The following is based on an essay appearing in Isaiah Berlin: A Celebration, edited by Edna and Avishai Margalit, to be published by the University of Chicago Press.

When Isaiah Berlin delivered his famous inaugural lecture as Chichele Professor of Social and Political Theory at Oxford, in 1958, he felt it necessary to acknowledge that politics did not attract the professional attention of most serious philosophers in Britain and America. They thought philosophy had no place in politics, and vice versa; that political philosophy could be nothing more than a parade of the theorist’s own preferences and allegiances with no supporting arguments of any rigor or respectability. That gloomy picture is unrecognizable now. Political philosophy thrives as a mature industry; it dominates many distinguished philosophy departments and attracts a large share of the best graduate students almost everywhere.

Berlin’s lecture, “Two Concepts of Liberty,” played an important and distinctive role in this renaissance. It provoked immediate, continuing, heated, and mainly illuminating controversy. It became, almost at once, a staple of graduate and undergraduate reading lists, as it still is. Its scope and erudition, its historical sweep and evident contemporary force, its sheer interest, made political ideas suddenly seem exciting and fun. Its main polemical message—that it is fatally dangerous for philosophers to ignore either the complexity or the power of those ideas—was both compelling and overdue. But chiefly, or so I think, its importance lay in the force of its central argument. For though Berlin began by conceding to the disdaining philosophers that political philosophy could not match logic or the philosophy of language as a theater for “radical discoveries,” in which “talent for minute analyses is likely to be rewarded,” he continued by analyzing subtle distinctions that, as it happens, are even more important now, in the Western democracies at least, than when he first called our attention to them.

I must try to describe two central features of his argument, though for reasons of space I shall have to leave out much that is important to them. The first is the celebrated distinction described in the lecture’s title: between two (closely allied) senses of liberty. Negative liberty (as Berlin came later to restate it) means not being obstructed by others in doing one might wish to do. We count some negative liberties—like the freedom to speak our minds without censorship—as very important and others—like driving at very fast speeds—as trivial. But they are both instances of negative freedom, and though a state may be justified in imposing speed limits, for example, on grounds of safety and convenience, that is nevertheless an instance of restricting negative liberty.

Positive liberty, on the other hand, is the power to control or participate in public decisions, including the decision how far to curtail negative liberty. In an ideal democracy—whatever that is—the people govern themselves. Each is master to the same degree, and positive liberty is secured for all.

In his inaugural lecture Berlin described the historical corruption of the idea of positive liberty, a corruption that began in the idea that someone’s true liberty lies in control by his rational self rather than his empirical self, that is, in control that aims at securing goals other than those the person himself recognizes. Freedom, on that conception, is possible only when people are governed, ruthlessly if necessary, by rulers who know their true, metaphysical, will. Only then are people truly free, albeit against their will. That deeply confused and dangerous, but nevertheless potent, chain of argument had in many parts of the world turned positive liberty into the most terrible tyranny. Of course, by calling attention to this corruption of positive liberty, Berlin did not mean that negative liberty was an unalloyed blessing, and should be protected in all its forms in all circumstances at all costs. He said, later, that on the contrary the vices of excessive and indiscriminate negative liberty were so evident, particularly in the form of savage economic inequality, that he had not thought it necessary to describe them in much detail.

The second feature of Berlin’s argument that I have in mind is a theme repeated throughout his writing on political topics. He insists on the complexity of political value, and the fallacy of supposing that all the political virtues that are attractive in themselves can be realized in a single political structure. The ancient Platonic ideal of some master accommodation of all attractive virtues and goals, combined in institutions satisfying each in the right proportion and sacrificing none, is in Berlin’s view, for all its imaginative power and historical influence, only a seductive myth. He later summed this up:

One freedom may abort another; one freedom may obstruct or fail to create conditions which make other freedoms, or a larger degree of freedom, or freedom for more persons, possible; positive and negative freedom may collide; the freedom of the individual or the group may not be fully compatible with a full degree of participation in a common life, with its demands for cooperation, solidarity, fraternity. But beyond all these there is an acuter issue: the paramount need to satisfy the claims of other, no less ultimate, values: justice, happiness, love, the realization of capacities to create new things and experiences and ideas, the discovery of the truth. Nothing is gained by identifying freedom proper, in either of its senses, with these values, or with the conditions of freedom, or by confounding types of freedom with one another.1

Berlin’s warnings about conflating positive and negative liberty, and liberty itself, with other values, seemed, to students of political philosophy in the great Western democracies in the 1950s, to provide important lessons about authoritarian regimes in other times and places. Though cherished liberties were very much under attack in both America and Britain in that decade, the attack was not grounded in or defended through either form of confusion. The enemies of negative liberty were powerful, but they were also crude and undisguised. Joseph McCarthy and his allies did not rely on any Kantian or Hegelian or Marxist concept of metaphysical selves to justify censorship or blacklists. They distinguished liberty not from itself, but from security; they claimed that too much free speech made us vulnerable to spies and intellectual saboteurs and ultimately to conquest.


In both Britain and America, in spite of limited reforms, the state still sought to enforce conventional sexual morality about pornography, contraception, prostitution, and homosexuality. Conservatives who defended these invasions of negative liberty appealed not to some higher or different sense of freedom, however, but to values that were plainly distinct from, and in conflict with, freedom: religion, true morality, and traditional and proper family values. The wars over liberty were fought, or so it seemed, by clearly divided armies. Liberals were for liberty, except, in some circumstances, for the negative liberty of economic entrepreneurs. Conservatives were for that liberty, but against other forms when these collided with security or their view of decency and morality.

But now the political maps have radically changed and some forms of negative liberty have acquired new opponents. Both in America and Britain, though in different ways, conflicts over race and gender have transformed old alliances and divisions. Speech that expresses racial hatred, or a degrading attitude toward women, has come to seem intolerable to many people whose convictions are otherwise traditionally liberal. It is hardly surprising that they should try to reduce the conflict between their old liberal ideals and their new acceptance of censorship by adopting some new definition of what liberty, properly understood, really is. It is hardly surprising, but the result is dangerous confusion, and Berlin’s warnings, framed with different problems in mind, are directly in point.

I shall try to illustrate that point with a single example: a lawsuit arising out of the attempt by certain feminist groups in America to outlaw what they consider a particularly objectionable form of pornography. I select this example not because pornography is more important or dangerous or objectionable than racist invective or other highly distasteful kinds of speech, but because the debate over pornography has been the subject of the fullest and most comprehensive scholarly discussion.

Through the efforts of Catharine MacKinnon, a professor of law at the University of Michigan, and other prominent feminists, Indianapolis, Indiana, enacted an antipornography ordinance. The ordinance defined pornography as “the graphic sexually explicit subordination of women, whether in pictures or words…” and it specified, as among pornographic materials falling within that definition, those that present women as enjoying pain or humiliation or rape, or as degraded or tortured or filthy, bruised or bleeding, or in postures of servility or submission or display. It included no exception for literary or artistic value, and opponents claimed that applied literally it would outlaw James Joyce’s Ulysses, John Cleland’s Memoirs of a Woman of Pleasure, various works of D.H. Lawrence, and even Yeats’s “Leda and the Swan.” But the groups who sponsored the ordinance were anxious to establish that their objection was not to obscenity or indecency as such, but to the consequences for women of a particular kind of pornography, and they presumably thought that an exception for artistic value would undermine that claim.2

The ordinance did not simply regulate the display of pornography so defined, or restrict its sale or distribution to particular areas, or guard against the exhibition of pornography to children. Regulation for those purposes does restrain negative liberty, but if reasonable it does so in a way compatible with free speech. Zoning and display regulations may make pornography more expensive or inconvenient to obtain, but they do not offend the principle that no one must be prevented from publishing or reading what he or she wishes on the ground that its content is immoral or offensive.3 The Indianapolis ordinance, on the other hand, prohibited any “production, sale, exhibition, or distribution” whatever of the material it defined as pornographic.


Publishers and members of the public who claimed a desire to read the banned material arranged a prompt constitutional challenge. The federal district court held that the ordinance was unconstitutional because it violated the First Amendment to the United States Constitution, which guarantees the negative liberty of free speech.4 The Circuit Court for the Seventh Circuit upheld the district court’s decision,5 and the Supreme Court of the United States declined to review that holding. The Circuit Court’s decision, in an opinion by Judge Easterbrook, noticed that the ordinance did not outlaw obscene or indecent material generally but only material reflecting the opinion that women are submissive, or enjoy being dominated, or should be treated as if they did. Easterbrook said that the central point of the First Amendment was exactly to protect speech from content-based regulation of that sort. Censorship may on some occasions be permitted if it aims to prohibit directly dangerous speech—crying fire in a crowded theater or inciting a crowd to violence, for example—or speech particularly and unnecessarily inconvenient—broadcasting from sound trucks patrolling residential streets at night, for instance. But nothing must be censored, Esterbrook wrote, because the message it seeks to deliver is a bad one, or because it expresses ideas that should not be heard at all.

It is by no means universally agreed that censorship should never be based on content. The British Race Relations Act, for example, forbids speech of racial hatred, not only when it is likely to lead to violence, but generally, on the grounds that members of minority races should be protected from racial insults. In America, however, it is a fixed principle of constitutional law that such regulation is unconstitutional unless some compelling necessity, not just official or majority disapproval of the message, requires it. Pornography is often grotesquely offensive; it is insulting, not only to women but to men as well. But we cannot consider that a sufficient reason for banning it without destroying the principle that the speech we hate is as much entitled to protection as any other. The essence of negative liberty is freedom to offend, and that applies to the tawdry as well as the heroic.

Lawyers who defend the Indianapolis ordinance argue that society does have a further justification for outlawing pornography: that it causes great harm as well as offense to women. But their arguments mix together claims about different types or kinds of harm, and it is necessary to distinguish these. They argue, first, that some forms of pornography significantly increase the danger that women will be raped or physically assaulted. If that were true, and the danger were clear and present, then it would indeed justify censorship of those forms, unless less stringent methods of control, such as restricting pornography’s audience, would be feasible, appropriate, and effective. In fact, however, though there is some evidence that exposure to pornography weakens people’s critical attitudes toward sexual violence, there is no persuasive evidence that it causes more actual incidents of assault. The Seventh Circuit cited a variety of studies (including that of the Williams Commission in Britain in 1979) all of which concluded, the court said, “that it is not possible to demonstrate a direct link between obscenity and rape….”6 A recent report based on a year’s research in Britain said: “The evidence does not point to pornography as a cause of deviant sexual orientation in offenders. Rather, it seems to be used as part of that deviant sexual orientation.”7

Some feminist groups argue, however, that pornography causes not just physical violence but a more general and endemic subordination of women. In that way, they say, pornography makes for inequality. But even if it could be shown, as a matter of causal connection, that pornography is in part responsible for the economic structure in which few women attain top jobs or equal pay for the same work, that would not justify censorship under the Constitution. It would plainly be unconstitutional to ban speech directly advocating that women occupy inferior roles, or none at all, in commerce and the professions, even if that speech fell on willing male ears and achieved its goals. So it cannot be a reason for banning pornography that it contributes to an unequal economic or social structure, even if we think that it does.

But the most imaginative feminist literature for censorship makes a further and different argument: that negative liberty for pornographers conflicts not just with equality but with positive liberty as well, because pornography leads to women’s political as well as economic or social subordination. Of course pornography does not take the vote from women, or somehow make their votes count less. But it produces a climate, according to this argument, in which women cannot have genuine political power or authority because they are perceived and understood unauthentically—that is, they are made over by male fantasy into people very different from, and of much less consequence than, the people they really are. Consider, for example, these remarks from the work of the principal sponsor of the Indianapolis ordinance. “[Pornography] institutionalizes the sexuality of male supremacy, fusing the eroticization of dominance and submission with the social construction of male and female….Men treat women as who they see women as being. Pornography constructs who that is. Men’s power over women means that the way men see women defines who women can be.”8

Pornography, on this view, denies the positive liberty of women; it denies them the right to be their own masters by recreating them, for politics and society, in the shapes of male fantasy. That is a powerful argument, even in constitutional terms, because it asserts a conflict not just between liberty and equality but within liberty itself, that is, a conflict that cannot be resolved simply on the ground that liberty must be sovereign. What shall we make of the argument understood that way? We must notice, first, that it remains a causal argument. It claims not that pornography is a consequence or symptom or symbol of how the identity of women has been reconstructed by men, but an important cause or vehicle of that reconstruction.

That seems strikingly implausible. Sadistic pornography is revolting, but it is not in general circulation, except for its milder, soft-porn manifestations. It seems unlikely that it has remotely the influence over how women’s sexuality or character or talents are conceived by men, and indeed by women, that commercial advertising and soap operas have. Television and other parts of popular culture use sexual display and sexual innuendo to sell virtually everything, and they often show women as experts in domestic detail and unreasoned intuition, and nothing else. The images they create are subtle and ubiquitous, and it would not be surprising to learn, through whatever research might establish this, that they indeed do great damage to the way women are understood and allowed to be influential in politics. Sadistic pornography, though much more offensive and disturbing, is greatly overshadowed by these dismal cultural influences as a causal force.

Judge Easterbrook’s opinion for the Seventh Circuit assumed, for the sake of argument, however, that pornography did have the consequences the defenders of the ordinance claimed. He said that the argument nevertheless failed because the point of free speech is precisely to allow ideas to have whatever consequences follow from their dissemination, including undesirable consequences for positive liberty. “Under the First Amendment,” he said, “the government must leave to the people the evaluation of ideas. Bald or subtle, an idea is as powerful as the audience allows it to be…. [The assumed result] simply demonstrates the power of pornography as speech. All of these unhappy effects depend on mental intermediation.”

That is right as a matter of American constitutional law. The Ku Klux Klan and the American Nazi Party are allowed to propagate their ideas in America, and the British Race Relations Act, so far as it forbids abstract speech of racial hatred, would be unconstitutional in the US. But does the American attitude represent the kind of Platonic absolutism Berlin warned against? No, because there is an important difference between the idea he thinks absurd, that all ideals attractive in themselves can be perfectly reconciled within a single utopian political order, and the different idea he thought essential, that we must, as individuals and nations, choose, among possible combinations of ideals, a coherent, even though inevitably and regrettably limited, set of these to define our own individual or national way of life. Freedom of speech, conceived and protected as a fundamental negative liberty, is the core of the choice modern democracies have made, a choice we must now honor in finding our own ways to combat the shaming inequalities women still suffer.

This reply depends, however, on seeing the alleged conflict within liberty as a conflict between the negative and positive senses of that virtue. We must consider yet another argument which, if successful, could not be met in the same way, because it claims that pornography presents a conflict within the negative liberty of speech itself. Berlin said that the character, at least, of negative liberty was reasonably clear, that although excessive claims of negative liberty were dangerous, they could at least always be seen for what they were. But the argument I have in mind, which has been offered by, among others, Frank Michelman of the Harvard Law School, expands the idea of negative liberty in an unanticipated way. He argues that some speech, including pornography, may be itself “silencing,” so that its effect is to prevent other people from exercising their negative freedom to speak.

Of course it is fully recognized in First Amendment jurisprudence that some speech has the effect of silencing others. Government must indeed balance negative liberties when it prevents heckling or other demonstrative speech designed to stop others from speaking or being heard. But Michelman has something different in mind. He says that a woman’s speech may be silenced not just by noise intended to drown her out but also by argument and images that change her audience’s perceptions of her character, needs, desires, and standing, and also, perhaps, change her own sense of who she is and what she wants. Speech with that consequence silences her, Michelman supposes, by making it impossible for her effectively to contribute to the process Judge Easterbrook said the First Amendment protected, the process through which ideas battle for the public’s favor. “[It] is a highly plausible claim,” Michelman writes, “[that] pornography [is] a cause of women’s subordination and silencing…. It is a fair and obvious question why our society’s openness to challenge does not need protection against repressive private as well as public action.”9

He argues that if our commitment to negative freedom of speech is consequentialist—if we want free speech in order to have a society in which no idea is barred from entry, then we must censor some ideas in order to make entry possible for other ones. He protests that the distinction that American constitutional law makes between the suppression of ideas by the effect of public criminal law and by the consequences of private speech is arbitrary, and that a sound concern for openness would be equally worried about both forms of control. But the distinction the law makes is not between public and private power as such, but between negative liberty and other virtues, including positive liberty. It would indeed be contradictory for a constitution to prohibit official censorship while also protecting the right of private citizens physically to prevent other citizens from publishing or broadcasting specified ideas. That would allow private citizens to violate the negative liberty of other citizens by preventing them from saying what they wish.

But there is no contradiction in insisting that every idea must be allowed to be heard, even those whose consequence is that other ideas will be misunderstood, or given little consideration, or even not be spoken at all because those who might speak them are not in control of their own public identities and therefore cannot be understood as they wish to be. These are very bad consequences, and they must be resisted by whatever means our Constitution permits. But acts that have these consequences do not, for that reason, deprive others of their negative liberty to speak, and the distinction, as Berlin insisted, is very far from arbitrary or inconsequential.

It is of course understandable why Michelman and others should want to expand the idea of negative liberty in the way they try to do. Only by characterizing certain ideas as themselves “silencing” ideas—only by supposing that censoring pornography is the same thing as stopping people from drowning out other speakers—can they hope to justify censorship within the constitutional scheme that assigns a preeminent place to free speech. But the assimilation is nevertheless a confusion, exactly the kind of confusion Berlin warned against in his original lecture, because it obscures the true political choice that must be made. I return to Berlin’s lecture, which put the point with that striking combination of clarity and sweep I have been celebrating:

I should be guilt-stricken, and rightly so, if I were not, in some circumstances, ready to make [some] sacrifice [of freedom]. But a sacrifice is not an increase in what is being sacrificed, namely freedom, however great the moral need or the compensation for it, Everything is what it is: liberty is liberty, not equality or fairness or justice or culture, or human happiness or a quiet conscience.

This Issue

August 15, 1991