplore the advantages of monopolists; sexual and racial discrimination, and the advantages of inherited position, share in the wickedness of monopolies, for they give undeserved advantages to their beneficiaries and undeserved handicaps to their victims. It may be that classical liberals suppose that once a “level playing field” has been achieved, it will remain level, while modern liberals suppose that it will need constant attention. It is certainly true that modern liberals emphasize the “equal” in equal opportunity, whereas their predecessors perhaps stress “opportunity” and have no particular liking for equality of any other kind. Still, the point remains that limited governments need not be inactive or lazy governments.