I
In 1993, in the early months of the presidency of Bill Clinton, the Clinton administration led a public campaign against the increasing quantity and extremity of portrayals of violence to be found on television, in motion pictures, and in popular music.1 President Clinton, his wife Hillary Rodham Clinton, and Attorney General Janet Reno all attacked much of the mass media for portraying violence both frequently and favorably, and thus for contributing to the level of violence existing in the United States.
In response to the president's public campaign against media violence, the entertainment industry, and especially the major television networks, all made the denial of the president's causal claim the centerpiece of their defense. Televised violence is causally inert, they claimed, at worst a symptom and not a cause of a violent society, and at best simply a form of entertainment with no effect on people's actual behavior. Consequently, they argued, efforts to control television and other media violence, whether through the law or through the forces of public opinion, would be an ineffectual use of public resources that would do little, if anything, to reduce the level of crime and violence.
The rhetoric of the entertainment industry regarding portrayals of violence is usefully contrasted with the rhetoric of the American Civil Liberties Union and others in their support, over the years, of the rights of racist and antisemitic organizations, most notoriously the Ku Klux Klan and the American Nazi Party, to speak, to march, to demonstrate, to parade, and to organize.2 Unlike the claim of causal inertness deployed by the defenders of television and motion picture violence, the American Civil Liberties Union and others have not, in their defense of Nazis and the Ku Klux Klan,3 claimed that the words, pictures, and symbols of such groups have no negative consequences. The defenders acknowledge that the racist and antisemitic images and discourse of these groups can harm the individuals against whom racist and antisemitic invectives might be directed, can harmfully corrupt the level and nature of civic discourse, and can at times increase the probability of violent and unlawful acts being committed against people on account of their race or religion.
Yet in these cases, the admission, by the American Civil Liberties Union and by the courts, of speech's causal propensities, and the admission of speech's harmful consequences, has not lessened the strength of the free speech claim. And that is because the free speech rights of Nazis and the Ku Klux Klan are not a consequence of the ineffectualness or the harmlessness of their utterances. Rather, the Nazis and their ilk have free speech rights not because what they say is harmless, but despite the harm they cause by what they say. Whether it be because allowing harmful speech fosters a desirable attitude of tolerance,4 or leads to the increased discovery of truth,5 or respects the process of democratic participation in public deliberation,6 or manifests the self-expression,7 self-realization, or autonomy8 of speaker or hearer, the values underlying the right to free speech are values not themselves derived from the fact of speech's harmlessness and are not therefore undercut by the fact of speech's harmfulness on particular occasions.9
II
This debate between those who defend speech because it is harmless10 and those who defend speech despite the harm it may cause reflects a deeper division in moral and political theory more generally. And this division may be thought of in terms of two different pictures of rights, pictures that are largely coincident with the two different postures about free speech I have just illustrated.
One picture of rights, the one that connects with the claims of the defenders of television and motion picture violence, is that rights are what we have to keep bad people from doing bad things. Under this view, there are bad people in the world, or at least there are people who do bad things, and often such people occupy positions of power, commonly in the government. Sometimes these people do bad things by putting their own selfish interests ahead of the interests of others, or ahead of the common good. Sometimes they do bad things out of ill will or spite. And sometimes these people may be well-meaning but do bad things out of ignorance of the facts or confusion about goals. Under one picture of rights, rights address these pathologies, and rights are the instruments that block the implementation of such ill-guided or simply confused policies or practices. Rights keep authoritarian police officers from targeting innocent suspects, they keep corrupt politicians from implementing policies for private financial gain, they keep power-hungry ideologues from imposing their views and their morals on others, they keep policymakers from enacting policies designed only to entrench the personal power of the policymakers, and they keep those in power from discriminating on the basis of race, religion, gender, sexual orientation, and the like for no reason other than a dislike of those who are different from themselves.
This picture of rights is well known, in part from the rhetoric routinely employed by various civil liberties organizations as they seek funds to combat these forces of evil, and in part from more serious scholarly efforts to understand rights as impediments to something like what Cass Sunstein has referred to as “naked preferences.”11 But although this picture of rights is well known, one of its necessary implications is far more rarely acknowledged: If rights are what keep bad people from doing bad things, then the recognition and enforcement of rights brings great benefits at no cost. If rights, for example, prevent evil police officers from pursuing the innocent, and if the pursuit of the innocent brings no social benefit, then keeping evil police officers from pursuing innocent citizens is an unalloyed social good. Similarly, if the diversion of public funds to the pockets of corrupt politicians benefits no one except the corrupt politician, then rights that might prevent corrupt politicians from engaging in self-serving behavior would again bring nothing but good.12 And, finally, if rights prevent people from imposing their moral or religious views on others or discriminating for no reason other than irrational prejudice, then preventing the moral imposition or preventing the irrational discrimination is a public benefit with no costs except perhaps to the unjustifiable preferences of a small number of people, preferences that themselves bring no good and are thus justifiably stifled.
Now let us consider a different picture of rights, one partly captured by Ronald Dworkin's idea of rights as trumps13, and by Robert Nozick's related formulation in terms of side constraints.14 In both of these versions, the central idea is a Kantian conception of rights as impediments to otherwise optimizing welfare or utility maximization. More simply, rights represent limits on the means that may be used to pursue certain ends. If our baseline principle of policy-making is some variety of utilitarian or consequentialist welfare maximization, it may turn out that at times the aggregate or general welfare might be maximized by engaging in, for example, torture, or murder, or racial discrimination, or gender discrimination, and so on. Yet if there are certain things that it is simply wrong to do, and if people have rights that mirror these wrongs such that they have rights not to be wronged in this way, then people will have rights against certain things being done to them even if doing those things will actually increase the aggregate welfare or utility.15 So if it is wrong to torture, to take the stock example, and if people have rights not to be tortured, then they have rights not to be tortured even if and when the torture will, say, reveal the plans for terrorist attacks and thus increase the aggregate welfare. Similarly, if people have a right to the privacy of their home, they may, as in the United States under the Fourth Amendment to the Constitution, have a right to impose high burdens on police officers who would seek to enter the home for the purpose of obtaining evidence and thus impose costs on the apprehension of crime and consequent costs on society at large.
Although the side-constraint picture of rights is commonly associated with nonconsequentialist moral theories of the kind advanced by Kant, Nozick, and Dworkin, a very similar picture can be derived from rule consequentialism as well. It might be the case that the complexity of calculation, the distortions of ignorance, or the complications of coordination among agents with different interests would cast doubt on the ability of individuals to engage in successful act-based consequentialist calculations. When that is so, various rules might be put in place to prevent such likely suboptimal case-by-case calculations and thus maximize the aggregate welfare in the long term even if not in every individual case.16 One form of such a rule is a right, and thus a picture of rights that prevents individual decision makers from engaging in case-by-case decision making, but does so only in the service of long-run utility and not in the service of deep deontological commitments, can still be understood on the side-constraint model. For here, jut as in the deontological model, the rights will prevent even well-meaning decision makers from taking decisions that, in individual cases and in the short run, even if not in the long run, might produce a net increase in the general welfare.
Under this side-constraint picture of rights, whether of rule utilitarian or of deontological origin, rights are no longer exclusively targeted at bad people doing bad things but are instead constraints on even the welfare-increasing decisions of well-meaning and well-informed decision makers. Rights are no longer just an unqualifiedly desirable impediment to the evil and the ill-informed, but an impediment to what appear to be wise policies, an impediment whose virtues are either virtues in and of themselves independent of consequences, as in the deontological version, or virtues whose long-run benefits are less likely to be perceived in the face of more salient short-term costs.
III
As I suggested above, the first picture of rights is one under which rights have benefits but no substantial costs, since the costs (other than enforcement costs) of preventing bad people from doing bad things are negligible. To oversimplify only slightly, under this picture rights are free. Under the second picture, however, especially in its deontological and not its rule-utilitarian version, rights constrain the process of welfare maximization and thus produce a world in which there is less utility than there would otherwise be. When this is so, it is no longer possible to say that rights are free, because rights have a quite tangible cost. If the rights of people who might be charged with a crime impede even good police officers in the use of legitimate law enforcement methods, then the consequence is either more crime, or more resources used to produce the same crime level. In either case the rights come at a social price. So too with the price of a restriction on torture, or with recognition of the rights of religious practice, even those religious practices that many others would find intrusive.17 If there is a right not to be discriminated against on account of age, and if individualized ability testing is more costly than categorial age qualifications and disqualifications, then again recognition and enforcement of the right comes only at a social price.
Let us look more closely at the question of cost, or price. Suppose I buy a bicycle. If I do so, I have less money, and that fact causes me some distress. But I also have the new bicycle, and so it is often the case that the pain I feel from having less money is counterbalanced by the pleasure I gain from what the money has purchased. But suppose instead I buy you a new bicycle or, even worse, am forced against my will to buy you a new bicycle. In that case my own utility calculation is in deficit, even though it still may be the case that the pleasure you gain from having the new bicycle is equal to or greater than the pain or utility loss I suffer from having less money.
This little example is directly applicable to the question of the cost of rights. If rights are costly, as I have just argued they frequently are, then someone must pay for them. Often, however, as in the bicycle example, the people who pay for the rights are not the ones who enjoy them, and thus the payers suffer a utility deficit. And if the number of people who pay for rights is greater than the number who benefit, then there is likely to be a net decrease in happiness. And, most importantly, the existence of this net decrease in happiness will occur even if and when recognizing the rights is the right thing to do.
Many of the most familiar rights in liberal democracies fit this mold. Sometimes these rights are a product of judicial action, as when courts recognize, enforce, or create the rights of criminal defendants to be free from warrantless search and seizure or to keep silent in the face of police or prosecutorial questioning, or when courts grant to members of religious groups the privilege of exemption from otherwise generally applicable governmental requirements. Even more commonly, the recognition of rights that have aggregate utility costs comes from legislative action. Laws protecting endangered species or requiring the construction of special facilities for the handicapped are examples, as would be the case of legislative recognition of minority languages. In all of these cases, and others, there is likely to be a general loss of aggregate utility, even though many people, including myself, believe that recognizing such rights is the right thing to do.
If the recognition of these and other rights, whether by courts, by legislatures, or by the pressure of public opinion and social sanction, is the right thing to do, then a society is, in the broadest sense, better off for doing so. Yet it is also the case that appreciating the utility cost of recognizing these rights provides a possibly different lens through which to view the worldwide question of satisfaction or dissatisfaction with government. For it is not only the case that rights have costs measured in utility, but also that the number of people paying the cost is typically greater than the number of people enjoying the benefit. If it is the case that public security is decreased because of the rights of those suspected of crimes, then it is quite possible that large numbers of people are somewhat worse off in order that society not do the wrong thing to a smaller number of people. Similarly, the number of people paying to provide special facilities for the handicapped is typically greater than the number of people benefiting from the existence of those facilities. And in most of these cases, the outcome is the same even when we take into account the possibility that in some cases the aggregate of the large utility gain to a smaller number of people is greater than the aggregate of the small utility loss to a larger number of people. It is probably the case with ramps and reserved parking places for the physically handicapped, but in most others—rights of suspects, for example—the weighted utility calculus is still more likely to produce a loss rather than a gain.
If the recognition of rights is costly in just this way, then one consequence is that a society that recognizes such rights will be a society in which large numbers of people not only feel worse off because of the recognition of those rights, but will in fact be worse off for the recognition of those rights. And, to repeat, this is so even when recognition of the rights is the right thing for the society to do. Accordingly, people's having a decreasing level of satisfaction with their level of utility may not be the surprising and unfortunate failing of a government or a society that is doing the wrong thing, but may instead be the unsurprising and inevitable consequence of a situation in which society has embarked on the morally worthy enterprise of paying in the currency of its own aggregate welfare (and, therefore, satisfaction) for the rights of those individuals who are often in the minority.
The calculus of this is often ignored, as I have suggested above, in the contemporary debates about rights. On the one side we tend to see the critics of a strongly rights-protective or rights-based culture. These critics properly identify the cost of rights and then typically (and often but not always justifiably) go on to suggest that the strength or scope of the rights recognized ought to be decreased. And on the other side we have the defenders of rights, who often stress the perniciousness of rights violators and thus often fail to see the cost of rights and fail to understand the consequence of their recognition. All too often, what is missing from the debate is the voice that recognizes the cost of rights and thus recognizes that unhappiness with that cost, and often unhappiness with the government that is seen as the creator or the protector of rights, is the unsurprising by-product of a society that takes rights seriously.
Once we recognize that rights have costs, we can think in different ways about payment of those costs. If we were thinking about the cost of building a new bridge, or building public tennis courts, we might think carefully about who was to pay the cost. Should it be the users, or should it be the entire community? If the latter, then should the cost imposition be progressive or regressive, universal or selective? And so on. Yet once we recognize that rights have costs just as do bridges and tennis courts, we can and should ask the same kinds of questions about the allocation of those costs. Does recognizing the rights of criminal defendants impose costs disproportionately on those likely to be the victims of crime, and, if so, who are those people? What is the incidence of the cost of granting rights to the physically handicapped, or the cost of protecting endangered species?
It is possible that one reason we have thought less than we should have about the allocation of these costs is that we believe that all of us benefit equally from the existence of rights, and that all of us consequently do and should share equally the cost of those rights. But this is simply false. More plausibly, we may believe that some of us enjoy the benefits of certain rights more than others, that some of us pay for some rights more than do others, but that when we put all of the rights together things pretty much work out fairly. But this may be false as well. And if it is, then it may be appropriate to focus as much attention on how we allocate the cost of rights as on how we allocate the cost of other publicly provided or guaranteed goods, services, and facilities. Even if more recognition of utility-trumping rights will inevitably produce less aggregate benefit from and satisfaction with the state, it may still be the case that a fairer allocation of the cost of those rights, and a better discussion of the fact that there are costs to be allocated, will reduce some degree of public dissatisfaction. It will probably not eliminate it. If rights constrain utility maximization, then recognizing rights will produce less utility than would otherwise be the case, and the relationship between less utility and less satisfaction is likely a close one. Still, it is possible that turning attention to the way in which dissatisfaction with government may just be a function of the fact that rights have costs will open up a fruitful new line of inquiry.
IV
In pursuing this inquiry, the right to free speech seems an ideal starting point. And it is so because so much of traditional free speech debate has deflected the question by refusing, as I suggested above, to acknowledge the other regarding and consequent harm-producing capacity of speech. Accordingly, discussions of free speech rights have often taken place within the first model I described above, in which rights are essentially costless because they keep bad people from doing bad things. Yet, increasingly, a combination of scientific evidence and tragic public events has made the claim about the harmlessness of speech more and more difficult to maintain. In terms of specific events, there is the murder of Prime Minister Yitzhak Rabin in Israel and the bombing of the federal building in Oklahoma City in the United States, as well as a larger number of less prominent but also tragic “copycat” crimes, all of which make it harder to deny the fact that those with violent propensities are often encouraged to manifest those propensities by acts of speech, and that the propensities themselves may have been created, at least in part, by the public communications of others.18 Moreover, scientific studies of the effect of written and printed materials endorsing racial violence, violence against women, and violence generally have tended to support the proposition that the incidence of such materials bears a causal relationship to the level of parallel violent propensities or attitudes, and that the level of those violent propensities or attitudes in turn bears a relationship to the level of such forms of violence themselves.19
The issues in the previous paragraph are complex, and the empirical questions surrounding them no less so. For purposes of my analysis here, however, I do not want to delve into them, but rather simply to assume that speech often has the capacity to, and does in fact, increase the probability (in an individual case) and the incidence (for a population) of behavior that is noncontroversially harmful, such as assassination, bombing of public buildings, and rape. To the extent that this is so, the next question would be whether any of the speech producing these effects either is or ought to be legally protected. And once again, addressing this topic is not my primary agenda here. Although there can and should be debates about the extent to which speech that bears a causal relationship to acts of violence should be restricted, and although no country in the world protects such speech as much as it is protected in the United States, it is hard to imagine a robust free speech principle that protected none of such speech. Although Justice Holmes undoubtedly overstated the case in claiming that “every idea is an incitement,”20 it is almost certainly the case that a wide range of strong political and social rhetoric would plausibly be found to be (probabilistically) causally related to the level of illegal and antisocial acts consistent with, even if not directly and immediately inspired by, that rhetoric. It seems safe to assume, therefore, that the existence of any free speech principle, any principle of communicative tolerance, will, ceteris paribus, produce a higher level of uncontroversially harmful activities than would exist without the existence of such a principle.
When we view the issue in this manner, we can thus see, consistent with the side-constraint view of rights I sketched above, that recognizing a right to free speech will produce, especially in the short term but perhaps even in the long term as well, less utility, or less general welfare, than would otherwise exist. If an entrenched free speech principle will produce more harm-encouraging speech than will the absence of such a principle, and if more harm-encouraging speech will produce more harm, then a free speech principle will likely yield more harm and, consequently, less overall welfare or utility. Like a nonutilitarian or rule-utilitarian side constraint on act-based utility maximization, therefore, a right to free speech will not be costless, but can only be bought at some cost, at least in the short term and perhaps in the long term, to the general welfare.
It turns out, therefore, that free speech is a good example of a right that is purchased at some short-term social cost. And if we think that the question of how to pay for a bridge or a tennis court that brings long-term gain at short-term cost is one that is best addressed by considering all of the possible cost payers in a society, then so too might we address the question of the cost of free speech in much the same way.
Most societies with strong free speech principles have not, however, addressed the question in this fashion. Instead, they have simply assumed that the cost must be borne by the victims of harmful speech. In the context of libelous speech, they have, in the interest of keeping the press from being excessively “chilled,” denied remedies to those whose reputations and careers have been damaged by false and public statements about them.21 In the context of speech inciting to racial hatred, they have increased the distress of those who are the targets of racial hatred in order to prevent laws prohibiting the incitement to hatred from inhibiting legitimate political disagreement.22 And in the context of speech bearing a causal relationship to the level of political violence, they have been willing to impose upon those more likely to be the targets of such violence an increased risk of that violence in the interest, again, of not excessively stifling the expression of legitimate political opinion.23 In all of these situations, the assumption has been that the only way to pay for the cost of free speech is to deny a remedy to those who would otherwise (under prevailing rules of tort law) have had one, and thus to impose the cost of free speech on the victims—direct or indirect—of harmful speech.
Yet it is by no means necessary that the costs of free speech must be allocated in this way. Societies, especially the United States, that have denied a remedy to those injured by defamatory or abusive speech, or injured by the consequences of the advocacy of violence, have premised their actions on the plausible hypothesis that speakers or publishers who fear monetary liability will refrain from publishing or speaking in order to reduce the risk of such liability. The optimal level of freedom to speak and publish, it is thought, is achieved by withdrawing the victim's monetary remedy in some situations and restricting it in others.
The converse of this, however, is that if the publisher's or speaker's fear of monetary liability were to be eliminated, then there would be no reason to think that publishers or speakers would engage in a suboptimal level of speech, or, to put the same point slightly differently, there would be no reason to believe that publishers or speakers would have an excessive degree of risk aversion compared to the degree of risk aversion that would produce the socially optimal amount and variety of speech.24 And this suggests that a mechanism whereby victims could be paid but publishers and speakers immunized from liability might achieve the simultaneous goals of freeing speech to the requisite degree while not requiring the victims of harmful speech to pay for that social goal.
If speakers and publishers are not to pay for the injuries they cause, but if those injured are nevertheless to receive compensation, then plainly the compensation must come from elsewhere. This is not as implausible as it might seem. Since freedom of speech and freedom of the press are social goals, one possibility might be for the public that enjoys the benefit of these freedoms to pay for them, in this case by providing some sort of victim's compensation. A number of countries provide for compensation for the victims of crime, the assumption being that the typical committer of a crime is unlikely to be a fruitful target for civil liability. When the target is without funds, therefore, some countries and some American states provide for financial compensation to the victim. An analogy is applicable here. If speakers and publishers are immunized from liability not because they are without funds, but because of the social goal of freeing their speech, the effect is the same from the perspective of the victim, and we could imagine some sort of victim's compensation scheme to recompense, even if partly, the otherwise unrecompensable victims of the harmful speech whose protection is, by stipulation, taken to be a valuable social goal. The compensation would be paid for by the public, and thus we would have the presumably desirable goal of the public paying for the public benefit.
We could also imagine a similar system of cost shifting by means of publicly subsidized media liability insurance, somewhat analogous to the program of federally subsidized flood insurance in the United States. In the United States it is thought socially desirable to encourage people to live in floodplains, areas in which there is a particularly high risk of flooding.25 Because the cost of flood insurance, if controlled strictly by market forces, would be above the cost that most residents could afford, there is a publicly funded scheme of flood insurance subsidies, the effect of which is for the public to support the presumed social goal of allowing people to remain in areas of high flood risk. Much the same could be done with media liability insurance, since such insurance could, in theory, eliminate the financial risk, or the excess financial risk, for the speaker or publisher, while still providing a source of financial recompense for those who might be injured in some way by publisher or speaker behavior.
Even apart from programs of financial recompense to victims, there are creative ways of thinking about shifting the costs of a free speech system. As the events in Israel and Oklahoma City tragically indicate, the risks to political leaders and to government employees (as well as those who use government facilities) increase when a certain form of speech flourishes, and that form of speech is likely to flourish, or flourish to a greater degree, when, in the service of free speech goals, it is not subject to legal liability, whether civil or criminal. Yet there is no reason to believe that the increased risk might not, to some extent, be partly alleviated, and, indeed, that is one of the things that we see, albeit unfortunately after and not before the fact, in the wake of recent events of political assassination, political terrorism, and political violence. We see increased security, we see decreased access of public officials, and we might, perhaps, see a decreased willingness of individuals to enter public life or government employ. Every one of these involves a cost borne by the public, in some cases financial and in others not, and we can even suppose that some of these costs might be better anticipated in advance than they historically have been. But regardless of when they are imposed, it is still possible to imagine that the increased risk of political violence might generate increased security, paid for by the public, resulting, in theory, in the optimal amount of free speech and also the optimal amount of public security, but might also generate increased payment by the public, their payment for the presumed public benefits of a free speech system.
V
My goal here has not been to provide a comprehensive list of the ways in which a society might think about reallocating the cost of communicative tolerance. The few I have suggested—victim's compensation programs; subsidized media liability insurance; and anticipatory increased security to accommodate the expected increase in political violence that comes when speech urging political violence is protected from liability—are but first-cut ideas to prompt a new way of thinking. I have no doubt that others can come up with more and better ones. But in many countries, including my own, the development of free speech thinking has been stifled by the pervasive belief that free speech must be purchased at a cost necessarily paid only by its victims. At times this belief produces undercompensation and underattention to victims, as it has in the United States. At times the belief has produced less free speech than would be socially desirable, as has been the case in many other countries throughout the world. But if we can set aside the belief that free speech must be paid for by its victims, we might wind up with a world in which we are both more attentive to victims and more attentive to free speech. Such a world might be more costly to all of its inhabitants, but it is hardly unreasonable to suppose that such a world would be one with both higher aggregate utility and more equity as well.
NOTES
This is the written version of the Keynote Lecture delivered on 28 January 1997 at the University of Haifa, Israel, on the occasion of the opening of the Conference on Ethics, Law, and Communication in an Era of Political Violence and Extremism. I am grateful to the University for its hospitality, to the other participants for their helpful comments, and especially to Raphael Cohen-Almagor, the organizer of the conference, for inviting me to deliver the Keynote Lecture, and for his support throughout. An earlier version of this essay was presented on 31 October 1995 at the Masters of Public Policy Program, Department of Industrial Engineering, University of Chile, Santiago, Chile, and I have benefited from the comments of Professors Eduardo Engel and Carlos Pena. I am also pleased to acknowledge the support for this essay from the Visions of Governance Project of the Kennedy School of Government, Harvard University, as well as the opportunity to present a version at the Visions Project's 1996 Bretton Woods Conference.
1. The public statements and proposals of the president and the attorney general focused more on television than on motion pictures or popular music, largely because existing American constitutional doctrine makes governmental regulation of broadcast (and, to a lesser extent, cable) television content possible under circumstances in which regulation of motion picture or music CDs would be unconstitutional. See Denver Area Educational Telecommunications Consortium, Inc. v. Federal Communications Commission, 116 S. Ct. 2374 (1996); Federal Communications Commission v. Pacifica Foundation, 438 U.S. 726 (1978); Red Lion Broadcasting Company v. Federal Communications Commission, 395 U.S. 367 (1969). The president, who was primarily urging “voluntary” self-restraint by producers and broadcasters, thus had a more credible threat of official control in the case of television than in the case of motion pictures or popular music, and the existence of this credible threat likely influenced the president's choice of focus.
2. The best-known example is the central role the American Civil Liberties Union played in defending the right of the American Nazi Party to march in Skokie, Illinois, a community with a notably large number of Holocaust survivors among its residents. For the relevant caselaw, see Collin v. Smith, 578 F.2d 1197 (7th Cir.), cert, denied 439 U.S. 915 (1978); Village of Skokie v. National Socialist Party of America, 69 111. 2d 605, 373 N.E.2d 21 (1978). For accounts and analysis, see Lee C. Bollinger, The Tolerant Society: Freedom of Speech and Extremist Speech in America (New York: Clarendon Press, 1986); Aryeh Neier, Defending My Enemy (New York: Dutton, 1979); Donald Downs, “Skokie Revisited: Hate Group Speech and the First Amendment,” 60 Notre Dame Lawyer (1985), 629-64.
3. On the free speech rights of the Ku Klux Klan and related organizations, see Forsyth County v. Nationalist Movement, 505 U.S. 123 (1992); Brandenburg v. Ohio, 395 U.S. 444 (1969); cf R.A. V. v. City of St. Paul, 505 U.S. 377 (1992).
4. Bollinger, The Tolerant Society.
5. John Stuart Mill, On Liberty, ed. D. Spitz (New York: W. W. Norton, 1975); Walter Bagehot, “The Metaphysical Basis of Toleration,” in Literary Studies (London: R. H. Hutton, 1884), 422-38; Abrams v. United States, 250 U.S. 616, 630 (1919) (Holmes, J., dissenting) (“the best test of truth is the power of the idea to get itself accepted in the competition of the market”); William P. Marshall, “In Defense of the Search for Truth as a First Amendment Justification,” 30 Georgia Law Review (1995), 1-39. For a skeptical challenge to this argument for free speech, see Alvin I. Goldman and James C. Cox, “Speech, Truth, and the Free Market for Ideas,” 2 Legal Theory (1996), 1-32.
6. Cass R. Sunstein, The Partial Constitution (Cambridge, Mass.: Harvard University Press, 1993), 197-318; Robert C. Post, “The Constitutional Concept of Public Discourse: Outrageous Opinion, Democratic Deliberation, and Hustler Magazine v. Falwell,” 103 Harvard Law Review (1990), 601-86.
7. C. Edwin Baker, Human Liberty and Freedom of Speech (New York: Oxford University Press, 1989).
8. Thomas Nagel, “Personal Rights and Public Space,” 24 Philosophy and Public Affairs (1995), 83-107; David Strauss, “Persuasion, Autonomy, and Freedom of Expression,” 91 Columbia Law Review (1991), 334-71. For an important critique, see Susan Brison, “The Autonomy Defense of Free Speech,” 106 Ethics, no. 1 (January 1998).
9. For the full elaboration of the claims in the previous two sentences, especially the claim that a Free Speech Principle protects speech not because it is harmless, but despite the harm it may cause, see Frederick Schauer, Free Speech: A Philosophical Enquiry (Cambridge: Cambridge University Press, 1982). For the closely related idea that free speech represents a side constraint on otherwise permissible restrictions, see Thomas Scanlon, “A Theory of Freedom of Expression,” 1 Philosophy and Public Affairs (1972), 204-23.
10. For examples from the academic literature, see Michael D. Bayles, “Mid-Level Principles and Justification,” in J. Roland Pennock and John W. Chapman, eds., NOMOS XXVII: Justification (New York: New York University Press, 1986), 49-67, 54 (freedom of speech “is less likely to interfere with the exercise of other liberties than is, say, liberty of action”); Martin Redish, Freedom of Expression: A Critical Analysis (Charlottesville, Va.: Michie, 1984), 5 (“it is almost certainly true in the overwhelming majority of cases that speech is less immediately dangerous than conduct”). Indeed, the harmlessness of speech is central to Ronald Dworkin's persistent use of freedom of speech as an example of a right whose protection is necessary in order to respect the equal dignity of the speaker. Ronald Dworkin, Taking Rights Seriously (Cambridge, Mass.: Harvard University Press, 1978), 200-203; Ronald Dworkin, Freedom's Law: The Moral Reading of the American Constitution (Cambridge, Mass.: Harvard University Press, 1996), 195-243. For my own argument against what I call the “lesser harm hypothesis,” see Frederick Schauer, “The Phenomenology of Speech and Harm,” 103 Ethics (1993), 635-53.
11. Cass R. Sunstein, “Naked Preferences and the Constitution,” 84 Columbia Law Review (1984), 1689-1738.
12. An example of such a right would be a free press right designed to make it easier for the press to expose official misbehavior. See New York Times Co. v. Sullivan, 376 U.S. 254 (1964); Vincent Blasi, “The Checking Value in First Amendment Theory,” American Bar Foundation Research Journal (1977), 521-96.
13. Dworkin, Taking Rights Seriously, 150-205.
14. Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), 26-53. See also Judith Jarvis Thomson, The Realm of Rights (Cambridge, Mass.: Harvard University Press, 1990).
15. See Bernard Williams, Morality: An Introduction to Ethics (Cambridge: Cambridge University Press, 1972).
16. See David Lyons, “Human Rights and the General Welfare,” 6 Philosophy and Public Affairs (1977), 113-29; Frederick Schauer, Playing by the Rules: A Philosophical Examination of Rule-Based Decision-Making in Law and in Life (Oxford: Clarendon Press, 1991); Frederick Schauer, “A Comment on the Structure of Rights,” 27 Georgia Law Review (1993), 415-34.
17. Consider, for example, the possibility that the right to freedom of religious practice might permit people to engage in socially offensive forms of animal sacrifice that are generally prohibited. The example is suggested by Church of the Lukumi Babalu Aye, Inc. v. City of Hialeah, 508 U.S. 520 (1993).
18. For a good survey of the research, in the context of televised images of violence, see Haejung Paik, “The Effects of Television Violence on Antisocial Behavior: A Meta-analysis,” 21 Communication Research (1994), 516-46. For a sample of the commentary suggesting a link between violence-encouraging public discourse and violence itself, see Phil Duncan, “Aggressive Rhetoric and Civil Violence,” Congressional Quarterly Weekly Report, 6 January 1996, 78. For an analysis in the context of sexual violence, see Frederick Schauer, “Causation Theory and the Causes of Sexual Violence,” American Bar Foundation Research Journal (1987), 737-70.
19. Schauer, “Causation Theory.”
20. Gitlow v. New York, 268 U.S. 652, 671 (1925) (Holmes and Brandeis, JJ., dissenting).
21. See, for example, Ocala Star-Banner Co. v. Damron, 221 So.2d 459 (Fla. Dist. Ct. App. 1969), cause dismissed, 231 So.2d 822 (Fla. 1970), reversed, 401 U.S. 295 (1971), which I discuss at length in “Uncoupling Free Speech,” 92 Columbia Law Review (1992), 1321-57.
22. The incidents in Skokie, Illinois, discussed above provide a central example of this phenomenon, and so too with marches, parades, and demonstrations by neo-Nazis and members of the Ku Klux Klan and various other white supremacist groups.
23. Consider Olivia N. v. National Broadcasting Co., 178 Cal. Rptr. 888 (Ct. App. 1981), cert, denied sub. nom. Niemi v. National Broadcasting Co., 458 U.S. 1108 (1982), in which the California courts denied the victim of a sexual assault a remedy against the television network whose depiction of a sexual assault had, to the standards of causation otherwise sufficient in tort law, caused a virtually identical type of sexual assault.
24. It is true that publishers and speakers might fear criticism as much as civil liability, and that the fear of criticism could inhibit some amount of speech. But criticism is not eliminated even when civil liability is eliminated—indeed, criticism is itself a form of speech—and thus the chilling effect of criticism, albeit real, drops out of the equation.
25. Perhaps more accurately, it is thought socially desirable not to compel people to move out of such areas just because of the prohibitive cost of flood insurance.